UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF GEORGIA ATLANTA DIVISION NAACP, et al., ) ) 4

Transcription

1 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 1 of UNITED STATES DISTRICT COURT FOR THE NORTHERN DISTRICT OF GEORGIA ATLANTA DIVISION NAACP, et al., ) ) 4 Plaintiffs, ) )Case No: 1:17-CV vs. BRIAN KEMP, in his official ) capacity as Secretary of ) State for the State of ) Georgia, ) ) Defendant. ) ) AUSTIN THOMPSON, et al., )TCB-WSD-BBM ) Plaintiffs, ) ) vs. ) ) BRIAN KEMP, in his official ) capacity as Secretary of ) State for the State of ) Georgia, ) ) Defendant. ) Deposition of Expert JOWEI CHEN, Ph.D February 27, :00 a.m. )CONSOLIDATED CASES ) BRYAN CAVE, LLP One Atlantic Center, 14th Floor 1201 W. Peachtree Street, N.W. Atlanta, GA Marianne Vargas, CCR, CVR-M VARGAS REPORTING SERVICES, INC N. Hillbrooke Trace Johns Creek, GA

2 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 2 of APPEARANCES OF COUNSEL: ON BEHALF OF THE PLAINTIFFS GEORGIA STATE CONFERENCE OF THE NAACP, as an organization; LAVELLE LEMON, MARLON REID, LAURETHA CELESTE SIMS, PATRICIA SMITH, COLEY TYSON: JON M. GREENBAUM, ESQ. LAWYERS' COMMITTEE FOR CIVIL RIGHTS UNDER LAW 1401 New York Avenue, N.W. Suite 400 Washington, DC jgreenbaum@lawyerscommittee.org ON BEHALF OF THE PLAINTIFFS JAMAL BROOKS, an individual; AUSTIN THOMPSON, an individual; WAYNE SWANSON, an individual; DARRYL PAYTON, an individual; AUDRA CUNNINGHAM, an individual; SABRINA McKENZIE, an individual; JAMIDA ORANGE, an individual; ANDREA SNOW, an individual; SAMMY ARREY-MBI; LYNNE ANDERSON, an individual; and CORETTA JACKSON, an individual: ARIA C. BRANCH, ESQ. (via telephone) PERKINS COIE, LLP th Street, N.W. Suite 600 Washington, DC abranch@perkinscoie.com ON BEHALF OF THE DEFENDANT THE STATE OF GEORGIA; and BRIAN KEMP, in his official capacity as Secretary of State for the State of Georgia: JOHN "JACK" PARK, JR., ESQ. STRICKLAND, BROCKINGTON, LEWIS, LLP Midtown Proscenium, Suite Peachtree Street, N.E. Atlanta, Georgia jjp@sbllaw.net VARGAS REPORTING SERVICES, INC

3 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 3 of INDEX OF EXAMINATIONS 2 3 JOWEI CHEN PAGE Examination by MR. PARK 4 Examination by MR. GREENBAUM 67 Examination by MR. PARK INDEX OF EXHIBITS DEFENDANT'S NO. DESCRIPTION PAGE 01 Dec. 22, 2017 Original Report, Jowei Chen 7 02 Jowei Chen, Reply Report VARGAS REPORTING SERVICES, INC

4 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 4 of P R O C E E D I N G S 10:03 AM MR. PARK: Would you please swear the witness? (The court reporter swore in the witness.) JOWEI CHEN, Ph.D, having first been duly sworn, was examined and testified as follows: EXAMINATION BY MR. PARK: Q. Would you please state your name for the record? A. I'm Jowei Chen. Q. And how do you spell your name? A. J-O-W-E-I C-H-E-N. Q. Thank you. And you're a Ph.D, right? A. Yes, sir. Q. Dr. Chen, have you been deposed before? A. Yes, sir, I have. Q. And so you know how they work? A. Yes, sir. Q. Just a couple of things. If I ask you a question that you don't understand, would you please VARGAS REPORTING SERVICES, INC

5 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 5 of feel free to ask me to rephrase it and I will attempt to do so? A. Yes, sir. Q. And if you need to take a break at any time, we'll be happy to accommodate you. I'd ask only that you not take a break while a question is pending. A. Yes, sir. Q. What did you do to prepare for your deposition? A. I met with Plaintiffs' counsel last night, yesterday, mostly yesterday evening. I reviewed my two reports in this case, as well, and I reviewed the rebuttal report written by Dr. Alford. Q. When you prepared your report, did you look at anything other than the electoral results and the demographic data? A. Which report are you asking about? Are you asking about my original -- Q. Your original -- A. -- December report? Q. Yes, sir. A. Okay. We'll start there. In preparing my December report, I certainly reviewed electoral data. I also reviewed demographic data, census data. I also reviewed various voter registration and voter history VARGAS REPORTING SERVICES, INC

6 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 6 of turnout files. And I will just add that all of these files were turned over in connection with my report. So I had those files. I also reviewed the deposition transcript of Ms. Gina Wright. I believe there was another document that was -- that outlined legislative reapportionment criteria, I can't remember the precise name of it, but it, too, was referenced in my original expert report, and I reviewed that document, as well. So those are all the documents I can remember off the top of my head right now that I used in connection for the original expert report. I believe there were additional documents for my second report. Q. And do you recall any of those additional documents that you reviewed for your second report? A. Sure. There was an additional deposition transcript that I reviewed and I referenced in my second February report. If you'll just give me a moment, I'll find -- I'll find the precise place where I referenced it. Q. Was it the deposition of Rob Strangia? A. Yes, sir, that's correct. So I reviewed Dr. Strangia's deposition. I, again, reviewed Ms. Gina Wright's deposition, and VARGAS REPORTING SERVICES, INC

7 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 7 of obviously I reviewed Dr. Alford's expert reports, as well as the data files and computer codes turned over in connection with Dr. Alford's rebuttal report. I should add a few more things that I just remembered. For my rebuttal report I, of course, also considered the new election results of the January 2018 Special House Election, and of course I analyzed those, those election results in my rebuttal -- sorry, in my reply report of February. So those were certainly new documents that I didn't have access to that I did not use in my original December report. Q. I'd like to hand you what's been marked as Defendant's Exhibit 1, Dr. Chen. Can you tell me what that is? A. This exhibit is my original December 22nd, 2017 expert report in this case. (Exhibit No. 01 was marked/identified.) BY MR. PARK: Q. Does it appear to be a true and correct copy? A. Yes, sir. Q. Turning your attention to Page 1 of this report, you identify a number of cases. In Missouri VARGAS REPORTING SERVICES, INC

8 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 8 of National Association for the Advancement of Colored People vs. Ferguson-Florissant School District and St. Louis County Board of Election Commissioners, which side did you support? A. I didn't support other side, sir. Q. What was the nature of your expert report in that case? A. I was engaged by the Defendants to work on a rebuttal report. Q. Was a rebuttal report filed? A. Yes, sir. I believe that report was filed, so I worked on that rebuttal report. That's what I was hired to do. Q. What about René Romo, et al. vs. Ken Detzner? A. I was engaged by the Plaintiffs. Q. What about League of Women Voters of Florida and others vs. Detzner? A. The Plaintiffs, sir. Q. What about Raleigh Wake Citizens Association vs. Wake County Board of Elections? A. The Plaintiffs. Q. What about Corrine Brown vs. Detzner? A. I was hired by the Defendant Intervenors. Q. And what was the nature of your work in VARGAS REPORTING SERVICES, INC

9 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 9 of René Romo? Did you prepare -- A. In Romo -- Q. -- an expert report? A. Sure. In Romo vs. Detzner I prepared a few different expert reports. Q. Did you testify? A. I did not testify at trial. Q. Were you deposed? A. Yes, sir, I was. Q. What about in League of Women Voters? A. In The League of Women Voters vs. Detzner, I believe that was the challenge to the Florida Senate map, and, again, I was hired by Plaintiffs. I wrote an expert report. I can't remember right now off the top of my head if I was deposed or not, but I recall that the case was somehow settled -- settled before trial. It was somehow resolved. I'm not sure exactly the nature of how it was resolved. Q. What about Raleigh Wake Citizens Association? A. The Plaintiff's counsel, League of Women Voters -- sorry, Plaintiff's counsel was the Southern Coalition for Social Justice. They hired me. I wrote an expert report, and I believe there was no deposition. I testified at trial in that case. VARGAS REPORTING SERVICES, INC

10 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 10 of That's correct, right. I was not deposed. I testified at trial. Q. And then in Corrine Brown you worked for the Defendant Intervenors; is that right? A. Correct. Defendant Intervenors. I wrote an expert report, and I was not deposed. I don't know whether or not -- I don't believe that ever went to trial. I'm not -- I couldn't tell you off the top of my head. Q. Okay. What about City of Greensboro and others vs. Guilford County Board of Elections? A. Okay. In the Greensboro case I was hired by Plaintiffs, I was deposed, and I -- the case went to trial in, I believe, February of 2017, and I testified at that trial. Q. Okay. What about Common Cause and others vs. Robert Rucho? A. Rucho, yes, sir. Sure, Common Cause vs. Rucho, I wrote an expert report. I was engaged by Plaintiffs. The deposition was, I believe, sometime in the spring of 2017, I think in April of And that case went to trial in October, just this past October, and I testified at trial, as well. Q. And then League of Women Voters of Pennsylvania vs. Commonwealth of Pennsylvania? VARGAS REPORTING SERVICES, INC

11 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 11 of A. Sure. The Plaintiff's counsel hired me in that case to write an expert report, as well as a supplemental report later on. I was not deposed, and I testified at trial this past December in that case. Q. Have you ever worked with Maptitude? A. I'm familiar with the software. And I've used it just a little bit just to become familiar with it, but I don't use it in my regular research process. Q. In your familiarity with Maptitude, do you know whether the Pending Changes box reflects a user's choice? A. Whether the -- I'm not sure. If you could clarify for me what you mean by the "pending box." Q. Do you know anything about the Pending Changes box that are discussed in some of the depositions? A. I'm not exactly sure what you're referring to. MR. PARK: Can we go off just a second? (An off-the-record discussion was held.) BY MR. PARK: Q. Do you recall that in her deposition Gina Wright answered a question about the Pending Changes box by saying, "When I work on a plan, I use the VARGAS REPORTING SERVICES, INC

12 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 12 of Pending Changes box, which is a feature in Maptitude that shows how your numbers change as you select what geography you're selecting"? A. I don't recall reading that portion of Ms. Wright's deposition, but I take your word for it. Q. Okay. But you can't tell me what -- When you've used Maptitude, you haven't relied on the Pending Changes box at all? A. Again, I would just reiterate that I don't use this -- I don't use Maptitude as part of my normal research process. I accept that you're describing a feature of Maptitude, and I don't use it very commonly as part of my normal research process, so I couldn't give you first-hand information about exactly what that box looks like. Q. I d like to, if you would, on page -- if you would turn to Page 4 of your report. A. Yes, sir. Q. And you describe your methodology. You say that you used ecological inference? A. Yes, sir. Q. And you use a procedure known as Maximum Likelihood Estimation, combined with Duncan and Davis' 1953 Method of Bounds? A. Yes, sir. VARGAS REPORTING SERVICES, INC

13 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 13 of Q. That's not a Bayesian approach, is it? A. It's a little bit different. Q. Is the procedure you're using like what King describes? A. It is exactly what Gary King describes. It is -- In fact, I used his software package, his computer code. So it is a package or set of computer codes, code files, called EI, and it's a program for our programming language. And professor King makes that code, that computer code readily available, and that is what I used for all of my ecological inference calculations in this report, as well as in my February response report. Q. When you reviewed the results, how many runs of data did you do? A. If I could ask you to clarify your question. Q. Well, when you put your data in and you run it through the software, that would be one iteration, right? A. You're saying if I analyze one election and run it once, that would be one iteration. That's what you want to refer to by "iteration"? Q. Well, in order to come up with your estimates in, say, Table 2, for example, in the 2012 VARGAS REPORTING SERVICES, INC

14 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 14 of House election, EI on Page 6 of your report, just turning to it, you report 98.2 percent of -- And is that a single run through the software, or is it -- how many -- how many runs through the software do you do to get your -- A. Sure. Q. -- confidence level? A. Each coefficient that you see here on Table 2, it represents one run of the software. And, again, that software was the EI package that I referred to a moment ago. Q. Given that you say each of these coefficients represents one run of the software, is that -- did I understand that correctly? A. Yes, sir, one run of the software. Q. You don't burn any, do you? A. If I could ask you to define what you mean by "burn." Q. Do you discard any of your runs? A. I do not run the software and then discard the run that I just did. That would not be part of my normal research practice, or any social scientist's normal research practice. Q. If you turn back to Page 2 of your -- Page 3 of your report, in the middle of the paragraph VARGAS REPORTING SERVICES, INC

15 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 15 of that starts "finally," there's a sentence toward the end: "I found that the Legislature's primary map drawer for the 2015 Plan had access only to racial data, but not partisan data, at the sub-precinct level;" is that right? A. Let me just orient myself to where you are. You're on Page 2? Q. Page 2, Paragraph "finally," then the line starts "account," and the sentence starts toward the end of it. So it's about seven or ten lines down. A. You're on -- Q. On Page 3. A. Okay. You're on Page 3. MR. PARK: Let's go off. (An off-the-record discussion was held.) MR. PARK: Let's go back on the record. BY MR. PARK: Q. On Page 3 in the paragraph that starts "finally," you have the sentence that says, "I found that the Legislature's primary map drawer for the 2015 Plan had access only to racial data, but not partisan data, at the sub-precinct level." A. Yes, sir, I see that sentence. Q. Who do you understand to be the VARGAS REPORTING SERVICES, INC

16 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 16 of Legislature's primary map drawer? A. When I wrote this original report -- and you're referring to me -- you're referring me to a sentence here from my December report -- I had, at that time, only reviewed the deposition transcript of Ms. Gina Wright. And so I was basing this sentence on Ms. Gina Wright -- Ms. Wright's deposition transcript regarding her process in drawing the map. Q. And that was based on your understanding of Ms. Wright's testimony? A. It was -- Yes, sir. It was based on my reading of Ms. Wright's deposition. Q. I d like you to turn to Page 5, pages 5 and 6 of your report. There are tables that address your ecological inference and ecological regression estimates for House District 105 and 111; is that right? A. Yes, sir. Q. Looking at the ecological regression estimates for Black for each of the three elections in Table 1, you report 100 percent with 100 percent confidence limits. A. Yes, sir. Q. Does that mask a result that was over 100 percent? VARGAS REPORTING SERVICES, INC

17 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 17 of A. How we report ecological regression estimates is that if the estimate is over 100 percent, then you report 100 percent, or you round it down to 100 percent, because that is logically the maximum -- the maximum coefficient that can explain the data. So I wouldn't say that it masks, but certainly that is -- that is what we do when the regression coefficient is over 100 percent. Q. Do you recall how high the regression coefficient was for the 2012 House election? A. You're asking for the Table 1 results here on the first row? Q. Correct. A. I don't recall the precise number, but obviously, as I explained in my report, the regression estimate would have been over 100 percent, and logically reported as 100 percent. Q. And in Tables 1 and 2 you're looking exclusively at any part Black and non-black; is that right? A. In Tables 1 and 2 when I conduct the EI and the ER estimates, you're asking about those tables? Q. Yes. A. And your question was whether Black VARGAS REPORTING SERVICES, INC

18 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 18 of includes any part Black; is that right? Q. Well, does Black include any part Black? A. Okay. I'll just explain. For the purpose of creating Tables 1 and 2, the data that I used to identify the racial composition of voters came from the voter registration files and the turnout files. Those files are not as detailed as the Census Bureau's breakdown of individuals' racial composition. So we don't know, for example, because voters are not given the opportunity to identify -- to identify as precisely their racial composition. So here the identification of Black versus non-black is anybody who checked the box for African-American as opposed to checking one of the other boxes. Q. And everyone who checked the box other than African-American is in the non-black? A. That's correct, sir. Q. On Page 7 of your report -- A. Yes, sir. Q. -- in the Paragraph No. 2 you say, "Non-African-American voters within House District 105 as drawn under the 2012 Plan became somewhat more likely to favor a Black Democratic House candidate in November 2016." VARGAS REPORTING SERVICES, INC

19 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 19 of A. Yes, sir, I see that sentence. Q. And is that -- Do you draw that conclusion from Table 1? A. If you could just give me a moment to orient myself. (Reviews document.) Yes, sir. I was describing the results that were reported in the table. Q. And Table 2 doesn't really show that, does it? You can't draw the same conclusion about House District 111 from Table 2? A. What specifically are you asking me about Table 2's results? Q. That non-african-american voters within House District 111, as drawn under the 2012 Plan, became somewhat more likely to favor a Black Democratic House candidate in November of 2016 compared to previous elections. A. (No response.) Q. If you substitute 111 for 105 with respect to -- and ask that question with respect to Table 2, it doesn't show that, does it? A. I did not draw that conclusion regarding 111. I just want to be as complete as possible, and if you'll just give me a moment -- VARGAS REPORTING SERVICES, INC

20 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 20 of Q. Sure. A. -- let me just review this section of my report and -- Q. Please take your time. A. -- make sure that I didn't answer your question in my report. (Reviews document.) I just wanted to make sure that I didn't directly say anything about it in the text of my report, and it appears that there's nothing in the text of my report that directly answers your question. I simply reported the numbers that I do for Table 2 in House District 111. Certainly what those numbers show to me is that there is a bit of an increase in non-african-american support for the Democratic candidate between 2014 to 2016, but obviously not such an increase in 2012 compared to So I just wanted to answer that using Table 2 of my report; answer your questions as completely as possible. Q. On Page 7 when you talk about the increase in the African-American population in House District 105 and you illustrate that in Table 3 -- this is your text -- which shows that African-Americans comprise 35.2 of the election day turnout in November 2012, VARGAS REPORTING SERVICES, INC

21 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 21 of percent in November 2014, and 37.0 percent in November 2016, counting only voters who reside within the 2012 Plan's House District 105 boundaries; is that right? A. Yes, sir, I see that. Q. And the African-Americans, they weren't a majority in the district, were they? A. You're asking me if those numbers that you just read out comprise a majority of the district. Q. Correct. A. Well, obviously the numbers that you just read out are under 50 percent. Q. And the same is true for the African-American population in House District 111; is that right? A. It's addressed in that next sentence, and I -- as I stated in the report, the numbers were 36.1 percent -- the comparable numbers were 36.1 percent in 2012, 37.6 percent in 2014, and 40.3 percent in And obviously, again, I confirmed that those numbers are under 50 percent. Q. I'd like to turn your attention to Page 10 of your report, your initial report. A. Yes, sir. Q. In the first paragraph, talking about VARGAS REPORTING SERVICES, INC

22 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 22 of House District 105, the prediction that non-black support for a Black Democratic candidate increased from 21.4 percent in November 2012 to 25.2 percent in November A. Yes, sir I see that. Q. And you state, "This increase is partially attributable to an increase in the Hispanic and Asian shares of the non-black population of the electorate in House District 105 as illustrated in Table 7..." A. Yes, sir, I see that sentence. Q. "...as Hispanic and Asian voters were more likely than non-hispanic White voters to support Black Democratic candidates." It's the last sentence -- A. Yes, sir -- Q. -- in that paragraph. A. -- I see that sentence. I see that sentence you're referring to. Q. Have you done any ecological inference or ecological regression runs to see about the political cohesion of Hispanic voters in House District 105? A. No, sir, I did not use EI or ER to specifically answer that question. Q. Did you try? A. Did I try doing what? Q. Using EI or ER to see about the cohesion VARGAS REPORTING SERVICES, INC

23 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 23 of of Hispanic voters in House District 105. A. No, sir, I did not. Q. What do you rely on, then, to say that they're more likely to vote Democrat than non-hispanic Whites? A. Okay. You're asking me what the basis of that sentence is in the paragraph that we were just reading? Q. Yes, Dr. Chen. A. Sure. Okay. I am purely relying on my general knowledge of political behavior in the South. If you would just give me a moment, let me just make sure I haven't -- I've given you a complete answer. So if you'll just give me a moment. (Reviews document.) Okay. I want to give as complete an answer as I can. So, again, the basis for my answer is my general knowledge -- for my answer for how I came to the statement that I -- that we were just talking about in that final sentence of that paragraph is my knowledge of political behavior in the South generally. What I saw here specifically, and why I reference Table 7 is because, as I stated in that sentence and as I can see here in Table 7, there was a VARGAS REPORTING SERVICES, INC

24 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 24 of noticeable increase in the Asian and Hispanic shares of the non-black portion of the electorate in House District 105 between 2012 to Now, logically I combined that information with what we just discussed a moment ago, which is that I can see in Table 1 that there was a noticeable increase between 2012 to 2016 in the ecological inference estimate of the percentage of non-blacks supporting the Democratic candidate. So I put those two pieces of information together and came to the conclusion that I stated in that final sentence that these Hispanic and Asian voters were more likely than non-hispanic White voters to support Black Democratic candidates. That confirms my general intuition and knowledge about political behavior in the South. Q. Turning back to Table 1, what part of the non-black estimate would be attributable to Hispanic and Asian voters? How do I see that in Table 1? A. Okay. You're asking me in Table 1, what part of the EI estimate for non-black voter support for Democratic candidates would be attributable to -- Q. Hispanics or Asians. A. -- to Asian or Hispanic voters? And my answer is that I, in Table 1, analyzed all non-blacks VARGAS REPORTING SERVICES, INC

25 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 25 of as a group. Q. And that could include White Democrats, couldn't it? A. Oh. You're asking me if non-blacks include White Democratic voters? Q. Yes, sir. A. Obviously they're always -- Anytime you have a group that includes non-blacks, there are going to be some number of White Democratic voters. Q. And the census has a category of Others, doesn't it? A. You're asking me if the census allows a respondent to check a box for Other; is that right? Q. Yes, sir. A. I think that's right. I'll take your word for it. I couldn't just tell you off the top of my head what that Other category specifies, but I'll take your word for it. Q. Well, if you'll turn to Table 7, the last box down there are Others or Unknown, correct, for turnout on Page 14? A. Okay. You're asking me now about Table 7, and you're asking me what the Other or Unknown line represents? Q. Right. VARGAS REPORTING SERVICES, INC

26 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 26 of A. I just want to be clear. This is not -- This is not census data. This is voter registration and voter history turnout data. So what I'm reporting here in Table 7 are the racial identifications of individuals who turned out in these various elections. So this is not census data, but that bottom line indeed represents people who checked Other or did not check a box at all. Q. Well, going back to Table 1, would the Other be included in the non-black? A. Yes, sir. Non-Blacks includes everybody who did not check the box for African-American. Q. And we don't really have any way of knowing how they voted, right? A. Who are you referring to? Q. The Others and the Unknowns. A. Okay. You're asking me if I can -- if I'm able to draw any conclusions about who specifically the Others are voting for, right? Q. Yes, sir. A. Okay. And the answer is that generally I -- let me turn off my phone. Okay. Your question is whether we can say anything, or whether I can say anything about how the Others or Unknowns voted, and my answer is that I VARGAS REPORTING SERVICES, INC

27 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 27 of analyzed them as part of my EI estimates by grouping voters into Black versus non-black, so in fact we can say something about those -- about those individuals. Q. But none of the voting behavior of Hispanics or Asians or Others or Unknown is reflected in an ecological inference or ecological reflection other than on Table 1; is that right? A. They are included in my analysis, and the behavior of those individuals is, in fact, reflected in my analysis in Table 1. Q. But they don't -- none of them has an independent ecological inference or an ecological regression. A. What exactly, sir, do you mean by "independent ecological inference" or -- Q. Well, you didn't -- A. -- "ecological regression"? Q. Well, you told us you didn't do an ecological reference or ecological regression for Hispanic voters in -- run for Hispanic voters in House District 105, correct? A. I included all Hispanic voters in my analysis, so my results do, in fact, account for the behavior of Hispanic voters. Q. But you didn't try to do Hispanics alone, VARGAS REPORTING SERVICES, INC

28 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 28 of correct? A. What do you mean by "try to do Hispanics alone"? I just want to understand your question. Q. If there were enough Hispanics, you could -- and they were properly distributed, could you try to identify the voting behavior of Hispanics as a group with using ecological inference or ecological regression? A. My answer is that my ecological inference and ecological regression analyses do, in fact, account for the behavior of Hispanics. Q. In the non-black portion of Table 1? A. That is correct. Q. And that's the only place it's accounted for in this report? A. That's not the only place where I've looked at Hispanic voters, but I think you're -- you know, I think you're asking in the context of conducting ecological inference estimates here. I think you're trying to ask whether what I reported in Table 1 and Table 2 represents the approach that I used, and it is. That is exactly how I conducted EI and ER analysis. Q. If you'll look at Table 8 of your report. A. Yes, sir. VARGAS REPORTING SERVICES, INC

29 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 29 of Q. The demographics of House District 111 as reflected in the turnout show a pretty small Hispanic and Asian turnout; is that right? A. They show specifically that between 2012 to 2016 within the old boundaries, within the 2012 Plan's boundaries for House District 111, that Hispanics increased from 1.8 percent to 2.4 percent, and that Asians increased from 0.9 percent to 1.3 percent. So I'd say that comparatively those are quite significant increases, but the point of this table is to report exactly numerically what those increases are. Q. But the demographics of Henry County are different from the demographics of Gwinnett County, right? A. I accept that. Q. In Henry County it's more African-American than White, isn't that right, than Gwinnett County? A. I did not specifically study the comparative demographics of Henry and Gwinnett counties as a whole, so I can't answer that question accurately. Q. Well, it's even -- A. I'll take your word for it. Q. It's even reflected in the difference VARGAS REPORTING SERVICES, INC

30 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 30 of between Tables 7 and 8, isn't it? A. Well, what I specifically studied in Table 7 and Table 8 were the demographics of the election turnout in House Districts 105 and 111 specifically, so that is what I can report in these tables. MR. PARK: Let's take a break for a couple of minutes. (Off the record at 10:54 AM) (On the record at 10:59 AM) BY MR. PARK: Q. When you say you rely -- for your conclusion about the behavior of Asian and Hispanic voters in House District 105, you're purely relying on your general knowledge of political behavior in the South, or something like that. Is that what you said? A. I gave a bit of an a longer answer than that, so I'm happy to revisit that answer again, if it would help you out. Would you like me to do that? Q. When you talk about "the South," what do you mean? A. I mean the South of the United States. Q. Romo vs. Detzner was in Florida, right? A. Yes, sir. VARGAS REPORTING SERVICES, INC

31 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 31 of Q. And League of Women Voters vs. Detzner was also in Florida? A. Yes, sir. Q. And Raleigh Wake was in North Carolina? A. Yes, sir. Q. And Corrine Brown was in Florida? A. Yes, sir. Q. The City of Greensboro was in North Carolina? A. Yes, sir, it was. Q. And Rucho was in North Carolina? A. Yes, sir. Q. Have you done any work in Georgia? A. Have I -- You're asking if I've worked on a redistricting case -- Q. Yes, sir. A. -- in Georgia as an expert witness before, and the answer is no, outside of this case. No. Obviously my work as an expert witness is not the extent of my expertise as a political scientist. I was a political scientist long before I became an expert witness in any redistricting cases. Q. Well, what goes to your general knowledge of political behavior of Asians and Hispanics in Georgia? Where do you draw your conclusion -- What do VARGAS REPORTING SERVICES, INC

32 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 32 of you use to draw that conclusion? A. Okay. You're asking me what is the basis of me saying in this sentence that Hispanic and Asian voters were more likely than non-hispanic White voters to support Black Democratic candidates, right? Q. In Georgia. A. Okay. So I think you're asking me the same question that you asked me some time ago, so I'm going to revisit my answer, because I think this is responsive to your question. So the basis for my answer is, as I said, I looked at Table 7 and I looked at Table 1 and I noticed a couple of things. The first thing that I noticed is that in Table 7 I saw that the percentage of the district that's Asian-Hispanic has increased in House District 105 between 2012 to 2016 within the boundaries -- within the boundaries of the 2012 districts. And then I saw the EI numbers that I reported and discussed some time ago in Table 1. I put those two things together, and that confirmed -- that confirmed my general knowledge about voting behavior of various minority groups, which is that Hispanics and Asians don't vote quite the same as non-hispanic Whites and Asian voters, in general. And it confirmed that these demographic changes had, in VARGAS REPORTING SERVICES, INC

33 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 33 of fact, been part of the partisan -- the results reported in Table 1, which were that there has been a slight change in the percentage of non-whites that support Democratic candidates. Q. How many times have you been to Georgia? A. How many times have I been in the state of Georgia? Q. Yes, sir. A. Well, too many to count. I ll tell you that there's no way I could possibly count. I grew up about 1 mile from the border, the northern border of Georgia, just outside of Chattanooga, Tennessee. So Atlanta was the closest metropolitan -- large metropolitan area to where we lived, so certainly my family and I would've gone on trips down to Atlanta, since that's where all the large museums, supermarkets and other cultural amenities were. We certainly would've gone down to Atlanta, I would say, at least once every one or two months when I was growing up. So I grew up for 18 years just outside of Chattanooga, and so it's easily in the hundreds the number of times that I've been into the state of Georgia. Actually, I'm pretty sure it's in the thousands, because I recall that when I was in elementary school our school bus actually had to make VARGAS REPORTING SERVICES, INC

34 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 34 of a detour into Georgia to pick up some students every day. So every day I would have traveled into the state of Georgia. But obviously, as I said, I grew up very close to this state. Q. How many times to Gwinnett County or Lawrenceville? A. How many times have I been in Gwinnett County? Q. Yes. A. Well, I love going to see the Gwinnett Braves. They have a really nice baseball stadium out there, and so I've been out there at least a few times whenever I'm in conferences in the area. So I really couldn't tell you off the top of my head. Q. What about McDonough in Henry County? A. How many times have I been to McDonough or Henry County? Q. Yes. A. I recall going to Henry County when I was -- when I was in high school. One time there was a school trip down there. I really couldn't tell you off the top of my head, but I recall being to Henry County one time. Q. Are you saying that Hispanic and Asian VARGAS REPORTING SERVICES, INC

35 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 35 of voters behave the same way in Georgia that they do in Florida and North Carolina? A. That's not a question I analyzed. Q. When you were doing the work in North Carolina, were you looking -- did you look at the behavior of Hispanic voters? A. I'm just going to have to ask you to be more specific. Which work in North Carolina are you referring to? Q. City of Greensboro or Common Cause vs. Rucho? A. Okay. So let's take those one at a time. You're asking me in the expert work that I did in Greensboro -- Q. Yes, sir. A. -- whether I looked at the voting behavior of Hispanic voters? Q. Yes. A. There certainly are a significant number of Hispanic voters within the city limits of Greensboro, and I had to analyze the partisan behavior of those voters as part of my general analysis of partisan voting patterns in Greensboro. So Hispanic voters were certainly part of my analysis in Greensboro. VARGAS REPORTING SERVICES, INC

36 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 36 of Q. Were they reflected in the non-black portion of an ecological regression chart, or were they reflected in a Hispanic portion of the chart? A. What exactly are you referring to? You're asking about my report here in this case? Q. In Rucho. A. Oh, in Rucho? Q. No. In Greensboro. In Greensboro. A. Okay. You're asking me about Greensboro. And what specifically -- you're asking me -- Q. Looking at Table 1 you ve got Black and non-black. In Greensboro, did you -- were your ecological inference runs based on Black and non-black? A. And my answer is that I did not do an ecological inference run in my expert report in Greensboro. Q. How did you analyze voting behavior, then, in Greensboro? A. How did I analyze voting behavior in my Greensboro expert report? Q. Yes, sir. A. I looked at past election results. Q. But you didn't use ecological inference or ecological regression to make any inferences about the VARGAS REPORTING SERVICES, INC

37 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 37 of voting behavior of various blocks? A. You're asking about my expert report in Greensboro, right? Q. Yes. A. And the answer is no, I did not use ecological inference. Q. What about Rucho? Did you do ecological inference and ecological regression to determine the behavior of blocks of voters? A. I did not do so in my expert report for the Common Cause vs. Rucho case. Q. What about in Romo? Did you do -- was the behavior of Hispanic voters at issue in Romo vs. Detzner? A. I can't tell you if that was an issue that was litigated. I can tell you what's in my expert report, though. Q. What was in -- What was the general nature of your expert report? Did you do ecological inference/ecological regression analysis? A. I did not, sir. My report in Romo -- You're asking about Romo vs. Detzner, right? Q. Yes, sir. A. In that report I did not present an ecological inference analysis. VARGAS REPORTING SERVICES, INC

38 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 38 of Q. When you speak about your general knowledge of political behavior in the South, are you relying on any publications? A. I am relying upon my education and training as a political scientist, which included courses on political behavior in the United States. Certainly in those courses we would have read quite a few publications. Q. Can you remember any of the publications that address the voting behavior of Hispanic voters? A. I couldn't tell you off the top of my head, but I'm certainly confident in generally saying that that is a topic that is covered in political behavior. Q. What about Asian voters? A. Same answer there. I couldn't tell you a publication off the top of my head, but I'm confident in saying that that is a commonly-studied topic in political behavior. Q. When you talk about the Reapportionment Office and its treatment of sub-precinct behavior, as you do on Page 33 of your report and Page 34 of your report -- A. Yes, sir. Q. -- you're relying on your understanding of VARGAS REPORTING SERVICES, INC

39 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 39 of Ms. Wright's deposition; is that correct? A. I am relying on Ms. Wright's deposition, but in this section of the report I'm also drawing on my own expert knowledge of census data, census block-level data, precinct election data, and the relationship between those different geographic units. Q. Well, do you recall in Mr. Strangia's deposition him saying that the software allocated voters into split precincts by the voting age population of the splits? Do you recall whether he testified that way? A. I recall that Mr. Strangia stated in his deposition that his office would have estimated, using voting age population, the partisanship of census blocks within split -- or within any precinct. And specifically what was important for me for my understanding of its process was that Mr. Strangia confirmed Ms. Wright's testimony that the Reapportionment Office had no way to distinguish the relative partisanship of different census blocks within split precincts. It struck me that that is -- that made sense to me, that they had that limitation. And I read that part of both of their depositions, and they seemed to be saying the same thing. And I wanted to review those portions of the VARGAS REPORTING SERVICES, INC

40 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 40 of depositions to understand their process and what data they had available at the block level within split precincts, and that made sense to me. Q. In essence, they used voting age population as the basis for an estimate of who is in each part of the split, right? A. Well, I just want to make sure you understand what I had gleaned from Mr. Strangia's deposition, as well as from Ms. Wright's deposition. It's not that they were estimating the relative partisanship of census blocks within a split precinct. They had no way -- Both of them testified they had no way to distinguish the partisanship, the relative partisan meanings of various census blocks within a split precinct; therefore, they simply assumed, or their software assumed that the partisanship of all census blocks within a split precinct is exactly the same, is perfectly uniform throughout all of the census blocks within a split precinct. That's not really estimating the partisanship of different census blocks within a split precinct. That's simply assuming uniformity. And maybe that's what you're referring to, but I just wanted to make sure that I was clear that that is information that I gleaned from those two VARGAS REPORTING SERVICES, INC

41 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 41 of depositions. Q. Well, do you recall the portion of Mr. Strangia's testimony or deposition where he said the lowest level of geography for election results is the precinct level of geography, okay? So that needs to be allocated down to the block level of geography based on voting age population? A. I recall him discussing that topic. If you'd like to put that deposition transcript in front of me, I would be happy to confirm that he said that, but I don't have that transcript in front of me here. Q. Well, do you know whether -- do you recall whether he said whether other states use a similar formula? A. I recall there was some discussion about them. As I said, I generally recall the discussion. I just can't confirm the precise words that he used because I don't have the same transcript that you have in front of you. Q. Did you recall that Ms. Wright testified that she had political data at the block level also? (Reading) "It's an estimate. When we bring in our file from the Secretary of State's Office, it is completely accurate to the precinct. But when we allocate that data from them to the block level, it VARGAS REPORTING SERVICES, INC

42 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 42 of estimates that figure to the blocks based on the percentage and proportions of the population"? MR. GREENBAUM: Objection. Asked and answered. Go ahead. Answer the question one more time. BY MR. PARK: Q. Do you recall her testifying to that? A. My answer is the same, which is that I can't confirm the precise words because I don't have the same deposition transcript that you have in front of you. I don't have a copy of it here with me right now in front of me, but I obviously remember them discussing these general topics in describing their process. And if you'd like to put the transcript in front of me, I would be happy to help you understand exactly what they're talking about. Q. You would be happy to help me understand what they're talking about? A. I would be happy to explain to you the process that Ms. Wright and Mr. Strangia are describing, if you'd like to put the transcript in front of me. MR. GREENBAUM: And I think, in fact, you VARGAS REPORTING SERVICES, INC

43 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 43 of already have described the process. BY MR. PARK: 3 Q You talk about some of the precinct changes in your report with respect to House District 105, and one change that was made in 2015 was Harbins C was added to House District 105; is that right? A. Yes, sir. Q. Is Harbins C a rural area? A. If you'll just give me a moment and let me review my report. (Reviews document.) Q. If you'll look at Page 24 of your report. A. Yes, sir. Q. In the bottom paragraph you say that House District 105, a precinct Harbins C, was added to it? A. Yes, sir. Q. Do you know whether Harbins C is rural or urban? A. I m just going to have to ask you to define what you mean by those terms. Q. Do you know whether it's Republican? A. Do I know whether Harbins C is a Republican precinct? Q. Mainly Republican precinct. A. Let me see if that's reported in my report. So if you'll just give me a second to look up VARGAS REPORTING SERVICES, INC

44 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 44 of to orient myself in my report. (Reviews document.) I can't find the place in my report where I specifically reported the partisan composition of that particular precinct, but I'll take your word for it. Q. On Page 25 you talk about moving a portion of Lawrenceville M out of 105 into 104? A. Yes, sir. Q. And that portion contains a BVAP of 45.6 percent? A. Yes, sir, I see that sentence. Q. Does that 45.6 percent BVAP mean that the non-bvap population would be some 54.4 percent? A. You're asking me if -- Q. Isn't that a majority non-bvap move? A. Okay. You're asking me of this portion of the Lawrenceville M precinct is more than 50 percent non-african-american. Q. Yes, sir. A. And I see -- I think I understand the math you're asking me, and I confirm that, sure, it's a little bit over 50 percent non-african-american. I think what I refer to in this portion of my report is that clearly this portion of the Lawrenceville M precinct is more heavily African-American than the VARGAS REPORTING SERVICES, INC

45 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 45 of rest of the district. I think that's all I was referring to here. But I confirm your math. Q. If you turn to Table A. Yes, sir. Q. -- this shows that by the 2010 Census, voting age population, the Black population, proportion of the population inside House District 105 boundaries was 29.8 percent; is that right? A. If I could just ask you to repeat that question before. I want to make sure I follow. Q. All right. We're looking at 2010 Census VAP by race within and outside House District 105 for the 2015 Plan in split precincts; is that right? A. Yes, sir. You're looking at the bottom table. Q. And in the totals by race for 2016 election turnout within, it's 29.8 percent; is that right? A. You're asking about the number -- Q. The total by race -- A. -- that was reported -- Q. -- for the Black proportion. A. Sure. You're asking about the number that I reported in the bottom row where I'm looking at -- where I'm looking at the House District 105 boundaries VARGAS REPORTING SERVICES, INC

46 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 46 of in the 2015 Plan, and I confirm that it is, in fact, 29.8 percent that I reported as a Black proportion. Q. And outside it was 30.4 percent Black proportion? A. Now you're asking about the right column, and I looked at the portions of the House District 105 split precincts outside of the district boundaries themselves, outside of the House District 105 boundaries, and I reported that the Black proportion was 30.4 percent. So, yes, I reported that number. Q. Do you know why there was a lower turnout within House District 105 boundaries than without in 2016? A. Do I know why there was a lower turnout within than without? I'm not sure that I've actually reported that there was a lower turnout. Q. A lower percentage of turnout? A. I -- Q. Why would the minority population turn out at a slightly lower rate within? A. Okay. You had asked me a moment ago whether -- why there was a lower turnout, and I'm not sure that I reported there was, in fact, a lower turnout. I think what you're trying to ask is specifically about the Black proportion of the VARGAS REPORTING SERVICES, INC

47 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 47 of electorate, right? Q. Correct. A. Okay. I appreciate that. So you're asking me why the total VAP within the House District 105 boundaries in these split precincts is 29.8 percent, and why the portion outside is 30.4 percent. Have I got your question right? Q. Correct. A. Okay. Sure. What I did in calculating those numbers was I aggregated up across these three split precincts, the three split portions that were within House District 105 boundaries, and then I aggregated up in the right column the House District 105 portions of these three split precincts outside. Now, the aggregate racial composition of those three precincts put together means absolutely nothing. It's absolutely irrelevant for the purposes of analyzing how these three individual precincts were split. And when you take these three split precincts and aggregate up the racial composition, that tells you a number that is pretty meaningless. It's telling you something about an arbitrary three chosen precincts, but you're not even analyzing one at a time. You're just reporting something about the VARGAS REPORTING SERVICES, INC

48 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 48 of total racial composition of these three chosen precincts, the ones that were chosen to be split. So it's a pretty arbitrary choice of three precincts, and that is not the basis on which I analyze split precincts in my report. Q. With respect to House District 111, do you know where the then incumbent Strickland lived? A. I couldn't tell you off the top of my head. Q. One of the precincts -- On Page 361, one of the split precincts is Tussahaw. A. Yes, sir, I see that. Q. Do you know whether Tussahaw performs pretty strongly Republican? A. I couldn't tell you off the top of my head. Q. Would that be a reason to put Tussahaw in the district of a Republican, or that portion of the Tussahaw precinct into the district of a Republican? A. You're asking me if Tussahaw's strong Republican performance would be a reason to put that split portion of Tussahaw into House District 111? That's your question? Q. Correct. A. Okay. And my answer is that I accept your VARGAS REPORTING SERVICES, INC

49 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 49 of representation that there may be a portion of Tussahaw that is strongly performing for Republicans. I accept that because you've told me that. I think that's what you're telling -- you're asking me to accept. My answer to your question, though, is that from my reading of Ms. Wright's deposition, Mr. Strangia's deposition, it would be impossible for the Legislative Reapportionment Office to know the relative partisan performance of one split precinct of the Tussahaw precinct as opposed to another split portion of that same precinct, because they both testified that they lacked the granular data necessary to compare the relative partisan performance of two split portions of the same precinct. I also know from their depositions that the only data they would've had to distinguish between the split portions of a precinct like Tussahaw would have been census demographic data, including racial data. So from the perspective of the Legislative Reapportionment Office, it would not have been possible to do what I think you're suggesting or asking about. Q. Just completely impossible? Is that your testimony? VARGAS REPORTING SERVICES, INC

50 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 50 of A. It's my testimony, given my understanding of the data available to the Reapportionment Office. Now, I wouldn't go so far as to say it's completely impossible, because what I found in my report actually does make it possible to do a version of what you're suggesting. My report concluded that it is, in fact, possible in House District 111 to use racial data to effectively get at the kind of partisan consideration that I think you are asking about. So it certainly would have been possible to use the available racial data that Ms. Wright and Mr. Strangia had to effectively accomplish the partisan manipulation of a splitting of the precinct that I think you are suggesting in your question. Q. Do you know whether there are White Democrats in Henry County? A. Do I know whether there are any White Democrats at all? Q. Yes, sir. A. In Henry County? I assume there are. Q. Do you know whether there's a significant block of White Democrats in Henry County? A. Well, I analyzed that question specifically with respect to 111. And obviously, as VARGAS REPORTING SERVICES, INC

51 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 51 of I've discussed earlier, those specific numbers are reported in Table 2 of my report. So I was able to answer that question only in the context of House District 111, but I was able to answer it a bit more precisely. Q. Let's take a look at your reply report. Let's mark this as Defendant's 2. Is that a copy of your reply report? A. Yes, it is. Q. Does it appear to be a true and correct copy? A. Yes, sir, it does. (Exhibit No. 02 was marked/identified.) BY MR. PARK: Q. If you would turn to Page 15. A. Yes, sir. Q. In the paragraph that starts "moreover," you say, "The key issue I sought to analyze was whether Black voters generally support Democratic or Republican candidates." A. I just want to try and find where you are on the page here. Q. The fourth line or the fifth line in the middle, "The key issue"? VARGAS REPORTING SERVICES, INC

52 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 52 of A. "The key issue I sought to analyze." Yes, sir, I see where you are. Go ahead. Q. And you go on to say that the Democratic candidate for November 2014 House election was Caucasian, and that House election exhibited substantially the same level of racially-polarized voting as the other two HD 111 House elections I analyzed in my original expert report." A. Yes, sir, I see where I said that. Q. And so do I correctly understand that Black voters voted for that White Democratic candidate in November 2014 in the House District 111 election? Is that what you're saying? A. If I could ask you to repeat the question. Q. Are you saying that Black voters in House District 111 in the November 2014 election supported the White Democratic candidate? A. Okay. You're basically asking me what the race of the Democratic candidate was and whether or not Black supporters supported that candidate at a relatively-high rate, and the answer is yes. I, of course, said here that the Democratic candidate is -- Q. You said he was Caucasian, right? A. Yes, sir. I agree with you. The candidate is -- the candidate in 2014 was Caucasian, VARGAS REPORTING SERVICES, INC

53 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 53 of and I presented EI numbers for Black support of the Democratic candidate that were in line with what I had reported for 2012 and Q. In his report Dr. Alford suggested that Black voters generally supported Democratic candidates in other races, and I know you disagree with his analysis, but is that -- do you think that's correct, that Black voters support Democratic candidates in, say, U.S. Senate elections, without regard to whether the candidate is White or Black? A. Well, I analyzed the question with respect to the House district elections, because those were the most probative elections. And as we've just been talking about, clearly that includes one election in 2014 where Blacks supported a Caucasian Democratic candidate. That's the extent of my analysis. Q. Is that the extent of your -- That's the extent of your analysis? Are you aware of other consistent instances of that happening? A. Okay. You're asking more generally -- Q. Yes, sir. A. -- beyond my expert report here? Q. Yes, sir. A. Sure. I'm happy to confirm that indeed there are examples where Black voters have voted in VARGAS REPORTING SERVICES, INC

54 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 54 of favor of a White Democratic candidate. I confirm that as a general -- as a general matter. Q. Do you think that would include Jim Barksdale when he ran for U.S. Senate -- A. I did not -- Q. -- here in Georgia? A. I did not analyze the results of that election beyond looking at the data and analysis in Dr. Alford's reports. But it's not one that in my report, for my expert report, that I analyzed. Q. What about Stephen Oppenheimer in 2012 for Public Service Commission District 3? A. I have the same answer there. I -- Q. That it's not in your report? A. Sure. I have the same answer. I also don't have Dr. Alford's report in front of me right now, so I can't tell you the race of those various candidates. But I take your word for it that that was a White Democrat, and I'm just giving you the same answer that I did not analyze it in my expert report. Q. In his rebuttal report, again on Page 15, Dr. Alford talks about the portions of the split precincts inside -- MR. GREENBAUM: Are we talking about Dr. Alford's report now, or Dr. Chen's report? VARGAS REPORTING SERVICES, INC

55 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 55 of MR. PARK: Dr. Chen's report. MR. GREENBAUM: The rebuttal report? MR. PARK: Yes, sir. MR. GREENBAUM: Exhibit 2? MR. PARK: Yes, sir. MR. GREENBAUM: Okay. Thank you. Sorry about the lack of clarity. BY MR. PARK: Q. You say, "Dr. Alford's observation about the aggregate racial composition of these split precinct portions is irrelevant." A. Yes, sir. Q. Why is it irrelevant? A. Why is it relevant for my -- Q. Why is it irrelevant? A. Sure. Why is it irrelevant for my analysis? And I'm happy to explain that in some detail. I will start my answer by saying that I explained why it's irrelevant in the latter half of Page 15, Page 16, and into Page 17, but I'll give you a briefer version here. It's relevant for a number of different reasons. When we're understanding split precincts precisely in the context of this particular case, VARGAS REPORTING SERVICES, INC

56 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 56 of there are -- were a number of important issues that I noted. First, because I noted that there was strong racially-polarized voting, it was clear that even just a small change in the racial composition of a district could have significant political consequences. Now, specifically with respect to split precincts, this means that it's important to analyze these split precincts one at a time. The broader reason why those split precincts need to be analyzed one at a time rather than in the aggregate is because each individual decision to split a precinct is one that I sought to scrutinize. I sought to scrutinize each individual decision to split an individual precinct because there was nothing that compelled, there was nothing that forced the precincts to be split by the Legislature. It wasn't something that the reapportionment guidelines require, the splitting of precincts. It was something that the Legislature chose to do. And so it was specifically a deviation from the redistricting guidelines whenever a precinct was split, and so I analyzed those, one at a time. So the point is that to properly analyze them, it's necessary to go precinct by precinct rather than to aggregate together the arbitrary set of VARGAS REPORTING SERVICES, INC

57 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 57 of precincts that were chosen to be split by the Reapportionment Office. Instead, it was necessary to go individual precinct by individual precinct and analyze the racial composition of those individual split precincts. Another important consideration for me in determining that the relevant consideration was how each individual precinct was split rather than how the aggregate composition of those split precincts looked was simply that I, based on Ms. Wright's and Mr. Strangia's deposition testimony, knew that the only available data within a particular split precinct available to them was census demographic and racial data and not partisan data. So it was important for me for all of those reasons to analyze how each individual precinct was split rather than the aggregate composition of all of those split precincts. So those are the various reasons why I said that focusing on the aggregate of those split precincts was irrelevant and beside the point of how to properly analyze split precincts. And I laid those explanations, those reasons out in detail in my response report. Q. When we determine the winner of an VARGAS REPORTING SERVICES, INC

58 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 58 of election in House District 105 and House District 111, we aggregate the votes, don't we? A. I agree with you that that is what the Secretary of State does when determining the winner of an election contest after the district lines have already been drawn and are put into place. Q. And a vote from a split precinct counts just as much as a vote from a whole precinct, doesn't it? A. That is indeed how the Secretary of State would put together election results. I affirm your description of how votes are counted. Q. On Page 17 of your reply report you say that the aggregate racial composition is more heavily African-American inside because of -- than outside is entirely caused by two precincts, Lawrenceville D in 105 and Hickory Flat in 111. Is that -- Am I reading that right? A. I see where you're referring to. You're referring to that second full paragraph on Page 17. I do see where you're referring to. Q. And you talk about racial segregation in the next sentence. Is that de jure segregation? A. Is the racial segregation of these two precincts de jure segregation? VARGAS REPORTING SERVICES, INC

59 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 59 of Q. Yes, sir. A. I really couldn't tell you, because I did not conduct an analysis of the causes of racial segregation in precinct Lawrenceville D and precinct Hickory Flat. Q. But you say it's racially segregated, right? Why do you say it's racially segregated? MR. GREENBAUM: Objection. Argumentative. THE WITNESS: I am merely describing the racial geography of those precincts that I saw using census data, using census data and using the election -- or the voter registration data that I analyzed in my reports. I made -- I was not attempting at all, and I apologize if this portion of the report misled you, to say anything further than an empirical statement of the geographic patterns. BY MR. PARK: Q. Of residency, right? A. The geography of race in these two precincts. That's all I was talking about. Q. Where people live, right, and who lives there? A. Sure. What we're talking about with geography is data that measures where people live. VARGAS REPORTING SERVICES, INC

60 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 60 of Q. And the drafter put these two precincts, or these two more heavily African-American portions of precincts into 105 and 111, right? Mrs. Wright did that? A. Maybe we can take those precincts one at the time. Q. Okay. A. I just want to make sure I understand your question properly. Q. Somehow Lawrenceville D ended up in -- or a portion of Lawrenceville D ended up in House District 105, right? A. Indeed. It was a split precinct. Q. And Ms. Wright put it in Lawrenceville D, didn't she? A. I assume the Reapportionment Office did. I couldn't -- I couldn't speak for her, but I accept -- I accept what you're saying. Q. And the Reapportionment Office would have put Hickory Flat in 111, right? A. I agree that it was a split precinct. Q. On Page 13 of your reply report you, in Table 1 -- and you talk about Democratic candidates' share of the two-party vote in House District 111 in January of VARGAS REPORTING SERVICES, INC

61 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 61 of A. Yes, sir, I see that. Q. Was that a -- Was that a jungle primary? A. If I could ask you to define that term. Q. How many candidates were involved in that election? Do you recall? A. I believe there were four. Q. Two Democrats and two Republicans; is that right? A. I believe that's correct. Q. And, again, on the ecological regression for Black in January of 2018, you got a result that was greater than 100 percent and rounded it down; is that right? A. That's correct. You're talking about the right portion of Table 1. Same answer as when we discussed that issue earlier in my original report. Q. How do you get results that are over 100 percent? Do you know how that happens? A. You're asking why does ecological regression produce estimates for Black voting behavior that are over 100 percent, right? Q. For any behavior -- A. Any behavior. Q. -- but in this case it's consistent for Black voting behavior. VARGAS REPORTING SERVICES, INC

62 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 62 of A. Sure. So you're asking how an estimate over 100 percent occurs. And the answer, just to explain a little bit of the statistical details behind this, is that when you conducted an ecological regression estimate, you were fitting a regression line to the data. So it's estimating one parameter for Blacks and one parameter for non-blacks that will be applied to all precincts within a particular district, and in this case, 111. And when you conduct a regression estimate, a linear regression estimate, what you're doing is fitting all of the data to one line, to two parameters, one regarding Blacks and one regarding non-blacks. And what the regression analysis does by fitting that line is producing simply a straight line, a straight line between the racial composition of a precinct and its partisan voting patterns. So when you produce that straight line, sometimes regression analysis will result in estimates at the very extremes that are obviously outside the bounds of what is logically possible. In other words, outside of 100 percent and zero percent. That is something that does not happen with ecological inference, but it does happen with ecological regression. VARGAS REPORTING SERVICES, INC

63 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 63 of And so the convention has been that when you produce an estimate using regression analysis that results in, say, a logically-impossible conclusion, say, a particular racial group voting at over 100 percent or under zero percent for a particular candidate, then we simply round that number. We report that number as zero or 100 percent. So that's -- that's how it happens. That's just a bit of the statistical background behind how ecological regression works. Q. You criticize the plan drafters for splitting municipalities; is that right? A. I don't criticize the plan drawers for anything. I simply analyzed what the plan drawers did. Again, I think what you're asking -- Q. You note -- A. -- your question -- Q. You note -- A. -- is whether I analyzed -- Q. You note -- A. -- split municipalities. Q. You note that the drafters split municipalities, correct? A. That is a part of what I analyzed in my report, yes, sir. VARGAS REPORTING SERVICES, INC

64 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 64 of Q. Did the guidelines, to the best of your recollection, say anything about municipalities? A. I think I've listed all of the guidelines stated in my original report, so if you'll allow me a second just to review that section of my report. Q. Sure. A. (Reviews document.) The answer is that the redistricting guidelines are silent on the issue of municipalities and only talk about avoiding a splitting of precincts and counties. Q. Correct. And do you know what the General Assembly used as their allowable deviation in drawing these plans, allowable population deviation? A. I'm not sure I'm aware of that number. Q. Do you know whether they went plus/minus one percent? A. I'm not sure I know. Q. Do you think you can split more counties if you went plus/minus one than if you went plus/minus five? A. Would you split more counties in a district -- Q. Are you most likely to? A. Are you more likely to split more counties in a plus or minus one percent deviation world than if VARGAS REPORTING SERVICES, INC

65 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 65 of you allowed plus or minus five percent? Q. Correct. A. And my answer is that, in general, that doesn't make a difference because in the end typically a plan will, particularly with larger counties that have to be split into multiple districts, which obviously applies to 105 and 111, a districting plan in general only needs to split -- split a county only when necessary to achieve equal population, which specifically means that at most two districts might need to be involved in the splitting of -- in being split up. In other words, it's possible to comply with any reasonable equal population threshold in larger counties, such as Gwinnett and Henry, by only involving two districts and splitting up that -- and being split -- I m sorry -- in only involving two districts in being split across counties. So my answer is that, in general, it shouldn't matter. Q. If you'd look at Page 14 of your reply report. A. Yes, sir. Q. In Paragraph 4 in the first paragraph you say toward the end of the fifth line from the bottom, VARGAS REPORTING SERVICES, INC

66 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 66 of "In my original report I found that in both House District 105 and House District 111, 98 to 99 percent Black voters support the Democratic state legislative candidate in each election." A. Yes, sir, I see that. Q. And, "Meanwhile 75 to 85 percent of non-black support Republican candidates" -- A. Yes, sir, I see that. Q. -- "constituting a level of Black voting sufficient to defeat the Democratic candidate supported by Black voters"? A. Yes, sir. Q. And, again, in House District 105 and 111, the Black voters are a minority when compared to the non-black voters, right? A. I can affirm that indeed it is less than 50 percent. Q. So if it's less than 50 percent, then everyone who could vote voted. Is that still White-Black vote -- or non-black voting defeating them or is it just their minority status? A. I want to make sure I understand the question. I think there are a couple of things going on here. You're asking me about a hypothetical world in which every Black person shows up and votes? VARGAS REPORTING SERVICES, INC

67 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 67 of Q. If every Black showed up to vote in House District 105 and voted for the Democrat, you still need crossover from some of the non-blacks to win, right, because you're the minority? A. And you're -- again, you're asking me about a hypothetical world in which every Black person votes. And are you telling me anything about how the non-black voters turn out? I just want understand your scenario as accurately as I can here. Q. Well, if you take every Black voter who votes in House District 105, you have -- I'll withdraw the question? MR. PARK: Let's take a break. (Off the record at 12:05 PM) (On the record at 12:07 PM) MR. PARK: Dr. Chen, I believe that's all the questions I have at this time. MR. GREENBAUM: And I only have a couple of questions. EXAMINATION BY MR. GREENBAUM: Q. Dr. Chen, during Mr. Park's examination, you said that for each election you run King's code once, correct? VARGAS REPORTING SERVICES, INC

68 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 68 of A. Yes, sir, that is exactly what I do. Q. What is entailed in -- When you run King's code once, what does that mean? A. Well, I call the software, and the software, with its standard -- with its standard inputs and its standard parameters, are run once. Now, the software certainly does a substantial number of what are called "burn-ins" as part of how the software normally operates, but I physically only run the software once. Q. Just really quickly, what is a burn-in? A. A burn-in involves runs that, at the beginning of the algorithm, of the EI algorithm, that are not kept and not analyzed. And typically the software will do that by default. And certainly the EI package that I was describing earlier to Mr. Park does that. So a burn-in is simply a large number of iterations at the beginning, certainly at least several thousand or so, probably more, as part of any EI software, as part of any EI algorithm, that are conducted at the beginning and not kept, not used in the analysis and not reported. That doesn't mean that I am intentionally throwing away any results. I obviously explained to Mr. Park this morning that I reported all of my VARGAS REPORTING SERVICES, INC

69 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 69 of analyses. But the EI software itself, the EI algorithm, includes that as part of -- includes those burn-ins as a standard part of how the algorithm runs. Q. So each run contains -- Each time you do a run, you run the code, the software goes through several thousand iterations? A. Oh, absolutely. I couldn't tell you the exact number off the top of my head, but that is a standard parameter in any EI software that one would use in this sort of racially-polarized voting analysis. It's a given that that is how ecological inference is conducted. MR. GREENBAUM: No further questions. EXAMINATION BY MR. PARK: Q. Just so I'm clear, what is the trade name of the software? A. Sure. There's not a trade name. It's not a trademarked software. It's open source software, and it's by Professor Gary King. The software is just called EI. It's runn in the R programming language. It's one that's open source. It's freely available to anybody on the Internet. And it's -- obviously it's the open-source computer code that Professor King and various collaborators developed as part of his VARGAS REPORTING SERVICES, INC

70 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 70 of development and research on ecological inference back in the early 2000's. But I would just add to that that in the computer code that I turned over in connection with my original expert report, the layout, the installation of that package, of that computer code, as well as the use -- my use of that EI software, is all in the computer code, and it was laid out precisely how I uploaded the computer code, as well as how I used it and the standard parameters that were used in my EI analysis. So all of that is clearly laid out in the computer code that I turned over in connection with my expert reports. Q. And, again, you call the package EI? A. Yeah. It's just called EI. The letters E-I. I know it's not the best name and not the most creative name, but that is -- that's literally what it's called. It just stands for Ecological Inference. MR. PARK: Thank you, Dr. Chen. THE WITNESS: Thank you, sir. MR. GREENBAUM: Aria, any questions? MS. BRANCH: No. No, thank you. MR. GREENBAUM: All right. We're done. THE COURT REPORTER: Aria, would you like a copy of the deposition? VARGAS REPORTING SERVICES, INC

71 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 71 of MS. BRANCH: Yes, we would like a copy. THE COURT REPORTER: All right. And Jon, are you reserving signature? MR. GREENBAUM: Yes. He'll read and sign. (Time noted: 12:12 p.m.) VARGAS REPORTING SERVICES, INC

72 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 72 of DISCLOSURE STATE OF GEORGIA COUNTY OF FULTON WITNESS: JOWEI CHEN, Ph.D Pursuant to Article 10.B of the Rules and Regulations of the Board of Court Reporting of the Judicial Council of Georgia, I make the following disclosure: I am a Georgia Certified Court Reporter. I am not disqualified for a relationship of interest under the provisions of O.C.G.A (c). I am a representative of Vargas Reporting Services, Inc. Vargas Reporting Services, Inc. was contacted by the offices of STRICKLAND, BROCKINGTON & LEWIS, LLP to provide court reporting services for this proceeding. Vargas Reporting Services, Inc. will not be taking this proceeding under any contract that is prohibited by Georgia law Marianne Vargas, CCR, CVR-M Certified Court Reporter Certificate Number B VARGAS REPORTING SERVICES, INC

73 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 73 of C E R T I F I C A T E STATE OF GEORGIA: COUNTY OF FULTON: I hereby certify that the foregoing proceeding was taken down as stated in the caption, and the colloquies and questions and answers were reduced to typewriting under my direction; that the foregoing transcript is a true and correct record of the evidence given upon said hearing. I further certify that I am not of kin or counsel to the parties in the case, and am not in the regular employ of counsel for any of the said parties, nor am I in any way interested in the outcome of the case. This certification is expressly withdrawn and denied upon the alteration, disassembly, and/or photocopying of the foregoing proceedings, including exhibits, unless such is done by the undersigned certified court reporter and the signature and original seal is attached thereto. This day, March 4, Marianne Vargas, CCR, CVR-M Certified Court Reporter Certificate Number B-1832 VARGAS REPORTING SERVICES, INC

74 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 74 of ERRATA SHEET I, JOWEI CHEN, Ph.D, the witness herein, do hereby certify that I have read the transcript of February 27, 2018, of my deposition testimony, and the same is true and correct, to the best of my knowledge, with the exception of the following changes noted below, if any: Page Line should read: Reason for change: Page Line should read: Reason for change: Page Line should read: Reason for change: Page Line should read: Reason for change: Page Line should read: Reason for change: Page Line should read: Reason for change: Page Line should read: Reason for change: JOWEI CHEN, Ph.D Sworn to and subscribed before me, the undersigned Notary Public, on this day of,. Notary Public VARGAS REPORTING SERVICES, INC

75 BY MR. GREENBAUM: [1] 67/21 BY MR. PARK: [11] 4/11 7/19 11/21 15/17 30/11 42/6 43/1 51/14 55/7 59/17 69/14 MR. GREENBAUM: [12] 42/2 42/24 54/23 55/1 55/3 55/5 59/7 67/18 69/12 70/20 70/22 71/3 MR. PARK: [11] 4/2 11/18 15/13 15/16 30/6 54/25 55/2 55/4 67/12 67/16 70/18 MS. BRANCH: [2] 70/21 70/25 THE COURT REPORTER: [2] 70/23 71/1 THE WITNESS: [2] 59/8 70/19....as [1] 22/ [1] 29/8 01 [1] 7/ [1] 1/4 02 [1] 51/ percent [1] 29/9 1.8 [1] 29/7 10 [1] 21/22 10.B [1] 72/5 100 [1] 62/ percent [15] 16/21 16/21 16/25 17/2 17/3 17/4 17/8 17/16 17/17 61/12 61/18 61/21 62/2 63/5 63/7 104 [1] 44/7 105 [36] 16/16 18/22 19/19 20/23 21/3 22/1 22/9 22/20 23/1 24/3 27/21 30/4 30/15 32/16 43/5 43/6 43/14 44/7 45/7 45/12 45/25 46/6 46/8 46/12 47/5 47/12 47/14 58/1 58/17 60/3 60/12 65/7 66/2 66/13 67/2 67/11 10:00 [1] 1/18 10:03 [1] 4/2 10:54 [1] 30/9 10:59 [1] 30/ [26] 16/16 19/10 19/14 19/19 19/23 20/12 21/14 29/1 29/6 30/4 48/6 48/22 50/8 50/25 51/4 52/7 52/12 52/16 58/1 58/17 60/3 60/20 60/24 62/9 65/7 66/13 111, 98 to [1] 66/ [1] 2/ [1] 1/20 12:05 [1] 67/14 12:07 [1] 67/16 12:12 [1] 71/5 13 [1] 60/22 13th [1] 2/14 14 [2] 25/21 65/ [1] 2/5 14th [1] 1/20 15 [4] 45/3 51/16 54/21 55/21 16 [1] 55/21 17 [3] 55/21 58/13 58/20 18 [1] 33/ [2] 72/22 73/ [1] 12/24 1:17-CV [1] 1/4 2 2's [1] 19/ [1] 29/7 2000's [1] 70/ [1] 2/ [1] 2/ [2] 45/5 45/ [15] 13/25 17/10 18/23 19/14 20/17 20/25 21/3 21/19 22/3 24/3 24/7 29/6 32/17 53/3 54/ to 2016 [2] 29/5 32/ [8] 20/16 21/1 21/19 52/4 52/12 52/16 52/25 53/ [5] 15/3 15/21 43/5 45/13 46/ [14] 18/25 19/16 20/16 20/17 21/2 21/19 22/4 24/3 24/7 29/5 32/16 45/16 46/13 53/ [5] 3/11 7/17 10/14 10/21 10/ [6] 1/17 7/7 60/25 61/11 73/20 74/ [1] 2/ [1] 2/ percent [1] 22/ [1] 2/6 22 [1] 3/ [1] 2/20 22nd [1] 7/16 24 [1] 43/11 25 [1] 44/ percent [1] 22/3 27 [2] 1/17 74/3 28 [1] 72/ percent [4] 45/8 45/17 46/2 47/ percent [3] 46/3 46/10 47/ [1] 1/ [1] 1/ [1] 2/21 33 [1] 38/22 34 [1] 38/ [1] 1/ [1] 20/ percent [1] 21/ [1] 21/ percent [1] 21/ [1] 48/ percent [1] 21/ [1] 21/ [1] 2/ percent [1] 21/ [1] 2/ percent [2] 44/10 44/ [1] 44/22 50 percent [5] 21/12 21/21 44/17 66/17 66/ [1] 44/ [1] 1/ [1] 2/ [1] 2/ [1] 1/ [1] 2/ [1] 2/21 75 [1] 66/ [1] 66/ [1] 72/ percent [1] 14/2 99 [1] 66/2 A a.m [1] 1/18 able [3] 26/18 51/2 51/4 about [73] 5/17 5/18 8/14 8/17 8/20 8/23 9/10 9/19 10/10 10/16 11/14 11/24 12/14 15/10 17/22 19/9 19/11 20/8 20/21 21/25 22/19 22/25 23/20 24/15 25/22 26/18 26/24 27/3 27/3 30/14 30/21 32/21 33/11 34/16 36/5 36/9 36/25 37/2 37/7 37/12 37/22 38/1 38/15 38/20 41/15 42/18 42/20 43/3 44/6 45/19 45/23 46/5 46/25 47/23 47/25 49/23 50/10 53/14 54/11 Alford [3] 5/13 53/4 54/22 54/24 55/7 55/9 54/22 58/22 59/21 59/24 Alford's [6] 7/1 7/3 60/23 61/14 64/2 64/9 54/9 54/16 54/25 55/9 66/24 67/6 67/7 algorithm [5] 68/13 abranch [1] 2/16 68/13 68/20 69/2 69/3 absolutely [3] 47/16 all [26] 6/1 6/10 12/8 47/17 69/7 13/11 24/25 26/8 accept [7] 12/11 27/22 33/16 40/16 29/16 48/25 49/2 49/4 40/18 45/1 45/11 60/17 60/18 50/19 57/15 57/17 access [3] 7/11 15/3 59/14 59/21 62/8 15/22 62/12 64/3 67/17 accommodate [1] 68/25 70/7 70/11 5/5 70/23 71/2 accomplish [1] allocate [1] 41/25 50/13 allocated [2] 39/8 account [3] 15/9 41/6 27/23 28/11 allow [1] 64/4 accounted [1] 28/14 allowable [2] 64/12 accurate [1] 41/24 64/13 accurately [2] 29/22 allowed [1] 65/1 67/9 allows [1] 25/12 achieve [1] 65/9 alone [2] 27/25 28/3 across [2] 47/10 already [2] 43/1 58/6 65/18 also [9] 5/24 5/24 6/4 actually [4] 33/23 7/5 31/2 39/3 41/21 33/25 46/15 50/4 49/15 54/15 add [3] 6/1 7/4 70/3 alteration [1] 73/15 added [2] 43/6 43/14 always [1] 25/7 additional [3] 6/13 am [15] 4/2 23/10 6/15 6/17 30/9 30/11 38/4 39/2 address [2] 16/14 58/17 59/9 68/23 72/7 38/10 72/8 72/10 73/10 addressed [1] 21/16 73/11 73/12 Advancement [1] 8/1 amenities [1] 33/17 affirm [2] 58/11 American [14] 18/14 66/16 18/17 18/22 19/13 African [16] 18/14 20/15 20/22 21/14 18/17 18/22 19/13 26/12 29/17 44/18 20/15 20/22 20/24 44/22 44/25 58/15 21/6 21/14 26/12 60/2 29/17 44/18 44/22 Americans [2] 20/24 44/25 58/15 60/2 21/6 African-American [9] analyses [2] 28/10 18/14 18/17 20/22 69/1 21/14 26/12 29/17 analysis [20] 27/8 44/25 58/15 60/2 27/10 27/23 28/23 African-Americans 35/22 35/24 37/20 [2] 20/24 21/6 37/25 53/7 53/16 after [1] 58/5 53/18 54/8 55/17 59/3 again [13] 6/25 9/13 62/14 62/19 63/2 12/9 14/10 21/20 68/22 69/11 70/11 23/17 30/19 54/21 analyze [14] 13/21 61/10 63/15 66/13 35/21 36/18 36/20 67/5 70/14 48/5 51/19 52/1 54/7 age [5] 39/9 39/14 54/20 56/7 56/23 57/4 40/4 41/7 45/6 57/16 57/22 aggregate [10] 47/15 analyzed [15] 7/8 47/21 55/10 56/10 24/25 27/1 35/3 50/24 56/25 57/9 57/17 52/8 53/11 54/10 56/9 57/20 58/2 58/14 56/22 59/13 63/14 aggregated [2] 47/10 63/19 63/24 68/14 47/13 analyzing [2] 47/18 ago [5] 14/11 24/5 47/24 32/8 32/19 46/21 and/or [1] 73/15 agree [3] 52/24 58/3 ANDERSON [1] 2/11 60/21 ANDREA [1] 2/11 ahead [2] 42/5 52/2 another [3] 6/5 49/10 al [3] 1/3 1/10 8/14 57/6 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 75 of 166

76 A ARREY [1] 2/11 AUSTIN [2] 1/10 2/9 being [4] 34/23 65/11 boxes [1] 18/15 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 76 of 166 answer [37] 20/5 20/18 20/19 22/22 23/14 23/17 23/17 23/18 24/25 26/21 26/25 28/9 29/21 30/18 30/19 31/18 32/9 32/11 36/15 37/5 38/16 42/5 42/9 48/25 49/5 51/3 51/4 52/21 54/13 54/15 54/20 55/19 61/15 62/2 64/7 65/3 65/19 answered [2] 11/24 42/4 answers [2] 20/10 73/6 any [30] 5/4 6/15 14/16 14/19 14/22 17/19 18/1 18/2 22/18 26/13 26/18 31/13 31/22 36/25 38/3 38/9 39/15 50/18 61/22 61/23 65/14 68/19 68/20 68/24 69/9 70/21 72/14 73/12 73/12 74/4 anybody [2] 18/13 69/23 anything [9] 5/15 11/14 20/8 26/24 26/24 59/16 63/14 64/2 67/7 Anytime [1] 25/7 apologize [1] 59/14 appear [2] 7/21 51/10 APPEARANCES [1] 2/1 appears [1] 20/9 applied [1] 62/8 applies [1] 65/7 appreciate [1] 47/3 approach [2] 13/1 28/21 April [1] 10/21 arbitrary [3] 47/23 48/3 56/25 are [56] 5/17 5/17 6/10 11/15 15/7 16/14 18/7 18/9 19/11 21/12 21/21 25/8 25/20 26/4 26/15 26/19 27/8 29/9 29/12 29/13 34/25 35/8 35/19 36/4 38/2 42/22 50/10 50/15 50/16 50/18 50/21 51/1 51/22 52/2 52/15 53/18 53/25 54/24 56/1 57/19 58/6 58/12 61/17 61/21 62/20 64/8 64/23 64/24 66/14 66/23 67/7 68/6 68/8 68/14 68/20 71/3 area [3] 33/14 34/14 43/8 Argumentative [1] 59/8 ARIA [3] 2/13 70/21 70/24 ARREY-MBI [1] 2/11 Article [1] 72/5 as [83] Asian [13] 22/7 22/11 24/1 24/12 24/19 24/24 29/3 30/14 32/3 32/15 32/24 34/25 38/15 Asian-Hispanic [1] 32/15 Asians [5] 24/23 27/5 29/8 31/24 32/23 ask [13] 4/24 5/1 5/5 13/16 14/17 19/20 28/20 35/7 43/18 45/9 46/24 52/14 61/3 asked [3] 32/8 42/3 46/21 asking [41] 5/17 5/18 17/11 17/22 19/11 21/8 23/6 24/20 25/4 25/12 25/22 25/23 26/17 28/18 31/14 32/2 32/7 35/13 36/5 36/9 36/10 37/2 37/22 44/14 44/16 44/21 45/19 45/23 46/5 47/4 48/20 49/4 49/23 50/10 52/18 53/20 61/19 62/1 63/15 66/24 67/5 Assembly [1] 64/12 Association [3] 8/1 8/21 9/20 assume [2] 50/21 60/16 assumed [2] 40/15 40/16 assuming [1] 40/22 at [60] 5/4 5/15 9/7 9/25 10/2 10/15 10/23 11/4 12/8 15/4 15/23 16/4 16/19 17/19 26/8 28/17 28/24 30/9 30/11 32/12 32/12 33/18 34/13 35/5 35/12 35/16 36/11 36/23 37/13 40/2 41/21 43/11 45/11 45/14 45/24 45/25 46/6 46/20 47/24 50/9 50/19 51/6 52/20 54/8 56/8 56/10 56/22 59/14 60/5 62/20 63/4 65/10 65/21 67/14 67/16 67/18 68/12 68/18 68/18 68/21 ATLANTA [6] 1/2 1/21 2/21 33/13 33/15 33/18 Atlantic [1] 1/20 attached [1] 73/19 attempt [1] 5/1 attempting [1] 59/14 attention [2] 7/24 21/22 attributable [3] 22/7 24/18 24/22 AUDRA [1] 2/10 available [7] 13/10 40/2 50/2 50/12 57/12 57/13 69/22 Avenue [1] 2/5 avoiding [1] 64/9 aware [2] 53/18 64/14 away [1] 68/24 B B-1832 [2] 72/22 73/25 back [5] 14/24 15/17 24/17 26/9 70/1 background [1] 63/9 Barksdale [1] 54/4 baseball [1] 34/12 based [6] 16/9 16/11 36/13 41/7 42/1 57/10 basically [1] 52/18 basing [1] 16/6 basis [6] 23/6 23/17 32/2 32/11 40/5 48/4 Bayesian [1] 13/1 BBM [1] 1/5 be [33] 5/5 7/21 13/19 13/22 14/21 15/25 19/24 24/18 24/22 25/9 26/1 26/10 35/7 39/24 41/6 41/10 42/17 42/19 42/21 44/13 48/2 48/17 48/21 49/1 49/7 51/10 56/9 56/16 57/1 62/8 65/6 65/11 72/14 became [3] 18/23 19/15 31/21 because [18] 17/4 18/9 23/24 32/9 33/24 41/18 42/10 49/3 49/11 50/4 53/12 56/2 56/10 56/14 58/15 59/2 65/4 67/4 become [1] 11/7 been [18] 4/9 4/20 7/13 17/16 33/1 33/2 33/5 33/6 33/22 34/8 34/13 34/17 49/18 49/21 50/11 53/13 58/6 63/1 before [6] 4/20 9/16 31/17 31/21 45/10 74/22 beginning [3] 68/13 68/18 68/21 BEHALF [3] 2/2 2/9 2/17 behave [1] 35/1 behavior [30] 23/11 23/21 24/16 27/4 27/9 27/24 28/6 28/11 30/14 30/16 31/24 32/22 35/6 35/16 35/21 36/18 36/20 37/1 37/9 37/13 38/2 38/6 38/10 38/14 38/19 38/21 61/20 61/22 61/23 61/25 behind [2] 62/3 63/9 65/17 65/18 believe [11] 6/5 6/13 8/11 9/12 9/24 10/7 10/14 10/20 61/6 61/9 67/17 below [1] 74/4 beside [1] 57/21 best [3] 64/1 70/16 74/3 between [9] 20/16 24/3 24/7 29/4 30/1 32/16 39/6 49/16 62/16 beyond [2] 53/22 54/8 bit [8] 11/7 13/2 20/14 30/18 44/22 51/4 62/3 63/9 Black [61] 16/20 17/19 17/19 17/25 18/1 18/2 18/2 18/12 18/13 18/17 18/24 19/15 22/1 22/2 22/8 22/12 24/2 24/14 24/18 24/21 26/10 27/2 27/2 28/12 32/5 36/1 36/11 36/12 36/13 36/14 45/6 45/22 46/2 46/3 46/9 46/25 51/20 52/11 52/15 52/20 53/1 53/5 53/8 53/10 53/25 61/11 61/20 61/25 66/3 66/7 66/9 66/11 66/14 66/15 66/20 66/20 66/25 67/1 67/6 67/8 67/10 Blacks [11] 24/8 24/25 25/4 25/8 26/11 53/15 62/7 62/7 62/13 62/14 67/3 block [6] 39/5 40/2 41/6 41/21 41/25 50/23 block-level [1] 39/5 blocks [10] 37/1 37/9 39/15 39/20 40/11 40/14 40/17 40/19 40/21 42/1 Board [4] 8/3 8/21 10/11 72/5 border [2] 33/11 33/11 both [4] 39/23 40/12 49/11 66/1 bottom [5] 26/7 43/13 45/14 45/24 65/25 boundaries [12] 21/3 29/5 29/6 32/17 32/17 45/8 45/25 46/7 46/9 46/12 47/5 47/12 bounds [2] 12/24 62/21 box [13] 11/10 11/13 11/15 11/25 12/1 12/8 12/15 18/13 18/16 25/13 25/20 26/8 26/12 BRANCH [1] 2/13 Braves [1] 34/12 break [4] 5/4 5/6 30/7 67/13 breakdown [1] 18/8 BRIAN [3] 1/6 1/13 2/18 briefer [1] 55/22 bring [1] 41/22 broader [1] 56/8 BROCKINGTON [2] 2/20 72/12 BROOKS [1] 2/9 Brown [3] 8/23 10/3 31/6 BRYAN [1] 1/19 Bureau's [1] 18/7 burn [7] 14/16 14/18 68/8 68/11 68/12 68/17 69/3 burn-in [3] 68/11 68/12 68/17 burn-ins [2] 68/8 69/3 bus [1] 33/25 but [45] 6/8 9/15 11/8 12/5 12/6 15/4 15/22 17/6 17/14 20/16 25/17 26/6 27/4 27/11 27/25 28/17 29/10 29/13 34/4 34/23 36/24 38/12 38/17 39/3 40/24 41/11 41/24 42/13 44/4 45/2 47/24 51/4 53/7 54/9 54/18 55/21 59/6 60/17 61/24 62/24 68/9 69/1 69/8 70/3 70/17 BVAP [4] 44/9 44/12 44/13 44/15 C C-H-E-N [1] 4/17 calculating [1] 47/9 calculations [1] 13/12 call [2] 68/4 70/14 called [5] 13/8 68/8 69/21 70/15 70/18 came [3] 18/5 23/19 24/11 can [18] 6/10 7/14 11/19 17/5 23/17 23/25 24/6 26/17 26/23 26/24 27/2 30/5 37/16 38/9 60/5 64/18 66/16 67/9 can't [10] 6/7 9/14 12/6 19/9 29/21 37/15 41/17 42/10 44/2 54/17 candidate [20] 18/24 19/16 20/16 22/2 24/9 52/4 52/11 52/17 52/19 52/20 52/22 52/25 52/25 53/2 53/10 53/16 54/1 63/6 66/4 66/10

77 C CHEN [14] 1/16 3/11 complete [3] 19/24 53/7 61/9 61/14 63/23 54/8 57/12 57/14 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 77 of 166 candidates [11] 22/13 24/14 24/22 32/5 33/4 51/21 53/5 53/8 54/18 61/4 66/7 candidates' [1] 60/23 capacity [3] 1/6 1/13 2/18 caption [1] 73/5 Carolina [6] 31/4 31/9 31/11 35/2 35/5 35/8 case [20] 1/4 5/12 7/17 8/7 9/16 9/25 10/12 10/13 10/22 11/2 11/4 31/15 31/18 36/5 37/11 55/25 61/24 62/9 73/11 73/13 cases [3] 1/10 7/25 31/22 category [2] 25/10 25/17 Caucasian [4] 52/5 52/23 52/25 53/15 Cause [4] 10/16 10/18 35/10 37/11 caused [1] 58/16 causes [1] 59/3 CAVE [1] 1/19 CCR [3] 1/22 72/21 73/24 CELESTE [1] 2/3 census [21] 5/24 18/7 25/10 25/12 26/2 26/6 39/4 39/4 39/14 39/20 40/11 40/14 40/17 40/19 40/20 45/5 45/11 49/18 57/13 59/11 59/11 Center [1] 1/20 certainly [14] 5/23 7/10 17/6 20/13 33/14 33/17 35/19 35/24 38/7 38/12 50/11 68/7 68/15 68/18 Certificate [2] 72/22 73/25 certification [1] 73/14 certified [4] 72/7 72/21 73/18 73/24 certify [3] 73/4 73/9 74/2 challenge [1] 9/12 change [11] 12/2 33/3 43/5 56/4 74/6 74/8 74/10 74/12 74/14 74/16 74/18 changes [8] 11/10 11/15 11/24 12/1 12/8 32/25 43/4 74/4 chart [2] 36/2 36/3 Chattanooga [2] 33/12 33/21 check [3] 25/13 26/8 26/12 checked [3] 18/13 18/16 26/7 checking [1] 18/14 3/12 4/8 4/15 4/20 7/14 23/9 67/17 67/23 70/19 72/4 74/2 74/21 Chen's [2] 54/25 55/1 choice [2] 11/11 48/3 chose [1] 56/19 chosen [4] 47/24 48/1 48/2 57/1 Citizens [2] 8/20 9/19 city [4] 10/10 31/8 35/10 35/20 CIVIL [1] 2/5 clarify [2] 11/13 13/16 clarity [1] 55/7 clear [4] 26/1 40/24 56/3 69/16 clearly [3] 44/24 53/14 70/11 close [1] 34/5 closest [1] 33/13 Coalition [1] 9/23 code [13] 13/7 13/8 13/10 13/10 67/24 68/3 69/5 69/24 70/4 70/6 70/8 70/9 70/12 codes [2] 7/2 13/8 coefficient [4] 14/8 17/5 17/7 17/10 coefficients [1] 14/13 cohesion [2] 22/20 22/25 COIE [1] 2/13 COLEY [1] 2/3 collaborators [1] 69/25 colloquies [1] 73/6 Colored [1] 8/1 column [2] 46/5 47/13 combined [2] 12/23 24/4 come [1] 13/24 Commission [1] 54/12 Commissioners [1] 8/3 COMMITTEE [1] 2/5 Common [4] 10/16 10/18 35/10 37/11 commonly [2] 12/12 38/18 commonly-studied [1] 38/18 Commonwealth [1] 10/25 comparable [1] 21/18 comparative [1] 29/20 comparatively [1] 29/9 compare [1] 49/13 compared [3] 19/17 20/17 66/14 compelled [1] 56/15 23/13 23/16 completely [4] 20/20 41/24 49/24 50/3 comply [1] 65/13 composition [14] 18/5 18/8 18/11 44/4 47/15 47/21 48/1 55/10 56/4 57/4 57/9 57/17 58/14 62/16 comprise [2] 20/24 21/9 computer [10] 7/2 13/7 13/7 13/10 69/24 70/4 70/6 70/8 70/9 70/12 concluded [1] 50/7 conclusion [8] 19/2 19/9 19/22 24/11 30/14 31/25 32/1 63/3 conclusions [1] 26/18 conduct [3] 17/21 59/3 62/10 conducted [4] 28/22 62/4 68/21 69/12 conducting [1] 28/19 CONFERENCE [1] 2/2 conferences [1] 34/14 confidence [2] 14/7 16/22 confident [2] 38/12 38/17 confirm [8] 41/10 41/17 42/10 44/21 45/2 46/1 53/24 54/1 confirmed [5] 21/20 32/21 32/21 32/25 39/18 confirms [1] 24/14 connection [5] 6/2 6/12 7/3 70/4 70/12 consequences [1] 56/5 consideration [3] 50/9 57/6 57/7 considered [1] 7/6 consistent [2] 53/19 61/24 CONSOLIDATED [1] 1/10 constituting [1] 66/9 contacted [1] 72/11 contains [2] 44/9 69/4 contest [1] 58/5 context [3] 28/18 51/3 55/25 contract [1] 72/14 convention [1] 63/1 copy [6] 7/22 42/12 51/7 51/11 70/25 71/1 CORETTA [1] 2/12 correct [25] 6/23 7/21 10/1 10/5 17/13 18/18 21/10 25/20 27/21 28/1 28/13 39/1 47/2 47/8 48/24 51/10 64/11 65/2 67/25 73/8 74/3 correctly [2] 14/14 52/10 Corrine [3] 8/23 10/3 31/6 could [13] 11/12 13/16 14/17 19/4 25/2 28/5 28/5 33/10 45/9 52/14 56/5 61/3 66/19 couldn't [14] 10/8 12/13 25/3 25/16 34/15 34/22 38/11 38/16 48/8 48/15 59/2 60/17 60/17 69/7 Council [1] 72/6 counsel [7] 2/1 5/10 9/21 9/22 11/1 73/10 73/11 count [2] 33/9 33/10 counted [1] 58/12 counties [8] 29/21 64/10 64/18 64/21 64/24 65/5 65/15 65/18 counting [1] 21/2 counts [1] 58/7 county [19] 8/3 8/21 10/11 29/13 29/14 29/17 29/18 34/6 34/9 34/16 34/18 34/20 34/24 50/17 50/21 50/23 65/8 72/2 73/3 couple [5] 4/24 30/7 32/13 66/23 67/19 course [3] 7/5 7/7 52/22 courses [2] 38/6 38/7 court [8] 1/1 4/5 72/5 72/7 72/12 72/21 73/18 73/24 covered [1] 38/13 creating [1] 18/4 creative [1] 70/17 Creek [1] 1/24 criteria [1] 6/7 criticize [2] 63/11 63/13 crossover [1] 67/3 cultural [1] 33/17 CUNNINGHAM [1] 2/10 CV [1] 1/4 CVR [3] 1/22 72/21 73/24 CVR-M [3] 1/22 72/21 73/24 D DARRYL [1] 2/10 data [39] 5/16 5/23 5/24 5/24 7/2 13/15 13/18 15/4 15/4 15/22 15/23 17/5 18/4 26/2 26/3 26/6 39/4 39/5 39/5 40/1 41/21 41/25 49/12 49/16 49/18 49/19 50/2 50/8 50/12 57/14 59/11 59/11 59/12 59/25 62/6 62/12 Davis' [1] 12/23 day [5] 20/25 34/2 34/2 73/20 74/23 DC [2] 2/6 2/15 de [2] 58/23 58/25 Dec [1] 3/11 December [6] 5/20 5/23 7/11 7/16 11/4 16/4 December 22nd [1] 7/16 decision [2] 56/11 56/14 default [1] 68/15 defeat [1] 66/10 defeating [1] 66/20 Defendant [6] 1/8 1/15 2/17 8/24 10/4 10/5 DEFENDANT'S [3] 3/9 7/14 51/7 Defendants [1] 8/8 define [3] 14/17 43/19 61/3 Democrat [3] 23/4 54/19 67/2 Democratic [26] 18/24 19/16 20/15 22/2 22/13 24/9 24/14 24/22 25/5 25/9 32/5 33/4 51/20 52/3 52/11 52/17 52/19 52/22 53/2 53/5 53/8 53/15 54/1 60/23 66/3 66/10 Democrats [5] 25/2 50/17 50/19 50/23 61/7 demographic [5] 5/16 5/24 32/25 49/18 57/13 demographics [5] 29/1 29/13 29/14 29/20 30/3 denied [1] 73/15 deposed [7] 4/20 9/8 9/15 10/1 10/6 10/13 11/3 deposition [28] 1/16 5/9 6/4 6/17 6/22 6/24 6/25 9/25 10/20 11/23 12/5 16/5 16/7 16/12 39/1 39/2 39/8 39/13 40/9 40/9 41/3 41/9 42/11 49/6 49/7 57/11 70/25 74/3 depositions [5] 11/16 39/24 40/1 41/1 49/15 describe [1] 12/19 described [1] 43/1 describes [2] 13/4 13/5 describing [6] 12/11 19/6 42/14 42/23 59/9 68/16 description [2] 3/10

78 D 48/18 48/19 48/22 54/9 54/25 55/9 26/5 52/7 53/9 53/12 9/5 9/14 9/24 10/6 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 78 of 166 description... [1] 58/12 detail [2] 55/18 57/23 detailed [1] 18/7 details [1] 62/3 determine [2] 37/8 57/25 determining [2] 57/7 58/4 detour [1] 34/1 Detzner [9] 8/15 8/18 8/23 9/4 9/11 30/24 31/1 37/14 37/22 developed [1] 69/25 development [1] 70/1 deviation [4] 56/20 64/12 64/13 64/25 did [41] 5/8 5/14 7/11 8/4 9/1 9/6 9/7 13/15 14/14 14/21 19/22 22/21 22/23 22/24 23/2 26/8 26/12 29/19 35/5 35/13 36/12 36/15 36/18 36/20 37/5 37/7 37/10 37/12 37/19 37/21 37/24 41/20 47/9 54/5 54/7 54/20 59/2 60/3 60/16 63/15 64/1 didn't [9] 7/10 8/5 20/5 20/8 27/16 27/18 27/25 36/24 60/15 difference [2] 29/25 65/4 different [7] 9/5 13/2 29/14 39/6 39/20 40/20 55/23 direction [1] 73/7 directly [2] 20/8 20/10 disagree [1] 53/6 disassembly [1] 73/15 discard [2] 14/19 14/20 disclosure [2] 72/1 72/6 discussed [5] 11/15 24/5 32/19 51/1 61/16 discussing [2] 41/8 42/14 discussion [4] 11/21 15/16 41/15 41/16 disqualified [1] 72/8 distinguish [3] 39/19 40/13 49/16 distributed [1] 28/5 district [61] 1/1 1/1 8/2 16/16 18/22 19/10 19/14 20/12 20/22 21/3 21/7 21/9 21/14 22/1 22/9 22/20 23/1 24/3 27/21 29/1 29/6 30/15 32/15 32/16 43/4 43/6 43/14 45/1 45/7 45/12 45/25 46/6 46/7 46/8 46/12 47/4 47/12 47/13 48/6 50/8 51/4 52/12 52/16 Dr. Chen [6] 4/20 53/12 54/12 56/4 58/1 7/14 23/9 67/17 67/23 58/1 58/5 60/12 60/24 70/19 62/9 64/22 66/2 66/2 Dr. Chen's [1] 55/1 66/13 67/2 67/11 Dr. Strangia's [1] districting [1] 65/7 6/24 districts [6] 30/4 drafter [1] 60/1 32/18 65/6 65/10 drafters [2] 63/11 65/16 65/18 63/22 DIVISION [1] 1/2 draw [6] 19/2 19/9 do [73] 4/16 5/2 5/8 19/22 26/18 31/25 6/15 8/13 11/9 11/14 32/1 11/23 13/15 14/4 14/5 drawer [3] 15/3 14/16 14/19 14/20 15/21 16/1 15/25 17/7 17/9 19/2 drawers [2] 63/13 20/12 23/3 24/19 63/14 27/14 27/18 27/23 drawing [3] 16/8 27/25 28/2 28/2 28/10 39/3 64/12 30/20 30/21 31/25 drawn [3] 18/23 31/25 35/1 36/15 37/7 19/14 58/6 37/10 37/12 37/19 duly [1] 4/9 38/22 39/7 39/10 41/2 Duncan [1] 12/23 41/12 41/12 42/8 during [1] 67/23 43/16 43/20 43/21 46/11 46/14 48/6 E 48/13 49/22 50/5 E-I [1] 70/16 50/16 50/18 50/22 each [12] 14/8 14/12 52/10 53/7 54/3 56/19 16/20 40/6 56/11 58/21 59/7 61/5 61/17 56/13 57/8 57/16 66/4 61/18 64/11 64/15 67/24 69/4 69/4 64/18 68/1 68/15 69/4 earlier [3] 51/1 61/16 74/2 68/16 document [8] 6/6 6/9 early [1] 70/2 19/5 20/7 23/15 43/10 easily [1] 33/21 44/2 64/7 ecological [42] 12/20 documents [4] 6/10 13/11 16/15 16/15 6/13 6/16 7/10 16/19 17/1 22/18 does [17] 7/21 16/24 22/19 24/7 27/6 27/6 18/2 19/8 19/21 44/12 27/12 27/12 27/15 50/5 51/10 51/12 58/4 27/17 27/19 27/19 61/19 62/14 62/23 28/7 28/7 28/9 28/10 62/24 68/3 68/7 68/17 28/19 36/2 36/13 doesn't [6] 19/8 36/16 36/24 36/25 19/21 25/11 58/8 65/4 37/6 37/7 37/8 37/19 68/23 37/20 37/25 61/10 doing [3] 22/24 35/4 61/19 62/4 62/23 62/12 62/24 63/10 69/11 don't [21] 4/25 10/6 70/1 70/18 10/7 11/8 12/4 12/9 education [1] 38/4 12/10 12/12 14/16 effectively [2] 50/9 17/14 18/9 26/13 50/13 27/11 32/23 41/11 EI [23] 13/8 14/1 41/18 42/10 42/12 14/10 17/21 22/21 54/16 58/2 63/13 22/25 24/21 27/1 done [4] 22/18 31/13 28/22 32/18 53/1 70/23 73/17 68/13 68/16 68/20 down [9] 15/10 17/3 68/20 69/1 69/1 69/9 25/20 33/15 33/18 69/21 70/7 70/10 34/22 41/6 61/12 73/5 70/14 70/15 Dr [3] 7/3 54/16 election [26] 7/6 7/7 54/25 7/8 8/3 13/21 14/1 Dr. [15] 4/20 5/13 17/10 20/25 30/4 6/24 7/1 7/14 23/9 36/23 39/5 41/4 45/17 53/4 54/9 54/22 54/25 52/4 52/5 52/12 52/16 55/1 55/9 67/17 67/23 53/14 54/8 58/1 58/5 70/19 58/11 59/12 61/5 66/4 Dr. Alford [3] 5/13 67/24 53/4 54/22 elections [9] 8/21 Dr. Alford's [4] 7/1 10/11 16/20 19/17 53/13 electoral [2] 5/15 5/23 electorate [3] 22/8 24/2 47/1 elementary [1] 33/25 empirical [1] 59/16 employ [1] 73/11 end [4] 15/2 15/10 65/4 65/25 ended [2] 60/10 60/11 engaged [3] 8/8 8/16 10/19 enough [1] 28/4 entailed [1] 68/2 entirely [1] 58/16 equal [2] 65/9 65/14 ER [4] 17/22 22/21 22/25 28/23 ERRATA [1] 74/1 ESQ [3] 2/4 2/13 2/19 essence [1] 40/4 estimate [12] 17/2 17/16 24/8 24/18 24/21 40/5 41/22 62/1 62/5 62/11 62/11 63/2 estimated [1] 39/13 estimates [10] 13/25 16/16 16/20 17/2 17/22 27/1 28/19 42/1 61/20 62/19 estimating [3] 40/10 40/20 62/6 Estimation [1] 12/23 et [3] 1/3 1/10 8/14 even [4] 29/23 29/25 47/24 56/3 evening [1] 5/11 ever [2] 10/7 11/5 every [7] 33/19 34/1 34/2 66/25 67/1 67/6 67/10 everybody [1] 26/11 everyone [2] 18/16 66/19 evidence [1] 73/9 exact [1] 69/8 exactly [11] 9/17 11/17 12/14 13/5 27/14 28/22 29/11 36/4 40/17 42/18 68/1 examination [4] 4/11 67/21 67/23 69/14 EXAMINATIONS [1] 3/1 examined [1] 4/9 example [2] 13/25 18/9 examples [1] 53/25 exception [1] 74/4 exclusively [1] 17/19 exhibit [5] 7/14 7/16 7/18 51/13 55/4 Exhibit 1 [1] 7/14 exhibited [1] 52/5 exhibits [2] 3/8 73/17 expert [30] 1/16 6/9 6/12 7/1 7/17 8/6 9/3 10/19 11/2 31/17 31/19 31/22 35/13 36/16 36/21 37/2 37/10 37/16 37/19 39/4 52/8 53/22 54/10 54/20 70/5 70/13 expertise [1] 31/20 explain [5] 17/5 18/3 42/21 55/17 62/3 explained [3] 17/15 55/20 68/24 explanations [1] 57/23 expressly [1] 73/14 extent [4] 31/20 53/16 53/17 53/18 extremes [1] 62/20 F fact [10] 13/6 27/2 27/9 27/23 28/10 33/1 42/25 46/1 46/23 50/7 familiar [2] 11/6 11/7 familiarity [1] 11/9 family [1] 33/15 far [1] 50/3 favor [3] 18/24 19/15 54/1 feature [2] 12/1 12/12 February [6] 1/17 6/19 7/9 10/14 13/12 74/3 feel [1] 5/1 Ferguson [1] 8/2 Ferguson-Florissant [1] 8/2 few [4] 7/4 9/5 34/13 38/8 fifth [2] 51/24 65/25 figure [1] 42/1 file [1] 41/23 filed [2] 8/10 8/12 files [8] 6/1 6/2 6/3 7/2 13/8 18/6 18/6 18/7 final [2] 23/20 24/12 finally [3] 15/1 15/8 15/20 find [4] 6/20 6/20 44/2 51/22 first [7] 4/9 12/14 17/12 21/25 32/13 56/2 65/24 first-hand [1] 12/14 fitting [3] 62/5 62/12 62/15 five [2] 64/20 65/1 Flat [3] 58/17 59/5 60/20 Floor [1] 1/20 Florida [6] 8/18 9/12 30/24 31/2 31/6 35/2 Florissant [1] 8/2 focusing [1] 57/20 follow [1] 45/10 following [2] 72/6 74/4 follows [1] 4/10

79 F GREENBAUM [1] 2/4 58/14 60/2 61/18 62/1 63/8 63/10 increased [4] 22/2 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 79 of 166 forced [1] 56/16 foregoing [3] 73/4 73/7 73/16 formula [1] 41/14 found [4] 15/2 15/20 50/4 66/1 four [1] 61/6 fourth [1] 51/24 free [1] 5/1 freely [1] 69/22 front [8] 41/9 41/11 41/19 42/11 42/13 42/17 42/24 54/16 full [1] 58/20 FULTON [2] 72/2 73/3 further [3] 59/16 69/13 73/9 G GA [2] 1/21 1/24 Gary [2] 13/5 69/20 gave [1] 30/18 general [17] 23/11 23/18 24/15 30/16 31/23 32/21 32/24 35/22 37/18 38/1 42/14 54/2 54/2 64/11 65/3 65/8 65/19 generally [7] 23/22 26/21 38/12 41/16 51/20 53/5 53/20 geographic [2] 39/6 59/17 geography [7] 12/3 41/4 41/5 41/6 59/10 59/20 59/25 GEORGIA [24] 1/1 1/7 1/14 2/2 2/17 2/18 2/21 31/13 31/17 31/25 32/6 33/5 33/7 33/12 33/23 34/1 34/3 35/1 54/6 72/2 72/6 72/7 72/15 73/2 get [3] 14/5 50/9 61/17 Gina [5] 6/5 6/25 11/23 16/6 16/7 give [10] 6/19 12/14 19/4 19/25 23/12 23/14 23/16 43/9 43/25 55/21 given [6] 14/12 18/9 23/13 50/1 69/11 73/9 giving [1] 54/19 gleaned [2] 40/8 40/25 go [9] 11/19 15/14 15/17 42/5 50/3 52/2 52/3 56/24 57/3 goes [2] 31/23 69/5 going [8] 25/8 26/9 32/9 34/11 34/20 35/7 43/18 66/23 gone [2] 33/15 33/18 got [3] 36/11 47/7 61/11 granular [1] 49/12 greater [1] 61/12 Greensboro [16] 10/10 10/12 31/8 35/10 35/14 35/21 35/23 35/25 36/8 36/8 36/9 36/12 36/17 36/19 36/21 37/3 grew [3] 33/10 33/20 34/4 group [4] 25/1 25/8 28/7 63/4 grouping [1] 27/1 groups [1] 32/22 growing [1] 33/19 guidelines [5] 56/18 56/21 64/1 64/3 64/8 Guilford [1] 10/11 Gwinnett [7] 29/14 29/18 29/20 34/6 34/8 34/11 65/15 H had [18] 6/3 15/3 15/22 16/4 32/25 33/25 35/21 39/19 39/22 40/2 40/8 40/12 40/12 41/21 46/21 49/16 50/13 53/2 half [1] 55/20 hand [2] 7/13 12/14 happen [2] 62/23 62/24 happening [1] 53/19 happens [2] 61/18 63/8 happy [8] 5/5 30/19 41/10 42/17 42/19 42/21 53/24 55/17 Harbins [5] 43/5 43/8 43/14 43/16 43/21 has [5] 25/10 27/11 32/15 33/2 63/1 have [44] 4/20 4/21 7/10 11/5 15/20 17/16 22/18 25/8 26/13 31/13 31/14 33/5 33/6 34/2 34/8 34/12 34/17 35/7 38/7 39/13 41/11 41/18 41/18 42/10 42/11 42/12 43/1 43/18 47/7 49/18 49/21 50/11 53/25 54/13 54/15 54/16 56/5 58/5 60/19 65/6 67/11 67/18 67/19 74/2 haven't [2] 12/7 23/13 having [1] 4/9 HD [1] 52/7 he [7] 39/10 41/3 41/10 41/13 41/17 52/23 54/4 He'll [1] 71/4 head [11] 6/11 9/15 10/9 25/17 34/15 34/23 38/12 38/17 48/9 48/16 69/8 hearing [1] 73/9 heavily [3] 44/25 held [2] 11/21 15/16 help [3] 30/20 42/17 42/19 Henry [11] 29/13 29/17 29/20 34/16 34/18 34/20 34/23 50/17 50/21 50/23 65/15 her [4] 11/23 16/8 42/8 60/17 here [19] 14/8 16/4 17/11 18/12 23/23 23/25 26/4 28/19 36/5 41/11 42/12 45/2 51/23 52/22 53/22 54/6 55/22 66/24 67/9 hereby [2] 73/4 74/2 herein [1] 74/2 Hickory [3] 58/17 59/5 60/20 high [3] 17/9 34/21 52/21 Hillbrooke [1] 1/24 him [2] 39/8 41/8 hired [6] 8/13 8/24 9/13 9/23 10/12 11/1 his [11] 1/6 1/13 2/18 13/6 13/6 39/12 39/13 53/4 53/6 54/21 69/25 Hispanic [30] 22/7 22/11 22/12 22/20 23/1 23/4 24/1 24/12 24/13 24/18 24/24 27/20 27/20 27/22 27/24 28/17 29/2 30/14 32/3 32/4 32/15 32/24 34/25 35/6 35/17 35/20 35/23 36/3 37/13 38/10 Hispanics [10] 24/23 27/5 27/25 28/2 28/4 28/6 28/11 29/7 31/24 32/23 history [2] 5/25 26/3 House [55] 7/7 14/1 16/16 17/10 18/22 18/24 19/9 19/14 19/16 20/12 20/22 21/3 21/14 22/1 22/9 22/20 23/1 24/2 27/20 29/1 29/6 30/4 30/15 32/16 43/4 43/6 43/13 45/7 45/12 45/25 46/6 46/8 46/12 47/4 47/12 47/13 48/6 48/22 50/8 51/3 52/4 52/5 52/7 52/12 52/15 53/12 58/1 58/1 60/11 60/24 66/1 66/2 66/13 67/1 67/11 how [40] 4/16 4/22 9/18 12/2 13/14 14/4 14/4 17/1 17/9 23/18 24/19 26/14 26/24 28/22 33/5 33/6 34/6 34/8 34/17 36/18 36/20 47/18 57/7 57/8 57/16 57/21 58/10 58/12 61/4 61/17 67/7 68/8 69/3 69/11 70/8 70/9 hundreds [1] 33/21 hypothetical [2] 66/24 67/6 I I'd [4] 5/5 7/13 21/22 29/9 I'll [9] 6/20 6/20 18/3 25/15 25/17 29/24 44/5 55/21 67/12 I'm [27] 4/15 9/17 10/8 11/6 11/12 11/17 26/3 26/17 30/19 32/8 33/23 34/14 35/7 38/12 38/17 39/3 45/24 45/25 46/15 46/22 53/24 54/19 55/17 64/14 64/14 64/17 69/16 I've [9] 11/6 23/13 28/16 31/14 33/22 34/13 46/15 51/1 64/3 identification [1] 18/12 identifications [1] 26/4 identified [2] 7/19 51/14 identify [5] 7/25 18/5 18/10 18/10 28/6 if [52] 4/24 5/4 6/19 9/15 11/12 12/16 12/16 13/16 13/21 14/17 14/24 17/2 19/4 19/19 19/25 21/8 23/12 23/14 25/4 25/12 25/19 26/17 26/17 28/4 28/24 30/19 31/14 37/15 41/8 42/16 42/23 43/9 43/11 43/24 43/25 44/14 45/3 45/9 48/20 51/16 52/14 59/14 61/3 64/4 64/19 64/19 64/25 65/21 66/18 67/1 67/10 74/4 illustrate [1] 20/23 illustrated [1] 22/9 important [5] 39/16 56/1 56/7 57/6 57/15 impossible [4] 49/7 49/24 50/4 63/3 in [327] INC [4] 1/23 72/10 72/11 72/14 include [4] 18/2 25/2 25/5 54/3 included [4] 26/10 27/8 27/22 38/5 includes [6] 18/1 25/8 26/11 53/14 69/2 69/2 including [2] 49/18 73/16 increase [7] 20/14 20/17 20/21 22/6 22/7 24/1 24/7 29/7 29/8 32/15 increases [2] 29/10 29/12 incumbent [1] 48/7 indeed [5] 26/7 53/24 58/10 60/13 66/16 independent [2] 27/12 27/15 INDEX [2] 3/1 3/8 individual [19] 2/9 2/9 2/10 2/10 2/10 2/11 2/11 2/11 2/12 2/12 47/18 56/11 56/13 56/14 57/3 57/3 57/4 57/8 57/16 individuals [3] 26/5 27/3 27/9 individuals' [1] 18/8 inference [22] 12/20 13/11 16/15 22/18 24/8 27/6 27/12 27/15 28/7 28/9 28/19 36/13 36/16 36/24 37/6 37/8 37/20 37/25 62/24 69/12 70/1 70/18 inference/ecological [1] 37/20 inferences [1] 36/25 information [4] 12/14 24/4 24/10 40/25 initial [1] 21/23 inputs [1] 68/6 ins [2] 68/8 69/3 inside [3] 45/7 54/23 58/15 installation [1] 70/5 instances [1] 53/19 Instead [1] 57/2 intentionally [1] 68/23 interest [1] 72/9 interested [1] 73/13 Internet [1] 69/23 Intervenors [3] 8/24 10/4 10/5 into [12] 27/2 33/22 34/1 34/2 39/9 44/7 48/19 48/22 55/21 58/6 60/3 65/6 intuition [1] 24/15 involved [2] 61/4 65/11 involves [1] 68/12 involving [2] 65/16 65/17 irrelevant [7] 47/17 55/11 55/13 55/15 55/16 55/20 57/21 is [162] isn't [3] 29/18 30/1 44/15 issue [7] 37/13 37/15 51/19 51/25 52/1 61/16 64/8 issues [1] 56/1 it [84] it's [45] 13/2 13/8 15/10 21/16 22/13

80 I 52/1 52/6 66/9 50/14 Midtown [1] 2/20 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 80 of 166 it's... [40] 28/14 29/17 29/23 29/25 33/21 33/23 40/10 41/22 43/20 44/21 45/17 47/17 47/23 48/3 50/1 50/3 54/9 54/14 55/20 55/23 56/7 56/24 59/6 59/7 61/24 62/6 65/13 66/18 69/11 69/18 69/19 69/20 69/21 69/22 69/22 69/23 69/23 70/15 70/16 70/18 iteration [3] 13/20 13/22 13/23 iterations [2] 68/18 69/6 its [6] 38/21 39/17 62/17 68/5 68/5 68/6 itself [1] 69/1 J J-O-W-E-I [1] 4/17 JACK [1] 2/19 JACKSON [1] 2/12 JAMAL [1] 2/9 JAMIDA [1] 2/11 January [3] 7/7 60/25 61/11 January 2018 [1] 7/7 jgreenbaum [1] 2/7 Jim [1] 54/3 jjp [1] 2/22 JOHN [1] 2/19 Johns [1] 1/24 JON [2] 2/4 71/2 JOWEI [8] 1/16 3/11 3/12 4/8 4/15 72/4 74/2 74/21 JR [1] 2/19 Judicial [1] 72/6 jungle [1] 61/2 jure [2] 58/23 58/25 just [59] 4/24 6/1 6/19 7/4 10/22 11/7 11/7 11/19 12/9 14/1 14/21 15/6 18/3 19/4 19/24 19/25 20/2 20/7 20/18 21/9 21/11 23/7 23/12 23/13 23/14 23/19 24/5 25/16 26/1 28/3 33/12 33/20 35/7 40/7 40/24 41/17 43/9 43/18 43/25 45/9 47/25 49/24 51/22 53/13 54/19 56/3 58/8 60/8 62/2 63/9 64/5 66/21 67/8 68/11 69/16 69/20 70/3 70/15 70/18 Justice [1] 9/23 K KEMP [3] 1/6 1/13 2/18 Ken [1] 8/14 kept [2] 68/14 68/21 key [3] 51/19 51/25 kin [1] 73/10 kind [1] 50/9 King [5] 13/4 13/5 13/9 69/20 69/24 King's [2] 67/24 68/2 knew [1] 57/11 know [25] 4/22 10/6 11/10 11/14 18/9 28/18 41/12 43/16 43/20 43/21 46/11 46/14 48/7 48/13 49/8 49/15 50/16 50/18 50/22 53/6 61/18 64/11 64/15 64/17 70/16 knowing [1] 26/14 knowledge [10] 23/11 23/18 23/21 24/15 30/16 31/23 32/21 38/2 39/4 74/3 known [1] 12/22 L lack [1] 55/7 lacked [1] 49/12 laid [3] 57/22 70/8 70/11 language [2] 13/9 69/21 large [3] 33/13 33/16 68/17 larger [2] 65/5 65/15 last [3] 5/10 22/13 25/19 later [1] 11/3 latter [1] 55/20 LAURETHA [1] 2/3 LAVELLE [1] 2/2 law [2] 2/5 72/15 Lawrenceville [9] 34/7 44/7 44/17 44/24 58/16 59/4 60/10 60/11 60/14 LAWYERS' [1] 2/5 lawyerscommittee.or g [1] 2/7 layout [1] 70/5 League [6] 8/17 9/10 9/11 9/21 10/24 31/1 least [3] 33/18 34/13 68/18 legislative [4] 6/6 49/8 49/20 66/3 Legislature [2] 56/16 56/19 Legislature's [3] 15/2 15/21 16/1 LEMON [1] 2/2 less [2] 66/16 66/18 let [6] 15/6 20/2 23/12 26/22 43/9 43/24 let's [7] 15/14 15/17 30/7 35/12 51/6 51/7 67/13 letters [1] 70/15 level [12] 14/7 15/5 15/23 39/5 40/2 41/4 41/5 41/6 41/21 41/25 LEWIS [2] 2/20 72/12 many [10] 13/14 14/4 might [1] 65/10 like [14] 7/13 12/15 14/4 33/5 33/6 33/9 12/16 13/3 16/13 34/6 34/8 34/17 61/4 21/22 30/17 30/20 map [5] 9/13 15/2 41/9 42/16 42/23 15/21 16/1 16/8 49/17 70/24 71/1 Maptitude [6] 11/5 Likelihood [1] 12/23 11/9 12/1 12/7 12/10 likely [8] 18/24 19/15 12/12 22/12 23/4 24/13 32/4 March [1] 73/20 64/23 64/24 Marianne [3] 1/22 limitation [1] 39/22 72/21 73/24 limits [2] 16/22 35/20 mark [1] 51/7 line [20] 15/8 25/23 marked [3] 7/13 7/19 26/7 51/24 51/24 53/2 51/14 62/6 62/12 62/15 marked/identified [2] 62/15 62/16 62/18 7/19 51/14 65/25 74/5 74/7 74/9 MARLON [1] 2/3 74/11 74/13 74/15 mask [1] 16/24 74/17 masks [1] 17/6 linear [1] 62/11 math [2] 44/20 45/2 lines [2] 15/10 58/5 matter [2] 54/2 65/20 listed [1] 64/3 maximum [3] 12/22 literally [1] 70/17 17/4 17/5 litigated [1] 37/16 may [1] 49/1 little [4] 11/7 13/2 maybe [2] 40/23 60/5 44/22 62/3 MBI [1] 2/11 live [2] 59/22 59/25 McDonough [2] lived [2] 33/14 48/7 34/16 34/17 lives [1] 59/22 McKENZIE [1] 2/10 ll [1] 33/9 me [67] 5/1 6/19 7/14 LLP [4] 1/19 2/13 9/23 11/1 11/13 12/6 2/20 72/12 15/6 16/3 16/3 19/4 logically [5] 17/4 19/11 19/25 20/2 17/17 24/4 62/21 63/3 20/13 21/8 23/6 23/12 logically-impossible 23/12 23/14 24/20 [1] 63/3 25/4 25/12 25/22 long [1] 31/21 25/23 26/17 26/22 longer [1] 30/18 30/20 32/2 32/3 32/7 look [7] 5/15 28/24 32/8 35/13 36/9 36/10 35/5 43/11 43/25 51/6 39/16 39/21 39/22 65/21 40/3 41/10 41/11 looked [7] 28/17 42/12 42/13 42/17 32/12 32/12 35/16 42/19 42/24 43/9 43/9 36/23 46/6 57/9 43/24 43/25 44/14 looking [9] 16/19 44/16 44/21 46/21 17/18 35/5 36/11 47/4 48/20 49/3 49/4 45/11 45/14 45/24 52/18 54/16 57/6 45/25 54/8 57/15 64/4 66/24 67/5 looks [1] 12/15 67/7 74/22 Louis [1] 8/3 mean [10] 11/13 love [1] 34/11 14/17 27/14 28/2 lower [7] 46/11 46/14 30/22 30/23 43/19 46/16 46/17 46/20 44/12 68/3 68/23 46/22 46/23 meaningless [1] lowest [1] 41/4 47/22 LYNNE [1] 2/11 meanings [1] 40/14 means [3] 47/16 56/7 M 65/10 made [4] 39/22 40/3 Meanwhile [1] 66/6 43/5 59/13 measures [1] 59/25 Mainly [1] 43/23 merely [1] 59/9 majority [3] 21/7 met [1] 5/10 21/9 44/15 Method [1] 12/24 make [13] 20/5 20/7 methodology [1] 23/13 33/25 36/25 12/19 40/7 40/24 45/10 50/5 metropolitan [2] 60/8 65/4 66/22 72/6 33/13 33/14 makes [1] 13/9 middle [2] 14/25 manipulation [1] 51/25 mile [1] 33/11 minority [5] 32/22 46/19 66/14 66/21 67/4 minus [5] 64/15 64/19 64/19 64/25 65/1 minutes [1] 30/8 misled [1] 59/15 Missouri [1] 7/25 moment [9] 6/20 14/11 19/4 19/25 23/12 23/14 24/5 43/9 46/21 months [1] 33/19 more [21] 7/4 18/23 19/15 22/11 23/4 24/13 29/17 32/4 35/8 42/5 44/17 44/25 51/4 53/20 58/14 60/2 64/18 64/21 64/24 64/24 68/19 moreover [1] 51/18 morning [1] 68/25 most [4] 53/13 64/23 65/10 70/16 mostly [1] 5/11 move [1] 44/15 moving [1] 44/6 Mr. [12] 39/7 39/12 39/17 40/8 41/3 42/22 49/7 50/13 57/11 67/23 68/16 68/25 Mr. Park [2] 68/16 68/25 Mr. Park's [1] 67/23 Mr. Strangia [4] 39/12 39/17 42/22 50/13 Mr. Strangia's [5] 39/7 40/8 41/3 49/7 57/11 Mrs. [1] 60/3 Mrs. Wright [1] 60/3 Ms [3] 12/4 16/7 16/12 Ms. [15] 6/5 6/25 16/6 16/7 16/10 39/1 39/2 39/18 40/9 41/20 42/22 49/6 50/12 57/10 60/14 Ms. Gina [4] 6/5 6/25 16/6 16/7 Ms. Wright [4] 41/20 42/22 50/12 60/14 Ms. Wright's [7] 16/10 39/1 39/2 39/18 40/9 49/6 57/10 much [1] 58/8 multiple [1] 65/6 municipalities [5] 63/12 63/21 63/23 64/2 64/9 museums [1] 33/16 my [113] myself [3] 15/6 19/5 44/1

81 N 26/6 26/8 26/12 28/16 15/14 15/15 25/16 54/11 15/12 15/13 15/19 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 81 of 166 N.W [3] 1/20 2/5 2/14 NAACP [2] 1/3 2/2 name [7] 4/13 4/16 6/8 69/16 69/18 70/16 70/17 National [1] 8/1 nature [4] 8/6 8/25 9/18 37/18 necessary [4] 49/12 56/24 57/2 65/9 need [4] 5/4 56/9 65/11 67/3 needs [2] 41/5 65/8 new [3] 2/5 7/6 7/10 next [2] 21/16 58/23 nice [1] 34/12 night [1] 5/10 no [20] 1/4 3/10 7/18 9/24 18/21 19/18 22/21 23/2 31/18 31/18 33/10 36/8 37/5 39/19 40/12 40/12 51/13 69/13 70/22 70/22 non [39] 17/19 18/13 18/17 18/22 19/13 20/15 22/1 22/8 22/12 23/4 24/2 24/8 24/13 24/18 24/21 24/25 25/4 25/8 26/10 26/11 27/2 28/12 32/4 32/24 33/3 36/1 36/12 36/14 44/13 44/15 44/18 44/22 62/7 62/14 66/7 66/15 66/20 67/3 67/8 non-african-america n [5] 18/22 19/13 20/15 44/18 44/22 non-black [18] 17/19 18/13 18/17 22/1 22/8 24/2 24/18 24/21 26/10 27/2 28/12 36/1 36/12 36/14 66/7 66/15 66/20 67/8 non-blacks [8] 24/8 24/25 25/4 25/8 26/11 62/7 62/14 67/3 non-bvap [2] 44/13 44/15 non-hispanic [5] 22/12 23/4 24/13 32/4 32/24 non-whites [1] 33/3 none [2] 27/4 27/11 nor [1] 73/12 normal [4] 12/10 12/13 14/22 14/23 normally [1] 68/9 North [6] 31/4 31/8 31/11 35/2 35/4 35/8 northern [2] 1/1 33/11 not [69] 5/6 7/11 9/7 9/15 9/17 10/1 10/6 10/7 10/8 11/3 11/12 11/17 13/1 14/20 14/21 15/4 15/22 18/7 18/9 19/22 20/16 22/21 23/2 26/1 26/2 29/19 31/19 35/3 36/15 37/5 37/10 37/21 37/24 40/10 40/19 46/15 46/22 47/24 48/4 49/21 52/20 54/5 54/7 54/9 54/14 54/20 57/14 59/3 59/14 62/23 64/14 64/17 68/14 68/14 68/21 68/21 68/22 69/18 69/18 70/16 70/16 72/8 72/14 73/10 73/11 Notary [2] 74/23 74/25 note [4] 63/16 63/18 63/20 63/22 noted [4] 56/2 56/2 71/5 74/4 nothing [4] 20/10 47/17 56/15 56/15 noticeable [2] 24/1 24/6 noticed [2] 32/13 32/14 November [10] 18/25 19/16 20/25 21/1 21/2 22/3 22/4 52/4 52/12 52/16 November 2012 [1] 22/3 November 2014 [2] 52/12 52/16 November 2016 [3] 18/25 21/2 22/4 now [12] 6/11 9/14 24/4 25/22 42/13 46/5 47/15 50/3 54/17 54/25 56/6 68/7 number [19] 7/25 17/14 25/9 33/22 35/19 45/19 45/23 46/10 47/22 55/23 56/1 63/6 63/7 64/14 68/7 68/17 69/8 72/22 73/25 numbers [12] 12/2 20/11 20/13 21/8 21/11 21/17 21/18 21/20 32/18 47/10 51/1 53/1 numerically [1] 29/11 O O.C.G.A [1] 72/9 Objection [2] 42/3 59/8 observation [1] 55/9 obviously [14] 7/1 17/15 20/16 21/11 21/20 25/7 31/19 34/4 42/13 50/25 62/20 65/7 68/24 69/23 occurs [1] 62/2 October [2] 10/22 10/23 off [18] 6/11 9/14 10/8 11/19 11/20 26/22 30/9 34/15 34/23 38/11 38/17 48/8 48/15 67/14 69/8 off-the-record [2] 11/20 15/15 office [10] 38/21 39/13 39/19 41/23 49/8 49/21 50/2 57/2 60/16 60/19 offices [1] 72/12 official [3] 1/6 1/13 2/18 Oh [3] 25/4 36/7 69/7 okay [29] 5/22 10/10 10/12 10/16 12/6 15/13 18/3 23/6 23/10 23/16 24/20 25/22 26/17 26/21 26/23 32/2 32/7 35/12 36/9 41/5 44/16 46/21 47/3 47/9 48/25 52/18 53/20 55/6 60/7 old [1] 29/5 on [58] 2/2 2/9 2/17 8/8 8/12 11/3 11/25 12/7 12/16 14/1 14/8 15/7 15/11 15/12 15/13 15/17 15/19 16/6 16/9 16/11 17/12 18/19 20/21 23/3 23/10 25/21 27/7 30/11 30/15 31/14 33/15 36/13 38/3 38/6 38/22 38/25 39/2 39/3 41/7 42/1 44/6 48/4 48/10 51/23 52/3 54/21 57/10 57/20 58/13 58/20 60/22 61/10 64/8 66/24 67/16 69/23 70/1 74/23 once [6] 13/22 33/19 67/25 68/3 68/6 68/10 one [35] 1/20 13/19 13/21 13/22 14/9 14/13 14/15 18/14 33/19 34/21 34/24 35/12 42/5 43/5 47/24 48/10 48/10 49/9 53/14 54/9 56/8 56/10 56/11 56/22 60/5 62/6 62/7 62/12 62/13 62/13 64/16 64/19 64/25 69/9 69/22 one percent [1] 64/16 ones [1] 48/2 only [17] 5/5 15/3 15/22 16/5 21/2 28/14 28/16 49/16 51/3 57/12 64/9 65/8 65/8 65/15 65/17 67/19 68/9 open [3] 69/19 69/22 69/24 open-source [1] 69/24 operates [1] 68/9 Oppenheimer [1] opportunity [1] 18/10 opposed [2] 18/14 49/10 or [55] 9/15 10/7 13/7 14/3 14/22 15/10 17/3 22/18 22/21 22/25 24/23 24/24 25/20 25/23 26/8 26/24 26/25 27/5 27/5 27/5 27/6 27/12 27/15 27/19 28/7 30/17 33/19 34/6 34/17 35/10 36/2 36/24 39/15 40/15 41/3 43/16 48/18 49/22 51/20 51/24 52/19 53/10 54/25 59/12 60/2 60/10 63/5 63/7 64/25 65/1 66/20 66/21 68/19 73/10 73/15 ORANGE [1] 2/11 order [1] 13/24 organization [1] 2/2 orient [3] 15/6 19/5 44/1 original [14] 3/11 5/18 5/19 6/8 6/12 7/11 7/16 16/2 52/8 61/16 64/4 66/1 70/5 73/19 other [17] 5/15 8/5 18/15 18/16 25/13 25/17 25/23 26/7 26/10 27/7 33/17 41/13 52/7 53/6 53/18 62/21 65/13 others [9] 8/18 10/11 10/16 25/10 25/20 26/16 26/19 26/25 27/5 our [3] 13/9 33/25 41/22 out [12] 21/9 21/12 26/5 30/20 34/12 34/13 44/7 46/19 57/23 67/8 70/8 70/11 outcome [1] 73/13 outlined [1] 6/6 outside [12] 31/18 33/12 33/20 45/12 46/3 46/7 46/8 47/6 47/14 58/15 62/20 62/22 over [13] 6/2 7/2 16/24 17/2 17/8 17/16 44/22 61/17 61/21 62/2 63/4 70/4 70/12 own [1] 39/4 P p.m [1] 71/5 package [6] 13/6 13/7 14/10 68/16 70/6 70/14 page [40] 3/2 3/10 7/24 12/16 12/17 14/1 14/24 14/25 15/7 15/8 16/13 18/19 20/21 21/22 25/21 38/22 38/22 43/11 44/6 48/10 51/16 51/23 54/21 55/21 55/21 55/21 58/13 58/20 60/22 65/21 74/5 74/7 74/9 74/11 74/13 74/15 74/17 Page 1 [1] 7/24 Page 10 [1] 21/22 Page 13 [1] 60/22 Page 14 [2] 25/21 65/21 Page 15 [3] 51/16 54/21 55/21 Page 16 [1] 55/21 Page 17 [3] 55/21 58/13 58/20 Page 2 [2] 15/7 15/8 Page 2 of [1] 14/24 Page 24 [1] 43/11 Page 25 [1] 44/6 Page 3 [2] 14/25 15/12 Page 33 [1] 38/22 Page 34 [1] 38/22 Page 361 [1] 48/10 Page 4 [1] 12/17 Page 5 [1] 16/13 Page 6 [1] 14/1 Page 7 [2] 18/19 20/21 pages [1] 16/13 paragraph [13] 14/25 15/8 15/19 18/21 21/25 22/15 23/7 23/20 43/13 51/18 58/20 65/24 65/24 Paragraph 4 [1] 65/24 parameter [3] 62/6 62/7 69/9 parameters [3] 62/13 68/6 70/10 PARK [3] 2/19 68/16 68/25 Park's [1] 67/23 part [21] 12/10 12/13 14/21 17/19 18/1 18/2 24/17 24/21 27/1 33/1 35/22 35/24 39/23 40/6 63/24 68/8 68/19 68/20 69/2 69/3 69/25 partially [1] 22/6 particular [6] 44/4 55/25 57/12 62/8 63/4 63/5 particularly [1] 65/5 parties [2] 73/10 73/12 partisan [13] 15/4 15/22 33/1 35/21 35/23 40/14 44/4 49/9 49/13 50/9 50/14 57/14 62/17 partisanship [6] 39/14 39/20 40/11 40/13 40/16 40/20

82 P plus [5] 64/15 64/19 prepare [2] 5/8 9/1 quite [3] 29/10 32/23 referred [1] 14/11 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 82 of 166 party [1] 60/24 past [3] 10/22 11/4 36/23 PATRICIA [1] 2/3 patterns [3] 35/23 59/17 62/17 PAYTON [1] 2/10 Peachtree [2] 1/20 2/21 pending [7] 5/6 11/10 11/13 11/14 11/24 12/1 12/8 Pennsylvania [2] 10/25 10/25 people [4] 8/2 26/7 59/22 59/25 percent [52] 14/2 16/21 16/21 16/25 17/2 17/3 17/4 17/8 17/16 17/17 21/1 21/1 21/12 21/18 21/18 21/19 21/19 21/21 22/3 22/3 29/7 29/7 29/8 29/9 44/10 44/12 44/13 44/17 44/22 45/8 45/17 46/2 46/3 46/10 47/6 47/7 61/12 61/18 61/21 62/2 62/22 62/22 63/5 63/5 63/7 64/16 64/25 65/1 66/2 66/6 66/17 66/18 percentage [5] 24/8 32/14 33/3 42/2 46/17 perfectly [1] 40/18 performance [3] 48/21 49/9 49/13 performing [1] 49/2 performs [1] 48/13 PERKINS [1] 2/13 perkinscoie.com [1] 2/16 person [2] 66/25 67/6 perspective [1] 49/20 Ph.D [6] 1/16 4/8 4/18 72/4 74/2 74/21 phone [1] 26/22 photocopying [1] 73/16 physically [1] 68/9 pick [1] 34/1 pieces [1] 24/10 place [5] 6/20 28/14 28/16 44/3 58/6 Plaintiff's [3] 9/21 9/22 11/1 Plaintiffs [10] 1/4 1/11 2/2 2/9 8/16 8/19 8/22 9/13 10/13 10/20 Plaintiffs' [1] 5/10 plan [12] 11/25 15/3 15/22 18/23 19/14 45/13 46/1 63/11 63/13 63/14 65/5 65/7 Plan's [2] 21/3 29/6 plans [1] 64/13 please [4] 4/3 4/13 4/25 20/4 64/19 64/25 65/1 plus/minus [3] 64/15 64/19 64/19 PM [2] 67/14 67/16 point [3] 29/10 56/23 57/21 polarized [3] 52/6 56/3 69/10 political [15] 22/19 23/11 23/21 24/15 30/16 31/20 31/21 31/24 38/2 38/5 38/6 38/13 38/19 41/21 56/5 population [16] 20/22 21/14 22/8 39/10 39/14 40/5 41/7 42/2 44/13 45/6 45/6 45/7 46/19 64/13 65/9 65/14 portion [19] 12/4 24/2 28/12 36/2 36/3 41/2 44/6 44/9 44/16 44/23 44/24 47/6 48/18 48/22 49/1 49/11 59/15 60/11 61/15 portions [9] 39/25 46/6 47/11 47/14 49/14 49/17 54/22 55/11 60/2 possible [8] 19/24 20/20 49/22 50/5 50/8 50/11 62/21 65/13 possibly [1] 33/10 practice [2] 14/22 14/23 precinct [44] 15/4 15/23 38/21 39/5 39/15 40/11 40/15 40/17 40/19 40/21 41/5 41/24 43/3 43/14 43/22 43/23 44/4 44/17 44/25 48/19 49/9 49/10 49/11 49/14 49/17 50/14 55/11 56/11 56/14 56/21 56/24 56/24 57/3 57/3 57/8 57/12 57/16 58/7 58/8 59/4 59/4 60/13 60/21 62/17 precincts [39] 39/9 39/21 40/3 45/13 46/7 47/5 47/11 47/14 47/16 47/18 47/21 47/24 48/2 48/4 48/5 48/10 48/11 54/23 55/24 56/7 56/8 56/9 56/16 56/18 57/1 57/5 57/9 57/18 57/21 57/22 58/16 58/25 59/10 59/21 60/1 60/3 60/5 62/8 64/10 precise [5] 6/7 6/20 17/14 41/17 42/10 precisely [4] 18/11 51/5 55/25 70/8 prediction [1] 22/1 prepared [2] 5/14 9/4 preparing [1] 5/22 present [1] 37/24 presented [1] 53/1 pretty [5] 29/2 33/23 47/22 48/3 48/14 previous [1] 19/17 primary [4] 15/2 15/21 16/1 61/2 probably [1] 68/19 probative [1] 53/13 procedure [2] 12/22 13/3 proceeding [3] 72/13 72/14 73/4 proceedings [1] 73/16 process [9] 11/8 12/11 12/13 16/8 39/17 40/1 42/15 42/22 43/1 produce [3] 61/20 62/18 63/2 producing [1] 62/15 professor [3] 13/9 69/20 69/24 program [1] 13/8 programming [2] 13/9 69/21 prohibited [1] 72/15 properly [4] 28/5 56/23 57/22 60/9 proportion [6] 45/7 45/22 46/2 46/4 46/9 46/25 proportions [1] 42/2 Proscenium [1] 2/20 provide [1] 72/12 provisions [1] 72/9 Public [3] 54/12 74/23 74/25 publication [1] 38/17 publications [3] 38/3 38/8 38/9 purely [2] 23/10 30/15 purpose [1] 18/3 purposes [1] 47/17 Pursuant [1] 72/5 put [14] 13/18 24/10 32/20 41/9 42/16 42/23 47/16 48/17 48/21 58/6 58/11 60/1 60/14 60/20 Q question [29] 4/25 5/6 11/24 13/17 17/25 19/20 20/6 20/11 22/22 26/23 28/3 29/21 32/8 32/10 35/3 42/5 45/10 47/7 48/23 49/5 50/15 50/24 51/3 52/14 53/11 60/9 63/17 66/23 67/12 questions [6] 20/19 67/18 67/20 69/13 70/21 73/6 quickly [1] 68/11 38/7 R race [6] 45/12 45/16 45/20 52/19 54/17 59/20 races [1] 53/6 racial [23] 15/3 15/22 18/5 18/8 18/11 26/4 47/15 47/21 48/1 49/18 50/8 50/12 55/10 56/4 57/4 57/13 58/14 58/22 58/24 59/3 59/10 62/16 63/4 racially [5] 52/6 56/3 59/6 59/7 69/10 racially-polarized [3] 52/6 56/3 69/10 Raleigh [3] 8/20 9/19 31/4 ran [1] 54/4 rate [2] 46/20 52/21 rather [4] 56/10 56/24 57/8 57/17 read [13] 21/9 21/12 38/7 39/23 71/4 74/2 74/5 74/7 74/9 74/11 74/13 74/15 74/17 readily [1] 13/10 reading [6] 12/4 16/12 23/8 41/22 49/6 58/17 really [8] 19/8 26/13 34/12 34/14 34/22 40/20 59/2 68/11 reapportionment [10] 6/7 38/20 39/19 49/8 49/21 50/2 56/17 57/2 60/16 60/19 reason [10] 48/17 48/21 56/9 74/6 74/8 74/10 74/12 74/14 74/16 74/18 reasonable [1] 65/14 reasons [4] 55/24 57/16 57/19 57/23 rebuttal [9] 5/13 7/3 7/5 7/9 8/9 8/10 8/12 54/21 55/2 recall [20] 6/15 9/15 11/23 12/4 17/9 17/14 33/24 34/20 34/23 39/7 39/10 39/12 41/2 41/8 41/12 41/15 41/16 41/20 42/8 61/5 recollection [1] 64/2 record [9] 4/14 11/20 15/15 15/17 30/9 30/11 67/14 67/16 73/8 redistricting [4] 31/15 31/22 56/21 64/8 reduced [1] 73/6 refer [2] 13/23 44/23 reference [2] 23/24 27/19 referenced [3] 6/8 6/18 6/21 referring [12] 11/17 16/3 16/3 22/17 26/15 35/9 36/4 40/23 45/2 58/19 58/20 58/21 reflected [6] 27/5 27/9 29/2 29/25 36/1 36/3 reflection [1] 27/6 reflects [1] 11/10 regard [1] 53/9 regarding [4] 16/8 19/22 62/13 62/13 registration [4] 5/25 18/6 26/2 59/12 regression [27] 16/15 16/19 17/1 17/7 17/9 17/15 22/19 27/13 27/17 27/19 28/8 28/10 36/2 36/25 37/8 37/20 61/10 61/20 62/5 62/5 62/10 62/11 62/14 62/19 62/25 63/2 63/10 regular [2] 11/8 73/11 Regulations [1] 72/5 REID [1] 2/3 reiterate [1] 12/9 relationship [2] 39/6 72/8 relative [5] 39/20 40/10 40/13 49/9 49/13 relatively [1] 52/21 relatively-high [1] 52/21 relevant [3] 55/14 55/23 57/7 relied [1] 12/7 rely [2] 23/3 30/13 relying [6] 23/10 30/15 38/3 38/4 38/25 39/2 remember [5] 6/7 6/11 9/14 38/9 42/13 remembered [1] 7/5 René [2] 8/14 9/1 repeat [2] 45/9 52/14 rephrase [1] 5/1 reply [7] 3/12 7/9 51/6 51/8 58/13 60/22 65/21 report [107] reported [19] 17/17 19/7 20/11 28/20 32/19 33/2 43/24 44/3 45/21 45/24 46/2 46/9 46/10 46/16 46/23 51/2 53/3 68/22 68/25 reporter [5] 4/5 72/7 72/21 73/18 73/24 reporting [8] 1/23 26/3 47/25 72/5 72/10 72/11 72/12 72/14 reports [6] 5/12 7/1 9/5 54/9 59/13 70/13 representation [1] 49/1 representative [1]

83 R row [2] 17/12 45/24 22/16 22/16 22/19 17/5 18/8 18/12 20/18 56/6 56/8 56/9 56/11 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 83 of 166 representative... [1] 72/10 represents [5] 14/9 14/13 25/24 26/7 28/21 Republican [9] 43/20 43/22 43/23 48/14 48/18 48/19 48/21 51/21 66/7 Republicans [2] 49/2 61/7 require [1] 56/18 research [6] 11/8 12/11 12/13 14/22 14/23 70/1 reserving [1] 71/3 reside [1] 21/2 residency [1] 59/19 resolved [2] 9/17 9/18 respect [7] 19/19 19/20 43/4 48/6 50/25 53/11 56/6 respondent [1] 25/13 response [3] 13/13 19/18 57/24 responsive [1] 32/10 rest [1] 45/1 result [3] 16/24 61/11 62/19 results [16] 5/15 7/6 7/8 13/14 17/11 19/6 19/12 27/23 33/1 36/23 41/4 54/7 58/11 61/17 63/3 68/24 review [4] 20/2 39/25 43/10 64/5 reviewed [14] 5/11 5/12 5/23 5/24 5/25 6/4 6/9 6/16 6/18 6/24 6/25 7/1 13/14 16/5 Reviews [6] 19/5 20/7 23/14 43/10 44/2 64/7 revisit [2] 30/19 32/9 right [54] 4/18 6/11 9/14 10/1 10/4 13/20 15/5 16/17 17/20 18/1 21/4 21/15 25/13 25/15 25/25 26/14 26/19 27/7 29/3 29/15 29/18 30/24 32/5 37/3 37/22 40/6 42/12 43/6 45/8 45/11 45/13 45/18 46/5 47/1 47/7 47/13 52/23 54/16 58/18 59/7 59/19 59/22 60/3 60/12 60/20 61/8 61/13 61/15 61/21 63/12 66/15 67/4 70/23 71/2 RIGHTS [1] 2/5 Rob [1] 6/22 Robert [1] 10/17 Romo [9] 8/14 9/1 9/2 9/4 30/24 37/12 37/13 37/21 37/22 round [2] 17/3 63/6 rounded [1] 61/12 Rucho [9] 10/17 10/18 10/19 31/11 35/11 36/6 36/7 37/7 37/11 Rules [1] 72/5 run [17] 13/19 13/22 14/3 14/9 14/13 14/15 14/20 14/21 27/20 36/16 67/24 68/2 68/6 68/9 69/4 69/5 69/5 runn [1] 69/21 runs [7] 13/15 14/4 14/19 22/19 36/13 68/12 69/3 rural [2] 43/8 43/16 S SABRINA [1] 2/10 said [14] 30/17 32/11 34/4 41/3 41/10 41/13 41/16 52/9 52/22 52/23 57/20 67/24 73/9 73/12 said whether [1] 41/13 same [19] 19/9 21/13 32/8 32/23 35/1 38/16 39/24 40/18 41/18 42/9 42/11 49/11 49/14 52/6 54/13 54/15 54/19 61/15 74/3 SAMMY [1] 2/11 saw [4] 23/23 32/14 32/18 59/10 say [27] 12/20 13/25 14/12 17/6 18/21 20/8 23/3 26/23 26/24 27/3 29/9 30/13 33/18 43/13 50/3 51/19 52/3 53/9 55/9 58/13 59/6 59/7 59/15 63/3 63/4 64/2 65/25 saying [12] 11/25 13/21 32/3 34/25 38/12 38/18 39/8 39/24 52/13 52/15 55/19 60/18 says [1] 15/20 sbllaw.net [1] 2/22 scenario [1] 67/9 school [5] 8/2 33/25 33/25 34/21 34/22 scientist [3] 31/20 31/21 38/5 scientist's [1] 14/22 scrutinize [2] 56/12 56/13 seal [1] 73/19 second [7] 6/13 6/16 6/19 11/19 43/25 58/20 64/5 Secretary [6] 1/6 1/13 2/18 41/23 58/4 58/10 section [3] 20/2 39/3 64/5 see [25] 14/8 15/24 19/1 21/5 22/5 22/10 22/25 23/25 24/6 24/19 34/11 43/24 44/11 44/20 48/12 52/2 52/9 58/19 58/21 61/1 66/5 66/8 seemed [1] 39/24 segregated [2] 59/6 59/7 segregation [5] 58/22 58/23 58/24 58/25 59/4 select [1] 12/2 selecting [1] 12/3 Senate [3] 9/12 53/9 54/4 sense [2] 39/22 40/3 sentence [19] 15/1 15/9 15/20 15/24 16/4 16/6 19/1 21/16 22/10 22/13 22/16 22/17 23/7 23/20 23/25 24/12 32/3 44/11 58/23 Service [1] 54/12 services [5] 1/23 72/10 72/11 72/12 72/14 set [2] 13/7 56/25 settled [2] 9/16 9/16 seven [1] 15/10 several [2] 68/19 69/6 share [1] 60/24 shares [2] 22/8 24/1 she [2] 41/21 60/15 SHEET [1] 74/1 should [8] 7/4 74/5 74/7 74/9 74/11 74/13 74/15 74/17 shouldn't [1] 65/20 show [5] 19/8 19/21 20/13 29/2 29/4 showed [1] 67/1 shows [4] 12/2 20/24 45/5 66/25 side [2] 8/4 8/5 sign [1] 71/4 signature [2] 71/3 73/18 significant [4] 29/10 35/19 50/22 56/5 silent [1] 64/8 similar [1] 41/13 simply [8] 20/11 40/15 40/21 57/10 62/15 63/6 63/14 68/17 SIMS [1] 2/3 since [1] 33/16 single [1] 14/3 sir [80] slight [1] 33/3 slightly [1] 46/20 small [2] 29/2 56/4 SMITH [1] 2/3 SNOW [1] 2/11 so [68] 4/22 5/2 6/3 6/10 6/24 7/9 8/12 12/13 13/7 15/10 16/6 23/14 23/17 24/10 26/3 26/6 27/2 27/23 29/9 29/21 30/5 30/19 32/7 32/8 32/11 33/12 33/14 33/20 33/21 34/2 34/13 34/14 35/12 35/23 37/10 41/5 43/25 46/10 47/3 48/3 49/20 50/3 50/11 51/2 52/10 54/17 56/20 56/22 56/23 57/15 57/19 62/1 62/6 62/18 63/1 63/8 64/4 65/19 66/18 68/17 68/19 69/4 69/16 70/11 social [2] 9/23 14/22 software [27] 11/6 13/6 13/19 14/3 14/4 14/9 14/10 14/13 14/15 14/20 39/8 40/16 68/4 68/5 68/7 68/9 68/10 68/15 68/20 69/1 69/5 69/9 69/17 69/19 69/19 69/20 70/7 some [10] 11/15 25/9 32/8 32/19 34/1 41/15 43/3 44/13 55/17 67/3 somehow [3] 9/16 9/17 60/10 something [7] 27/3 30/17 47/23 47/25 56/17 56/19 62/23 sometime [1] 10/20 sometimes [1] 62/19 somewhat [2] 18/23 19/15 sorry [4] 7/9 9/22 55/6 65/17 sort [1] 69/10 sought [4] 51/19 52/1 56/12 56/13 source [3] 69/19 69/22 69/24 South [7] 23/11 23/21 24/16 30/17 30/21 30/23 38/2 Southern [1] 9/22 speak [2] 38/1 60/17 Special [1] 7/7 specific [2] 35/8 51/1 specifically [16] 19/11 22/22 23/23 26/18 29/4 29/19 30/2 30/5 36/10 39/16 44/3 46/25 50/25 56/6 56/20 65/10 specifies [1] 25/17 spell [1] 4/16 split [59] 39/9 39/15 39/21 40/2 40/6 40/11 40/15 40/17 40/19 40/21 45/13 46/7 47/5 47/11 47/11 47/14 47/19 47/20 48/2 48/5 48/11 48/22 49/9 49/10 49/14 49/17 54/22 55/10 55/24 56/14 56/16 56/22 57/1 57/5 57/8 57/9 57/12 57/17 57/18 57/20 57/22 58/7 60/13 60/21 63/21 63/22 64/18 64/21 64/24 65/6 65/8 65/8 65/12 65/17 65/18 splits [1] 39/10 splitting [6] 50/14 56/18 63/12 64/10 65/11 65/16 splitting of [1] 65/11 spring [1] 10/21 St [1] 8/2 stadium [1] 34/12 standard [6] 68/5 68/5 68/6 69/3 69/9 70/10 stands [1] 70/18 start [2] 5/22 55/19 starts [5] 15/1 15/9 15/9 15/19 51/18 state [19] 1/7 1/7 1/14 1/14 2/2 2/17 2/18 2/18 4/13 22/6 33/6 33/22 34/3 34/5 58/4 58/10 66/3 72/2 73/2 State's [1] 41/23 stated [6] 21/17 23/24 24/11 39/12 64/4 73/5 statement [2] 23/19 59/16 states [4] 1/1 30/23 38/6 41/13 statistical [2] 62/3 63/9 status [1] 66/21 Stephen [1] 54/11 still [2] 66/19 67/2 straight [3] 62/15 62/16 62/18 Strangia [5] 6/22 39/12 39/17 42/22 50/13 Strangia's [6] 6/24 39/7 40/8 41/3 49/7 57/11 Street [3] 1/20 2/14 2/21 STRICKLAND [3] 2/20 48/7 72/12 strong [2] 48/20 56/2 strongly [2] 48/14 49/2 struck [1] 39/21 students [1] 34/1 studied [2] 30/2 38/18 study [1] 29/19 sub [3] 15/4 15/23 38/21 sub-precinct [3] 15/4 15/23 38/21 subscribed [1] 74/22 substantial [1] 68/7 substantially [1]

84 S 60/23 64/9 25/20 28/4 33/2 34/13 thousand [2] 68/19 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 84 of 166 substantially... [1] 52/6 substitute [1] 19/19 such [3] 20/16 65/15 73/17 sufficient [1] 66/10 suggested [1] 53/4 suggesting [3] 49/22 50/6 50/15 Suite [3] 2/6 2/14 2/20 supermarkets [1] 33/16 supplemental [1] 11/3 support [14] 8/4 8/5 20/15 22/2 22/12 24/14 24/21 32/5 33/4 51/20 53/1 53/8 66/3 66/7 supported [5] 52/16 52/20 53/5 53/15 66/11 supporters [1] 52/20 supporting [1] 24/9 sure [33] 6/17 9/4 9/17 10/18 11/1 11/12 11/17 14/6 20/1 20/5 20/8 23/10 23/13 33/23 40/7 40/24 44/21 45/10 45/23 46/15 46/23 47/9 53/24 54/15 55/16 59/24 60/8 62/1 64/6 64/14 64/17 66/22 69/18 sure I [1] 45/10 SWANSON [1] 2/10 swear [1] 4/3 swore [1] 4/5 sworn [2] 4/9 74/22 T table [45] 13/25 14/9 16/21 17/11 19/3 19/7 19/8 19/10 19/12 19/20 20/12 20/19 20/23 22/9 23/24 23/25 24/6 24/17 24/19 24/20 24/25 25/19 25/22 26/4 26/9 27/7 27/10 28/12 28/21 28/21 28/24 29/11 30/3 30/3 32/12 32/12 32/14 32/19 33/2 36/11 45/3 45/15 51/2 60/23 61/15 tables [7] 16/14 17/18 17/21 17/23 18/4 30/1 30/6 take [16] 5/4 5/6 12/5 20/4 25/15 25/17 29/24 30/7 35/12 44/5 47/20 51/6 54/18 60/5 67/10 67/13 taken [1] 73/5 taking [1] 72/14 talk [8] 20/21 30/21 38/20 43/3 44/6 58/22 talking [9] 21/25 23/20 42/18 42/20 53/14 54/24 59/21 59/24 61/14 talks [1] 54/22 TCB [1] 1/5 TCB-WSD-BBM [1] 1/5 telephone [1] 2/13 tell [16] 7/14 10/8 12/6 25/16 33/9 34/15 34/22 37/15 37/16 38/11 38/16 48/8 48/15 54/17 59/2 69/7 telling [3] 47/23 49/4 67/7 tells [1] 47/22 ten [1] 15/10 Tennessee [1] 33/12 term [1] 61/3 terms [1] 43/19 testified [10] 4/10 9/25 10/2 10/15 10/23 11/4 39/11 40/12 41/20 49/12 testify [2] 9/6 9/7 testifying [1] 42/8 testimony [7] 16/10 39/18 41/3 49/25 50/1 57/11 74/3 text [3] 20/9 20/10 20/24 than [25] 5/15 18/17 22/12 23/4 24/13 27/7 29/18 29/18 30/18 32/4 44/17 44/25 46/12 46/15 56/10 56/25 57/8 57/17 58/15 59/16 61/12 64/19 64/25 66/16 66/18 thank [5] 4/18 55/6 70/19 70/20 70/22 that [391] that's [31] 6/23 8/12 10/1 13/1 13/22 18/18 25/15 28/14 28/16 32/15 33/16 35/3 40/19 40/21 40/23 43/24 45/1 48/23 49/3 53/7 53/16 53/17 59/21 61/9 61/14 63/8 63/8 63/9 67/17 69/22 70/17 their [8] 18/11 39/23 40/1 40/16 42/14 49/15 64/12 66/21 them [9] 27/1 27/11 40/12 41/16 41/25 42/13 56/24 57/13 66/20 themselves [1] 46/8 then [12] 10/3 10/24 14/20 15/8 17/3 23/3 32/18 36/18 47/12 48/7 63/6 66/18 there [38] 5/22 6/5 6/13 6/17 9/24 16/14 20/14 23/25 24/6 25/8 34/13 34/21 34/22 35/19 38/16 41/15 46/11 46/14 46/16 46/22 46/23 49/1 50/16 50/18 50/21 53/25 54/13 56/1 56/2 56/14 56/15 59/23 61/6 66/23 there's [5] 15/1 20/9 33/10 50/22 69/18 therefore [1] 40/15 thereto [1] 73/19 these [20] 6/1 14/12 24/12 26/5 30/5 32/25 42/14 47/5 47/10 47/14 47/18 47/20 48/1 55/10 56/8 58/24 59/20 60/1 60/2 64/13 they [25] 4/22 9/23 21/6 21/7 26/14 27/8 27/11 28/5 29/4 34/12 35/1 36/1 36/3 39/22 39/24 40/2 40/4 40/10 40/12 40/12 40/15 49/11 49/12 49/16 64/15 they're [4] 23/4 25/7 42/18 42/20 thing [2] 32/13 39/25 things [5] 4/24 7/4 32/13 32/20 66/23 think [22] 10/21 25/15 28/17 28/18 28/20 32/7 32/9 42/25 44/20 44/23 45/1 46/24 49/3 49/22 50/10 50/15 53/7 54/3 63/15 64/3 64/18 66/23 this [44] 5/12 7/16 7/17 7/24 10/22 11/4 12/10 13/12 16/2 16/6 20/2 20/23 22/6 26/1 26/2 26/2 26/6 28/15 29/10 31/18 32/3 32/9 34/5 36/5 39/3 44/16 44/23 44/24 45/5 51/7 55/25 56/7 59/14 61/24 62/4 62/9 67/18 68/25 69/10 72/12 72/14 73/14 73/20 74/23 THOMPSON [2] 1/10 2/9 those [43] 6/3 6/10 6/15 7/8 7/8 7/10 17/22 18/7 20/13 21/8 21/20 24/10 27/3 27/3 27/9 29/9 29/11 32/20 35/12 35/22 38/7 39/6 39/25 40/25 43/19 47/10 47/16 51/1 53/12 54/17 56/9 56/22 57/4 57/9 57/16 57/18 57/19 57/20 57/22 57/23 59/10 60/5 69/2 though [2] 37/17 49/5 69/6 thousands [1] 33/24 three [10] 16/20 47/10 47/11 47/14 47/16 47/18 47/20 47/23 48/1 48/3 threshold [1] 65/14 through [4] 13/19 14/3 14/4 69/5 throughout [1] 40/18 throwing [1] 68/24 time [17] 5/5 16/5 20/4 32/8 32/19 34/21 34/24 35/12 42/6 47/25 56/8 56/10 56/22 60/6 67/18 69/4 71/5 times [7] 33/5 33/6 33/22 34/6 34/8 34/13 34/17 together [5] 24/11 32/20 47/16 56/25 58/11 told [2] 27/18 49/3 too [2] 6/8 33/9 top [11] 6/11 9/15 10/8 25/16 34/15 34/23 38/11 38/17 48/8 48/15 69/8 topic [3] 38/13 38/18 41/8 topics [1] 42/14 total [3] 45/20 47/4 48/1 totals [1] 45/16 toward [3] 15/1 15/9 65/25 Trace [1] 1/24 trade [2] 69/16 69/18 trademarked [1] 69/19 training [1] 38/5 transcript [12] 6/4 6/18 16/5 16/7 41/9 41/11 41/18 42/11 42/16 42/23 73/8 74/2 traveled [1] 34/2 treatment [1] 38/21 trial [10] 9/7 9/17 9/25 10/2 10/8 10/14 10/15 10/22 10/23 11/4 trip [1] 34/22 trips [1] 33/15 true [5] 7/21 21/13 51/10 73/8 74/3 try [6] 22/23 22/24 27/25 28/2 28/6 51/22 trying [2] 28/20 46/24 turn [10] 12/17 14/24 16/13 21/22 25/19 26/22 45/3 46/19 51/16 67/8 turned [5] 6/2 7/2 26/5 70/4 70/12 turning [3] 7/24 14/2 24/17 turnout [15] 6/1 18/6 20/25 25/21 26/3 29/2 29/3 30/4 45/17 46/11 46/14 46/16 46/17 46/22 46/24 Tussahaw [8] 48/11 48/13 48/17 48/19 48/22 49/1 49/10 49/17 Tussahaw's [1] 48/20 two [19] 5/12 24/10 32/20 33/19 40/25 49/13 52/7 58/16 58/24 59/20 60/1 60/2 60/24 61/7 61/7 62/12 65/10 65/16 65/17 two-party [1] 60/24 typewriting [1] 73/7 typically [2] 65/4 68/14 TYSON [1] 2/3 U U.S [2] 53/9 54/4 under [9] 2/5 18/23 19/14 21/12 21/21 63/5 72/9 72/14 73/7 undersigned [2] 73/17 74/23 understand [13] 4/25 14/14 15/25 28/3 40/1 40/8 42/17 42/19 44/20 52/10 60/8 66/22 67/8 understanding [5] 16/9 38/25 39/17 50/1 55/24 uniform [1] 40/18 uniformity [1] 40/22 UNITED [3] 1/1 30/23 38/6 units [1] 39/6 Unknown [3] 25/20 25/23 27/5 Unknowns [2] 26/16 26/25 unless [1] 73/17 up [16] 13/24 33/10 33/19 33/20 34/1 34/4 43/25 47/10 47/13 47/21 60/10 60/11 65/12 65/16 66/25 67/1 up that [1] 65/16 uploaded [1] 70/9 upon [3] 38/4 73/9 73/15 urban [1] 43/17 us [1] 27/18 use [17] 7/11 11/8 11/25 12/10 12/10 12/12 12/22 22/21 32/1 36/24 37/5 41/13 50/8 50/12 69/10 70/7 70/7 used [14] 6/11 11/7 12/7 12/20 13/6 13/11 18/4 28/22 40/4 41/17 64/12 68/21 70/9 70/10

85 U wasn't [1] 56/17 which [15] 5/17 8/3 42/19 42/21 44/13 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 85 of 166 user's [1] 11/10 using [9] 13/3 20/18 22/25 28/7 39/13 59/11 59/11 59/11 63/2 V VAP [2] 45/12 47/4 Vargas [7] 1/22 1/23 72/10 72/11 72/14 72/21 73/24 various [8] 5/25 26/5 32/22 37/1 40/14 54/17 57/19 69/25 ve [1] 36/11 version [2] 50/5 55/22 versus [2] 18/12 27/2 very [3] 12/12 34/4 62/20 via [1] 2/13 vote [8] 23/4 32/23 58/7 58/8 60/24 66/19 66/20 67/1 voted [6] 26/14 26/25 52/11 53/25 66/19 67/2 voter [8] 5/25 5/25 18/6 24/21 26/2 26/3 59/12 67/10 voters [53] 8/17 9/10 9/11 9/22 10/24 18/5 18/9 18/22 19/13 21/2 22/11 22/12 22/20 23/1 24/13 24/13 24/19 24/24 25/5 25/9 27/2 27/20 27/20 27/22 27/24 28/17 30/15 31/1 32/4 32/4 32/24 35/1 35/6 35/17 35/20 35/22 35/24 37/9 37/13 38/10 38/15 39/9 51/20 52/11 52/15 53/5 53/8 53/25 66/3 66/11 66/14 66/15 67/8 votes [5] 58/2 58/12 66/25 67/7 67/11 voting [24] 26/19 27/4 28/6 32/22 35/16 35/23 36/18 36/20 37/1 38/10 39/9 39/14 40/4 41/7 45/6 52/7 56/3 61/20 61/25 62/17 63/4 66/9 66/20 69/10 W Wake [4] 8/20 8/21 9/19 31/4 want [11] 13/23 19/24 23/16 26/1 28/3 40/7 45/10 51/22 60/8 66/22 67/8 wanted [4] 20/7 20/18 39/25 40/24 was [117] Washington [2] 2/6 2/15 way [8] 26/13 33/10 35/1 39/11 39/19 40/12 40/13 73/12 WAYNE [1] 2/9 we [24] 11/19 17/1 17/7 18/8 23/7 23/19 24/5 26/13 26/23 27/2 33/14 33/17 38/7 41/22 41/24 54/24 57/25 58/2 58/2 60/5 61/15 63/6 63/6 71/1 we'll [2] 5/5 5/22 we're [4] 45/11 55/24 59/24 70/23 we've [1] 53/13 well [30] 5/12 6/9 7/2 10/23 11/2 13/12 13/18 13/24 18/2 21/11 25/19 26/9 27/16 27/18 29/23 30/2 31/23 33/9 34/11 39/7 40/7 40/9 41/2 41/12 50/24 53/11 67/10 68/4 70/6 70/9 went [6] 10/7 10/13 10/22 64/15 64/19 64/19 were [37] 6/2 6/13 7/10 9/8 19/7 21/7 21/17 21/18 22/11 23/7 23/19 24/13 28/4 28/5 30/3 32/4 33/2 33/17 35/4 35/5 35/24 36/1 36/2 36/12 40/10 47/11 47/18 48/2 53/2 53/12 56/1 57/1 61/4 61/6 62/5 70/10 73/6 weren't [1] 21/6 what [90] what's [2] 7/13 37/16 when [33] 5/14 11/25 12/6 13/14 13/18 16/2 17/7 17/21 20/21 30/13 30/21 33/19 33/24 34/20 34/21 35/4 38/1 38/20 41/22 41/24 47/20 54/4 55/24 57/25 58/4 61/15 62/4 62/10 62/18 63/1 65/9 66/14 68/2 whenever [2] 34/14 56/21 where [20] 6/20 15/6 28/16 31/25 33/14 33/16 41/3 44/3 45/24 45/25 48/7 51/22 52/2 52/9 53/15 53/25 58/19 58/21 59/22 59/25 whether [25] 10/7 11/10 11/12 17/25 26/23 26/24 28/20 35/16 39/10 41/12 41/13 41/13 43/16 43/20 43/21 46/22 48/13 50/16 50/18 50/22 51/20 52/19 53/9 63/19 64/15 12/1 20/24 24/5 32/22 33/2 35/8 38/5 42/9 48/4 65/6 65/9 66/25 67/6 while [1] 5/6 White [16] 22/12 24/13 25/2 25/5 25/9 29/18 32/4 50/16 50/18 50/23 52/11 52/17 53/10 54/1 54/19 66/20 White-Black [1] 66/20 Whites [3] 23/5 32/24 33/3 who [13] 15/25 18/13 18/16 21/2 26/5 26/7 26/12 26/15 26/18 40/5 59/22 66/19 67/10 whole [2] 29/21 58/8 why [16] 23/23 46/11 46/14 46/19 46/22 47/4 47/6 55/13 55/14 55/15 55/16 55/20 56/9 57/19 59/7 61/19 will [8] 5/1 6/1 55/19 62/7 62/19 65/5 68/15 72/14 win [1] 67/3 winner [2] 57/25 58/4 withdraw [1] 67/12 withdrawn [1] 73/14 within [26] 18/22 19/13 21/2 29/5 29/5 32/16 32/17 35/20 39/15 39/15 39/21 40/2 40/11 40/14 40/17 40/19 40/21 45/12 45/17 46/12 46/15 46/20 47/4 47/12 57/12 62/8 without [3] 46/12 46/15 53/9 witness [7] 4/4 4/6 31/17 31/19 31/22 72/4 74/2 Women [6] 8/17 9/10 9/11 9/21 10/24 31/1 word [6] 12/5 25/15 25/18 29/24 44/5 54/18 words [4] 41/17 42/10 62/21 65/13 work [9] 4/22 8/8 8/25 11/25 31/13 31/19 35/4 35/8 35/13 worked [4] 8/12 10/3 11/5 31/14 works [1] 63/10 world [3] 64/25 66/24 67/6 would [41] 4/3 4/13 4/25 12/9 12/16 12/17 13/19 13/22 14/21 17/16 23/12 24/18 24/22 26/9 30/20 30/20 33/18 34/2 38/7 39/13 41/10 42/17 46/19 48/17 48/21 49/7 49/17 49/21 50/11 51/16 54/3 58/11 60/19 64/21 69/9 70/3 70/24 71/1 would've [3] 33/15 33/18 49/16 wouldn't [2] 17/6 50/3 Wright [9] 6/5 11/24 16/6 16/7 41/20 42/22 50/12 60/3 60/14 Wright's [11] 6/25 12/5 16/7 16/10 16/12 39/1 39/2 39/18 40/9 49/6 57/10 write [1] 11/2 written [1] 5/13 wrote [5] 9/13 9/23 10/5 10/19 16/2 WSD [1] 1/5 Y Yeah [1] 70/15 years [1] 33/20 yes [83] yesterday [2] 5/11 5/11 York [1] 2/5 you [241] you'd [4] 41/9 42/16 42/23 65/21 you'll [9] 6/19 19/25 23/14 25/19 28/24 43/9 43/11 43/25 64/4 you're [70] 4/18 11/17 12/3 12/11 13/3 13/21 15/7 15/11 15/13 16/3 16/3 17/11 17/18 17/22 21/8 22/17 23/6 24/20 25/4 25/12 25/22 25/23 26/17 28/17 28/18 28/20 30/15 31/14 32/2 32/7 35/13 36/4 36/9 36/10 37/2 37/22 38/25 40/23 44/14 44/16 44/21 45/14 45/19 45/23 46/5 46/24 47/3 47/24 47/25 48/20 49/4 49/4 49/22 50/5 52/13 52/18 53/20 58/19 58/19 58/21 60/18 61/14 61/19 62/1 62/11 63/15 66/24 67/4 67/5 67/5 you've [2] 12/7 49/3 your [76] Z zero [3] 62/22 63/5 63/7 zero percent [1] 63/5

86 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 86 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 1 of 44 EXPERT REPORT OF JOWEI CHEN. Ph.D. I am an Associate Professor in the Department of Political Science at the University of Michigan, Ann Arbor. I am also a Faculty Associate at the Center for Political Studies of the Institute for Social Research at the University of Michigan as well as a Research Associate at the Spatial Social Science Laboratory at Stanford University. In 2007,1 received a M.S. in Statistics from Stanford University, and in 2009,1 received a Ph.D. in political science from Stanford University. I have published academic papers on political geography and districting in top political science journals, including The American Journal ofpolitical Science and The American Political Science Review, and Election Law Journal. My academic areas of expertise include spatial statistics, redistricting, racial politics, legislatures, elections, and political geography. I have unique expertise in the use of geographic information systems (GIS) data to study questions related to political geography and redistricting. 1 have provided expert reports in the following redistricting court cases: Missouri National Association for the Advancement of Colored People v. Ferguson-Florissant School District and St. Louis County Board of Election Commissioners (E.D. Mo. 2014); Rene Romo et al. V. Ken Detzner et al. (Fla. 2d Judicial Cir. Leon Cnty. 2013); The League of Women Voters of Florida et al. v. Ken Detzner et al. (Fla. 2d Judicial Cir. Leon Cnty. 2012); Raleigh Wake Citizens Association et al. v. Wake County Board ofelections (E.D.N.C. 2015); Corrine Brown et al. V. Ken Detzner et al. (N.D. Fla. 2015); City of Greensboro et al. v. Guilford County Board ofelections, (M.D.N.C. 2015); Common Cause et al. v. Robert A. Rucho et al. (M.D.N.C. 2016); League ofwomen Voters of Pennsylvania et al. v. Commonwealth of Pennsylvania et al. (No. 261 M.D. 2017). I have testified at trial in the following cases: Raleigh Wake Citizens Association et al. v. Wake County Board ofelections (E.D.N.C. 2015); City of Greensboro et al. V. Guilford County Board ofelections (M.D.N.C. 2015); Common Cause et al. v. Robert A. Rucho et al. (M.D.N.C. 2016); League of Women Voters ofpennsylvania et al. v. Commonwealth ofpennsylvania et al. (No. 261 M.D. 2017). 1 am being compensated $250 per hour for my work in this case. Research Questions and Summary of Findings: The attorneys for the plaintiffs in this case have asked me to analyze House Districts 105 and 111 in the 2012 Georgia House districting plan, as created by Act No. 277 (S.B. 513) of DEFENDANT'S EXi^lBIT /

87 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 87 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 2 of , and in the 2015 Georgia House districting plans, as created by Act No. 251 (2015 Ga. L. 1413) (H.B. 566) of2015. Specifically, I was asked to analyze: 1) Whether there is racially polarized voting within HD 105 and HD 111 under the two plans; 2) What the partisan results of the House races in HD 105 and HD 111 would have been in November 2016 if these two House races had been held using the boundaries of the 2012 House districting plan (Act No. 277); and 3) Whether race predominated in the drawing of HD 105 and HD 111 under the 2015 plan. I answered these questions by analyzing individual-level voter registration files, individual-level voter turnout history files, and precinct-level election results for Georgia's state house elections held in November 2012, 2014, and also analyzed 2010 Census data describing the racial and ethnic breakdowns of Georgia's precincts and Census blocks, as well as shapefiles depicting the district boundaries within the 2012 and 2016 Plans. In Georgia, residents are asked to select their racial identification when they register to vote. However, voters are not given the opportunity to select a partisan affiliation. Therefore, the publicly available voter registration list in Georgia contains information on the racial identification of each individual voter, along with the precinct and the House district in which each voter resides. I therefore analyzed this data in order to identify the number of voters of each racial identity residing within each precinct and within HD 105 and HD 111, as drawn by both the 2012 Plan and the 2015 Plans. Because Georgia voters are not asked to identify their partisan affiliation, it is not possible to obtain or analyze data regarding voter partisanship or election results at the sub-precinct level. In Georgia, election results are available only at the precinct level. By analyzing these precinct-level data, I concluded that voters in both HD 105 and HD 111 exhibit significant racially polarized voting. In both districts, virtually all Black voters supported Democratic House candidates in the 2012, 2014, and 2016 House elections, while 75-85% of non-black voters supported Republican candidates. Thus, race is an extremely strong proxy for partisanship in both districts. Next, I estimated the hypothetical outcomes of the November 2016 House elections, assuming they had been held under the old 2012 Plan boundaries for HD 105 and HD found that, under the 2012 Plan boundaries, a Black Democratic candidate would have defeated a

88 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 88 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filedl2/22/17 Page 3 of 44 White Republican candidate in November 2016, winning approximately 50.3%-54.4% of the vote in the two districts. Finally, I analyzed the motivations for the redrawing of HD 105 and HD 111 in the 2015 plan. First, I found that the 2015 plan decreased the African-American share of the turnout electorate by 4.0 percentage points in HD 105 and by 2.9 percentage points in HD 111. Overall, in HD 105 and HD 111, the 2015 Plan generally decreased compliance with traditional districting principles and with the principles set forth in the " Guidelines for the House Legislative and Congressional Reapportionment Committee" (Hereinafter:"Redistricting Guidelines"). Given that race and partisanship are highly correlated within these two districts, I also sought to analyze whether partisan considerations, rather than racial considerations, could account for the drawing of the new district boundaries in the 2015 plan. I found that the Legislature's primary map-drawer for the 2015 Plan had access only to racial data, but not partisan data, at the sub-precinct level. Yet strikingly, I also found that the 2015 Plan splits three precincts in HD 105 and five precincts in HD 111 in ways that consistently decreased the African-American share of the population in both districts. These two findings demonstrate that racial considerations, not partisanship, predominated in the drawing of the 2015 Plan boundaries within these eight split precincts. This report proceeds as follows. First, I describe my analysis of racially polarized voting in HD 105 and HD 111. Second, I illustrate how increasing racial minority proportions caused a pro-democratic shift during 2012 to 2016 within the 2012 Plan's boundaries for HD 105 and 111. Third, I produce vote estimates of hypothetical November 2016 House elections held using the previous 2012 Plan boundaries. Fourth, I describe how the 2015 Plan made a series of changes to the boundaries of HD 105 and HD 111 that altered the racial composition of these districts by subordinating traditional districting principles, including principles set forth in the Redistricting Guidelines. Finally, I describe how the 2015 Plan's boundaries for HD 105 and HD 111 within split precincts appears to have been driven by racial considerations. Racially Polarized Voting Analysis To analyze whether there was racially polarized voting within HD 105 and HD 111 under the 2012 and 2015 Plans, I first calculated precinct-level racial breakdowns of the turnout electorate - the set of registered voters who cast ballots - within the boundaries of HD 105 and

89 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 89 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 4 of 44 HD 111 during the November 2012, 2014, and 2016 general elections. I then compared these precinct-level racial breakdowns to the precinct-level House election results for HD 105 and 111 during these tliree elections. To estimate the partisan voting patterns of each racial group within each district, I use ecological inference (El), a commonly-used and widely-accepted statistical technique for estimating different racial groups' political behavior when racial breakdowns of such behavior is not directly reported in publicly-available data. El uses a procedure known as maximum likelihood estimation, combined with Duncan and Davis' (1953) method of bounds, to estimate the level support for a particular party's candidate among members of different racial groups across the different precincts contained within a district. The key advantage of El is that it uses observed election results and racial data jbrom all precincts within the district and estimates any differences across precincts in a particular racial group's voting behavior. Table 1 reports the El estimates of each racial group's tendency to support Democratic candidates during the November 2012, 2014, and 2016 House elections in HD 105, while Table 2 reports the El estimates for HD 111. It is clear that both HD 105 and HD 111 exhibited significantly racially polarized voting during each of these three elections. In HD 105, approximately 98-99% of Black voters supported the Democratic candidate during the three elections, whereas only 19-25% of non-black voters supported the Democratic candidate. HD 111 exhibited a similar pattern of racially polarized voting during each of the three elections. Approximately 98-99% of Black voters supported the Democratic candidate during the 2012, 2014, and 2016 elections, whereas only 16-18% of non-black voters supported the Democratic candidate.

90 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 90 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 5 of 44 Table 1: Ecological Inference and Ecological Regression Estimates of Democratic Candidates' Share of Two-Party Vote Among Among Blacks and Non-Blacks in House District 105 Ecological Inference Estimates Ecological Regression Estimates: Black Non-Black Black Non-Black 2012 House Election 99.0% [98.0%, 99.7%] 21.4% [21.0%, 22.0%] 100% [100%, 100%] 7.4% [0.4%, 10.3%] 2014 House Election 97.9% [93.3%, 99.6%] 19.2% [18.2%, 21.7%] 100% [100%, 100%] 6.6% [4.0%, 9.0%] 2016 House Election 99.3% [98.8%, 99.7%] [95% Confidence Intervals listed in brackets] 25.2% [25.0%, 25.4%] 100% [100%, 100%] 10.2% [7.8%, 12.6%]

91 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 91 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filedl2/22/17 Page 6 of 44 Table 2: Ecological Inference and Ecological Regression Estimates of Democratic Candidates' Share of Two-Party Vote Among Among Blacks and Non-Blacks in House District 111: Ecological Inference Estimates Ecological Regression Estimates: Black Non-Black Black Non-Black 2012 House Election 98.2% [90,1%, 99.8%] 18.1% [17.2%, 20.4%] 100% [100%, 100%] 8.0% [6.0%, 10.0%] 2014 House Election 98.6 % [94.9%, 99.8%] 15.7% [14.9%, 17.9%] 100% [100%, 100%] 7.4% [5.4%, 9.5%] 2016 House Election 99.3% [98.6%, 99.8%] [95% Confidence Intervals listed in brackets] 17.8% [17.5%, 18.1%] 100% [100%, 100%] 7.6% [5.3%, 10.0%]

92 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 92 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 7 of 44 Demographic and Partisan Changes in HD 105 and 111 under the 2012 Plan Having found that HD 105 and HD 111 both exhibited racially polarized voting in the 2012, 2014, and 2016 House elections, I next analyzed the racial composition and partisan performance of the two districts, as drawn by the 2012 Plan. Overall, my analysis revealed three findings: 1) The African-American share of the turnout electorate increased noticeably from November 2012 to November 2016 in both HD 105 and HD 111 under the 2012 Plan. 2) Non-African-American voters within HD 105, as drawn under the 2012 Plan, became somewhat more likely to favor a Black Democratic House candidate in November 2016, compared to previous elections. 3) As a result of these racial and partisan shifts. Democratic House candidates' vote share significantly increased among voters residing within the 2012 Plan boundaries for HD 105 and HD 111 from November 2012 to November Below, I describe and illustrate these three findings in greater detail: First, both HD 105 and HD 111, as drawn under the 2012 Plan, became more heavily African-American from 2014 to This increasing African-American share of the electorate within the 2012 Plan's HD 105 boundaries is illustrated in Table 3, which shows that African- Americans comprised 35.2% of the Election Day turnout in November 2012, 35.7% in November 2014, and 37,0% by November 2016 (counting only voters who reside within the 2012 Plan's HD 105 boundaries). Table 4 illustrates an even more significant increase in African-American share of the electorate within the 2012 Plan's HD 111 boundaries: African- Americans comprised 36.1% of the Election Day turnout in November 2012, 37.6% in November 2014, and 40.3% by November 2016 (counting only voters who reside within the 2012 Plan's HD 111 boundaries). During all three elections, voters in both districts exhibited significant racially polarized voting patterns, with African-Americans favoring Democratic House candidates at a rate of around 98-99%. Thus, it is clear that this demographic pattern of increasing African-American population within the 2012 Plan's HD 105 and HD 111 boundaries would have caused a substantial increase in Democratic vote share by the November 2016 House elections in both districts.

93 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 93 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 8 of 44 Table 3: HD 105 Precinct-Level Voter Turnout by Race under the 2012 and 2015 Plans 2012 Election Turnout 2014 Election Turnout 2016 Election Turnout 2016 Election Turnout Within HD 105 Within HD 105 Within HD 105 Within HD 105 Boundaries (2012 Plan) Boundaries (2012 Plan) Boundaries (2012 Plan) Boundaries (2015 Plan) Pet: Precinct Name: Black Non-Black Black Non-Black Black Non-Black Black Non-Black 001 Harbins A 329' 1651' 060 Lawrenceville D 897' 705' 474' 379' 938' 893' 1088' 933' 071 Lawrenceville F Baycreek K Baycreek C Baycreek D Baycreek F Lawrenceville M 1233' 1152' 865' 681' 1490' 1525' 146 Baycreek H 333' 1997' 246' 1396' 468' 2096' 468' 2096' 147 Baycreek I Harbins C Totals by Race: 7,586 13,975 5,180 9,335 9,199 15,660 8,443 17,198 (35.2%) (64.8%) (35.7%) (64.3%) (37.0%) (63.0%) (32.9%) (67.1%) Totals: 21,561 14,515 24,945 25,641 * Indicates that the precinct was split into multiple districts, including House District 105. Only those voters residing within HD 105 are included in this table's turnout numbers. In particular, note that the HD 105 portion of Lawrenceville D was different under the 2012 plan than under the 2015 plan. Therefore, the November 2016 turnout numbers for Lawrenceville D within HD 105 are different under the 2012 plan and under the 2015 plan

94 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 94 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 9 of 44 Table 4: HD 111 Precinct-Leve! Voter Turnout by Race under the 2012 and 2015 Plans 2012 Election Turnout 2014 Election Turnout 2016 Election Turnout 2016 Election Turnout Within HD 111 Within HD 111 Within HD 111 Within HD 111 Boundaries (2012 Plan) Boundaries (2012 Plan) Boundaries (2012 Plan)» Boundaries (2015 Plan) oo oo Precinct Name: Black Non-Black Black Non-Black Black Non-Black Black Non-Black 26 - Tussahaw 25' 498' 29 - Lowes North Hampton 505' 1208' 395' 823' 659' 1240' 32 - Mount Carmel ' 803' 34 - Wesley Lakes McDonough Hickory Flat 795' 583' 40 - Stockbridge West Stagecoach Unity Grove Pates Creek Oakland Flippen 1106' 57 - Dutchtown Grove Park McDonough Central 288' 731' 185' 562' 348' 775' ' Totals by Race: 9,320 16,517 6,898 11,443 11,735 17,352 10,794 18,057 (36.1%) (63.9%) (37.6%) (62.4%) (40.3%] (59.7%) (37.4%) (62.6%) Totals: 25,837 18,341 29,087 28,851 * Indicates that the precinct was split into multiple districts, including House District 111. Only those voters residing within HD 111 are included in this table's turnout numbers

95 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 95 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 10 of 44 Second, the non-african-american portion of the electorate in HD 105 exhibited a noticeable increase in its support for a Black Democratic candidate in November 2016, compared to earlier elections. This increase in Democratic support is illustrated by the Ecological Inference estimates in Table 1, which predict that non-blacks support for a Black Democratic candidate increased from 21.4% in November 2012 to 25.2% in November This increase is partially attributable to an increase in the Hispanic and Asian shares of the non-black portion of the electorate in HD 105, illustrated in Table 7, as Hispanic and Asian voters were more likely than non-hispanic white voters to support Black Democratic candidates. As a result of these two demographic shifts within the 2012 Plan's HD 105 and HD 111- the increase in African-Americans and other Democratic-supporting minority populations both districts would have exhibited a substantial increase in Democratic vote share in the November 2016 House elections, if not for the 2015 Plan's redrawing of the two districts' boundaries. This pro-democratic shift within the 2012 Plan's HD 105 and HD 111 boundaries is clearly seen in Tables 5 and 6. These Tables show the actual precinct-level House election vote counts for only those precincts that were assigned to HD 105 or HD 111 and whose district boundaries were identical under both the 2012 and the 2015 Plans. In other words, these precincts are the ones whose House district assignments were unaffected by the 2015 Plan redistricting. There were seven such precincts in HD 105 (Table 5) and six such precincts in HD 111 (Table 6). Table 5 illustrates that all seven precincts in HD 105 that were unaffected by the 2015 Plan substantially increased their Democratic vote shares in House elections from November 2012 to 2016; in fact, all seven precincts increased their Democratic vote share from November 2014 to For example, voters in Precinct 71 (Lawrenceville F) supported Democrat Renita Hamilton at a 66.7% rate in November 2012 and 2014, but by November 2016, the precinct's support for the Democratic candidate (Donna McLeod) increased to 70.2%. Overall, all seven precincts increased their respective Democratic vote shares by a margin of 3.5 to 9.6 percentage points between November 2012 and

96 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 96 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 11 of 44 Table 5: House Election Results in Precincts in which HD 105 Boundaries Remained Unchanged from the 2012 Plan to the 2015 Plan 2012 Election Results (HD 105) 2014 Election Results (HD 105) 2016 Election Results (HD 105) Precinct Name: Renita Hamilton Joyce Chandler Renita Hamilton Joyce Chandler Donna McLeod Joyce Chandler (Black (White (Black (White (Black (White Democrat) Republican) Democrat) Republican) Democrat) Republican) 71-Lawrenceville F Baycreek K Baycreek C Baycreek D Baycreek F Baycreek H Baycreek I Election Results (HD 105) 2014 Election Results (ED) 105) 2016 Election Results (ED) 105) Precinct Name: Democratic Candidate Vote Share Democratic Candidate Vote Share Democratic Candidate Vote Share 71-Lawrenceville F 66.7% 66.7% 70.2% 78-Baycreek K 35.8% 37.9% 45.2% 80-Baycreek C 47.4% 45.3% 54.0% 91-Baycreek D 34.4% 32.4% 40.0% 134-Baycreek F 48.9% 48.0% 58.5% 146-Baycreek H 24.1% 24.7% 32.9% 147-Baycreek I 48.0% 47.1% 54.2% Only includes precincts in which the boundaries of HD 105 did not change from the 2012 to the 2015 Plan. 11

97 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 97 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 12 of 44 Table 6: House Election Results in Precincts in which HD 111 Boundaries Remained Unchanged from the 2012 Plan to the 2015 Plan 2012 Election Vote Counts (HD 111) 2014 Election Vote Counts (HD 111) 2016 Election Vote Counts (HD 111) Precinct Name: Brian Strickland Brian Strickland Brian Strickland Bill Blackmon (White Jim Nichols (White Darryl Payton (White (Black Democrat) Republican) (White Democrat) Republican) (Black Democrat) Republican) 29 - Lowes Wesley Lakes Unity Grove Pates Creek Oakland Dutchtown Election Results (HD 111) 2014 Election Results (HD 111) 2016 Election Results (HD 111) Precinct Name: Democratic Candidate Vote Share Democratic Candidate Vote Share Democratic Candidate Vote Share 29 - Lowes 48.1% 48.5% 53.3% 34 - Wesley Lakes 63.6% 62.1% 65.4% 48 - Unity Grove 21.2% 20.2% 21.5% 50 - Pates Creek 42.9% 43.5% 49.8% 51 - Oakland 35.1% 35.6% 42.3% 57 - Dutchtown 31.6% 30.7% 35.9% Only includes precincts in which the boundaries of HD 105 did not change from the 2012 to the 2015 Plan. 12

98 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 98 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 13 of 44 Table 6 illustrates a similar pro-democratic pattern for the six precincts in HD 111 that were unaffected by the 2015 Plan: All six precincts substantially increased their Democratic vote shares in House elections from November 2012 to 2016; all six precincts also increased their Democratic vote share from November 2014 to For example, voters in Precinct 29 (Lowes) supported the Democratic House candidate at a 48.1% rate in November 2012 and a 48.5% rate in 2014, but by November 2016, the precinct's support for the Democratic candidate (Darryl Payton) increased to 53.3%. Overall, all six precincts increased their respective Democratic vote shares by a margin of 0.3 to 6.9 percentage points between November 2012 and Overall, these two Tables illustrate that HD 105 and 111, as drawn by the 2012 Plan, would have exhibited a substantial increase in Democratic vote share in the November 2016 House elections, if not for the 2015 Plan's redrawing of the two districts' boundaries. Among the seven unaffected precincts in HD 105, the Black Democratic candidate's vote share increased by 7.1 percentage points from November 2012 to November Among the six unaffected precincts in HD 111, the Black Democratic candidate's vote share increased by 4.2 percentage points from November 2012 to November These changes in the unaffected portions ofhd 105 and 111 are attributable primarily to the increasing minority proportions of the electorate within the two districts' boundaries under the 2012 Plan. Given that voting patterns in these two districts are highly racially polarized, it is not surprising that these partisan shifts coincided with racial shifts in the composition of the electorate. Tables 7 and 8 show how the partisan shifts within the 2012 Plan's boundaries for HD 105 and 111 are clearly attributable to the increasing African-American, Hispanic, and Asian proportions of the electorate within the two districts' boundaries under the 2012 Plan. For HD 105, Table 7 lists the racial breakdown of voters residing in the 2012 Plan's HD 105 boundaries who turned out to vote in November 2012, 2014, and The final column then lists the racial breakdown of November 2016 voters who turned out and who resided within the new HD 105 boundaries, as drawn by the 2015 Plan. Table 8 shows the analogous calculations for HD 111, as drawn by the 2012 and the 2015 Plans. 13

99 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 99 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filedl2/22/17 Page 14 of 44 Table 7: HD 105 District-Wide Turnout by Race under the 2012 and 2015 Plans Racial Group: 2012 Election Turnout Within HD 105 Boundaries (2012 Plan) 2014 Election Turnout Within HD 105 Boundaries (2012 Plan) 2016 Election Turnout Within HD 105 Boundaries (2012 Plan) 2016 Election Turnout Within HD 105 Boundaries (2015 Plan) White 10,885 (50.5%) 7,468 (51.5%) 10,800 (43.4%) 12,554 (49%) Black (non-hispanic) 7,586 (35.2%) 5,180 (35.7%) 9,199 (37%) 8443 (32.9%) Hispanic 747 (3.5%) 372 (2.6%) 1397 (5.6%) 1178 (4.6%) Asian or Pacific Islander 348 (1.6%) 175 (1.2%) 603 (2.4%) 552 (2.2%) American Indian or Alaskan Native 10 (0%) 7 (0%) 17 (0.1%) 22 (0.1%) Other or Unknown 1,985 (9.2%) 1,313 (9.0%) 2,843 (11.4%) 2,892 (11.3%) Total Turnout: 21,561 14,515 24,859 25,641 * Indicates that the precinct was split into multiple districts, including House District 105. Only those voters residing within HD 105 are included in this table's turnout numbers. In particular, note that the HD 105 portion of Lawrenceville D was different under the 2012 plan than under the 2015 plan. Therefore, the November 2016 turnout numbers for Lawrenceville D within HD 105 are different under the 2012 plan and under the 2015 plan. 14

100 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 100 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 15 of 44 Table 8: HD 111 District-Wide Turnout by Race under the 2012 and 2015 Plans Racial Group: 2012 Election Turnout Within HD 111 Boundaries (2012 Plan) 2014 Election Turnout Within HD 111 Boundaries (2012 Plan) 2016 Election Turnout Within HD 111 Boundaries (2012 Plan) 2016 Election Turnout Within HD 111 Boundaries (2015 Plan) White 13,349 (51.7%) 9,422 (51.4%) 13,251 (45.6%) 13,836 (48.0%) Black (non-hispanic) 9,320 (36.1%) 6,898 (37.6%) 11,735 (40.3%) 10,794 (37.4%) Hispanic 463 (1.8%) 259 (1.4%) 692 (2.4%) 679 (2.4%) Asian or Pacific Islander 235 (0.9%) 113 (0.6%) 379 (1.3%) 436 (1.5%) American Indian or Alaskan Native 1 (0%) 0 (0%) 10 (0%) 8 (0%) Other or Unknown 2,469 (9.6%) 1,649 (9.0%) 3,020 (10.4%) 3,098 (10.7%) Total Turnout: 25,837 18,341 29,087 28,851 * Indicates that the precinct was split into multiple districts, including House District 111. Only those voters residing within HD 111 are included in this table's turnout numbers. 15

101 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 101 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 16 of 44 Together, these Tables illustrate a similar pattern in both districts. From November 2012 to 2016, each racial group's share of the total election-day turnout increased significantly within the 2012 Plan's HD 105 boundaries: African-Americans increased from 35.2% to 37%, Hispanics increased from 3.5% to 5.6%, and Asians increased from 1.6% to 2.4%. Similarly, within the 2012 Plan's HD 111 boundaries, each racial group's share of the total election-day turnout also increased significantly from November 2012 to 2016: African-Americans increased from 36.1% to 40.3%, Hispanics increased from 1.8% to 2.4%, and Asians increased from 0.9% to 1.3%. These increases in racial minority population explain why every single precinct that remained within HD 105 and 111 in both the 2012 Plan and the 2015 Plan exhibited a noticeable increase in Democratic candidate vote share in the 2016 House elections, as compared to the 2012 House elections. Moreover, Tables 7 and 8 also illustrate how this trend of increasing racial minority populations in HD 105 and 111 was successfully reversed by the 2015 Plan's redrawing of the two districts. Within the 2012 Plan's boundaries for HD 105, the November 2016 turnout electorate consisted of 37% African-American voters and 5.6% Hispanic voters. But within the 2015 Plan's new boundaries for HD 105, the November 2016 turnout electorate consisted of only 32.9% African-Americans and 4.6% Hispanics. A similar reversal occurred in HD 111: Within the 2012 Plan's boundaries for HD 111, the November 2016 turnout electorate consisted of 40.3% African-American voters. But within the 2015 Plan's new boundaries for HD 111, the November 2016 turnout electorate consisted of only 37.4% African-Americans. By redrawing the boundaries of HD 105 and 111, the 2015 Plan decreased the racial minority proportions of the electorate, thus reversing the demographic changes that had occurred within the 2012 Plan's boundaries for HD 105 and 111 in recent years. Estimate of November 2016 Election Results Under 2012 Plan Boundaries Next, I estimated what the partisan results of the House races in HD 105 and HD 111 would have been in November 2016 if these two House races had been held using the boundaries of the 2012 House districting plan (Act No. 277). Specifically, I assumed that the set of electionday voters would have been exactly the same as the voters who actually turned out in November In total, I identified a total of 23,696 voters in HD 105 and 29,087 who satisfy the following two criteria: 16

102 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 102 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 17 of The voter cast a ballot in the November 2016 General Election. 2. As of November 2016, the voter resided within the 2012 Plan boundaries of HD 105 or HD 111. I then used this set of voters, along with their respective racial identifications on their voter registrations, to construct estimates of hypothetical November 2016 House election outcomes within the borders of HD 105 and HD 111 of the 2012 Plan. Specifically, I use Ecological Inference (El) to derive predicted voting patterns by race and to estimate the rate at which voters cast ballots in House elections. Because the November 2016 elections included a US presidential race, and because turnout levels differ significantly between presidential and non-presidential elections, 1 use the November 2012 House election results and precinct-level turnout counts by race in order to derive precinct-level El estimates about racial voting patterns. I then apply these racial voting estimates to precinct-level turnout counts by racial group in November 2016 in order to estimate how many votes would have been cast for each party's candidate in each precinct. The November 2012 House elections featured a Black Democratic candidate and a White Republican candidate in both the HD 105 (Renita Hamilton and Joyce Chandler) and HD 111 (Brian Strickland and Bill Blackmon) races, which were held using the boundaries of the 2012 Plan. Thus the El estimates derived using the results of this election give us reliable predictions regarding the racial voting patterns within each precinct in a House election featuring a Black Democratic and White Republican candidate during a presidential election year. Table 9 reports the El estimates for HD 105, while Table 10 reports the El estimates for HD 111. The first row of Table 9 reports, for example, that African-American voters in Gwinnett County's Precinct 60 ("Lawrenceville D") who turn out to vote exhibit a roll-off rate of 8.3%, meaning that 91.7% of those who turn out are expected to cast a vote in their House race; Among those who cast a vote, 98.2% are expected to vote for a Black Democratic candidate. Meanwhile, only 44.7% of non-black voters in Precinct 60 would have supported a Black Democratic candidate. 17

103 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 103 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 18 of 44 Table 9: El-Based Estimates of Hypothetical 2016 Election Results Within the Boundaries of HD 105 from the 2012 House Plan El Estimates of Voter El Estimates within HD 105 El-Based Estimates of Democratic Votes in Roll-Off within HD 105 Boundaries from the 2012 House November 2016 within HD 105 Boundaries from Boundaries from the 2012 Plan the 2012 House Plan House Plan Precinct Name: El El El Estimates of El Estimates of El-based Estimate of El-based Estimate of Estimates of Estimates of Black Voter Non-Black Nov Votes for a Nov Votes for a Black Voter Non-Black Support for a Voter Support Black Democratic White Republican Roll-Off: Voter Roll- Off: Black Democratic Candidate: for a Black Democratic Candidate: Candidate: Candidate: 60-Lawrenceville D' 8.3% 3.5% 98.2% 44.7% Lawrenceville F 6.2% 3.0% 98.2% 35.9% Baycreek K 8.6% 3.9% 97.4% 17.0% Baycreek C 7.5% 3.4% 97.9% 21.8% Baycreek D 6.6% 3.1% 98.1% 17.5% Baycreek F 5.6% 2.8% 98.8% 19.5% Lawrenceville M' 6.9% 3.2% 98.2% 39.6% Baycreek H ' 6.8% 3.1% 98.2% 12.3% Baycreek 1 7.5% 3.3% 97.9% 23.1% Total Estimated Votes Total Estimated Votes for a Black Democratic for a White Candidate: Republican Candidate: 11,933 11,763 * Indicates that the precinct was split into multiple districts, including House District 105. Only those voters residing within HD 105 are included in the El estimates and estimated vote totals reported in this Table. Note that an extremely small portion of precinct "Baycreek G" also lies within HD 105 from the 2012 Plan. However, there were no registered voters within this portion, so Baycreek G is not listed on this Table. 18

104 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 104 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 19 of 44 Table 10: El-Based Estimates of Hypothetical 2016 Election Results Within the Boundaries of HD 111 from the 2012 House Plan El Estimates of Voter Roll-Off within HD 111 Boundaries (2012 Plan) El Estimates within HD 111 Boundaries from the 2012 House Plan El-Based Estimates of Democratic Votes in November 2016 within HD 111 Boundaries from the 2012 House Plan Precinct Name: El El Estimates of Estimates of Black Voter Non-Black Roll-Off: Voter Roll- Off: El Estimates of El Estimates of Black Voter Non-Black Support for a Voter Support Black for a Black Democratic Democratic Candidate: Candidate: El-based Estimate of El-based Estimate of Nov Votes for a Nov Votes for a Black Democratic White Republican Candidate: Candidate: 29 - Lowes 1.0% 0.9% 98.5% 17.3% North Hampton' 3.1% 3.9% 98.3% 12.3% Mount Carmel 1.4% 1.9% 98.7% 23.2% Wesley Lakes 3.2% 4.6% 98.8% 26.2% Stockbridge West 4.6% 5.5% 99.2% 41.2% Stagecoach 3.5% 5.3% 98.7% 20.0% Unity Grove 3.4% 4.3% 98.8% 9.5% Pates Creek 3.0% 4.6% 98.6% 17.2% Oakland 4.6% 7.0% 98.4% 13.5% Dutchtown 3.5% 4.6% 98.4% 14.3% McDonough 3.6% 5.1% 98.4% 13.1% Central* * Indicates that the precinct was split into multip of the 2012 Plan are included in the El estimates Total Estimated Votes Total Estimated Votes for a Black Democratic for a White Candidate: Republican Candidate: 14,634 14,453 e districts, including House District 111. Only those voters residing within HD 111 and estimated vote totals reported in this Table. 19

105 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 105 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 20 of 44 From these El estimates, I am able to predict that Precinct 60 would have produced a total of 1,230 votes for a Black Democratic candidate and 492 votes for a White Republican candidate. Note that Precinct 60 was split by the 2012 Plan into HD 104 and HD 105, and this El analysis considers only voters who resided, as of November 2016, within the boundaries of HD 105 from the 2012 Plan. Applying this methodology to all precincts within the two districts, I find that both HD 105 and HD 111 would have been won by a Black Democratic candidate in November 2016 if these House races had been held using the boundaries of the 2012 Plan. Specifically, as Table 9 illustrates, voters in HD 105 would have favored the Democratic over the Republican candidate by 11,933 to 11,763 votes. Meanwhile, voters in HD 111 would have favored the Democratic over the Republican candidate by 14,634 to 14,453 votes, as illustrated in Table 10. Yet even these El estimates likely under-estimate the number of Democratic voters residing within both districts as of November The El estimates used in Tables 9 and 10 are based on voting patterns observed in the November 2012 House elections. From 2014 to 2016, these voting patterns shifted noticeably in a pro-democratic direction, due to increases in racial minority proportions in HD 105 and 111, as described earlier in this report. This pro-democratic shift in the two districts suggests that the use of El-based predictions likely under-estimates the true Democratic vote share in a hypothetical November 2016 election held using the 2012 Plan's boundaries. Hence, a more realistic method of estimating hypothetical November 2016 election outcomes involves using actual precinct-level House election outcomes from November 2016 for those precincts that were not removed from HD 105 or 111 by the 2015 Plan; for precincts that were removed, the same El predicted results are used. 20

106 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 106 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filedl2/22/17 Page 21 of 44 Table 11: Combined Estimates of Hypothetical 2016 Election Results Within the Boundaries of HD 105 from the 2012 House Plan Estimates of November 2016 Votes for a Black Democratic Candidate within the Boundaries of HD 105 of the 2012 House Plan Estimates of November 2016 Votes for a White Republican Candidate within the Boundaries of HD 105 of the 2012 House Plan Precinct Name: Actual Nov Votes for Dem. Donna McLeod (in precincts where HD 105 was not altered by the 2015 Plan): Bl-based Estimate of Nov Votes for a Black Democratic Candidate (in precincts where HD 105 was altered by the 2015 Plan): Actual Nov Votes for Rep. Joyce Chandler (in precincts where HD 105 was not altered by the 2015 Plan): El-based Estimate of Nov Votes for a Black Democratic Candidate (in precincts where HD 105 was altered by the 2015 Plan): 60-Lawrenceville D ' Lawrenceville F Baycreek K Baycreek C Baycreek D Baycreek F Lawrenceville M' Baycreek H' Baycreek Combined Total Estimated Votes for a Black Democratic Candidate: 12,780 (54.4%) Combined Total Estimated Votes for a White Republican Candidate: 10,702 (45.6%) * Indicates that the precinct was split into multiple districts, including House District 111. Only those voters residing within HD 111 of the 2012 Plan are included in the El estimates and estimated vote totals reported in this Table. 21

107 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 107 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Flledl2/22/17 Page 22 of 44 Table 12: Combined Estimates of Hypothetical 2016 Election Results Within the Boundaries of HD 111 from the 2012 House Plan Estimates of November 2016 Votes for a Black Democratic Candidate within the Boundaries of HD 111 of the 2012 House Plan Estimates of November 2016 Votes for a White Republican Candidate within the Boundaries of HD 111 of the 2012 House Plan Precinct Name: Actual Nov Votes for Dem. Darryl Payton (in precincts where HD 111 was not altered by the 2015 Plan): El-based Estimate of Nov Votes for a Black Democratic Candidate (in precincts where HD 111 was altered by the 2015 Plan): Actual Nov Votes for Rep. Brian Strickland (in precincts where HD 111 was not altered by the 2015 Plan): El-based Estimate of Nov Votes for a White Republican Candidate (in precincts where HD 111 was altered by the 2015 Plan): 29 - Lowes North Hampton' Mount Carmel Wesley Lakes Stockbridge West Stagecoach Unity Grove Pates Creek Oakland Dutchtown McDonough Central* Combined Total Estimated Votes for a Black Democratic Candidate: 14,561 (50.9%) Combined Total Estimated Votes for a White Republican Candidate: 14,023 (49.1%) * Indicates that the precinct was split into multiple districts, including House District 111. Only those voters residing within HD 111 of the 2012 Plan are included in the El estimates and estimated vote totals reported in this Table. 22

108 Case 1:17-cv TCB-WSD-BBM Document 140 Filed 03/27/18 Page 108 of 166 Case l:17-cv tcb-wsd-bbm Document 63-1 Filed 12/22/17 Page 23 of 44 Table 11 illustrates this method for HD 105. Under the 2012 Plan, HD 105 contained nine precincts, including three split precincts. Of these nine precincts, seven were unaffected by the 2015 Plan: The same portions of these seven precincts assigned to HD 105 under the 2012 Plan were again assigned to HD 105 under the 2015 Plan. The remaining two precincts were affected by the 2015 redistricting: Precinct 144 (Lawrenceville M) was completely removed from HD 105, while the borders ofhd 105 were altered within Precinct 60 (Lawrenceville D). Moreover, all of Lawrenceville M and portions of Lawrenceville D were reassigned to HD 104 in the 2015 Plan. In November 2016, HD 104 featured an uncontested House race with no Democratic candidate; therefore, no meaningful election results from November 2016 are available for these two reassigned precincts. For the seven precincts in HD 105 unaffected by the 2015 Plan, Table 11 simply counts the number of House election votes received by the Black Democratic candidate (Donna McLeod) and the White Republican candidate (Joyce Chandler) in November 2016, with no El estimates used. For the remaining two precincts affected by the 2015 Plan - Precincts 60 and Table 11 uses the same El-based predictions derived previously in Table 9. Table 11 sums together these actual election results and El-based predictions for the nine precincts within the 2012 Plan boundaries of HD 105. In total, a Black Democratic candidate is expected to receive 12,780 votes, whereas a White Republican would receive 10,702 votes in a November 2016 House election held using the 2012 Plan boundaries for HD 105. The Black Democratic candidate's predicted vote share of 54.4% accounts for the increases in African- American and other minority populations that occurred in HD 105 during , thus producing an even more Democratic-leaning prediction than the Table 9 estimates, which solely relied upon El predictions about racial voting patterns from 2012 election data. Table 12 uses this identical methodology for HD 111, yielding a similar prediction. The 2012 Plan boundaries for HD 111 contained all or split portions of 11 different precincts. Five of the precincts were affected by the 2015 Plan redistricting, while the House district assignments for the remaining six precincts were unaffected by the 2015 Plan. For the six unaffected precincts. Table 12 uses the actual election results from the November 2016 between Darryl Payton (Democrat) and Brian Strickland (Republican). For the remaining five precincts that affected by the 2015 Plan's redrawing of the HD 111, Table 12 uses the El-based predictions 23