An Analysis of American Attitudes Toward Immigration

An Analysis of American Attitudes Toward Immigration Daniel Chavez January 2016 1 Abstract This paper analyzes individuals opinions towards how they believe the number of immigrants coming to the United States should change. Whereas an individual believes that the amount of immigrants coming to the United States should be increased, decreased, or remain the same. The objective of this paper s analysis is to ultimately investigate how various demographic, behavioral, and attitudinal factors shape, or are related to, individuals beliefs on how the number of immigrants coming to the United States should change. Beyond a positive analysis, this paper presents the various policy implications which stem from the analysis findings. The paper uses General Social Survey (GSS) data in its investigation. By analyzing GSS data over the past decade this paper also uncovers how individuals opinions on this issue have changed over time, and proposes potential explanations for the changes. 2 Introduction: American Attitudes Toward Immigration Over the past century, and perhaps even further before, immigration has remained an integral part of American political discourse. Immigration reform, one of the most partisan issues on legislative dockets nationwide has been, and will most likely continue to be, the topic of heated debates across the nation. Outside of the political realm, everyday Americans take part in the discussion within their homes, schools, churches, and communities. This paper seeks to investigate what factors underlie the national discourse by studying how the American public views the current amount of immigrants entering the United States. This analysis not only provides important insights into how Americans from di erent demographic groups and backgrounds believe that the number of immigrants entering the U.S. should change, and how immigrants have a ected society, but also provide information as to how these beliefs are formed. Because attitudes toward immigration will a ect immigration in years to come, in order to fully understand the implications these attitudes hold one 1

must first understand the implications of immigration. For starters, many prevalent social and economic e ects we see today have arisen as a direct product of immigration. Within day-to-day life immigration shapes various community institutions all around us. Take schools for example, where the rapidly increasing in cultural diversity is changing how and what we teach to our children. In some areas of the country this increase has led to the emergence of multicultural education methods in some schools. These methods prioritize educating students in a manner which develops the knowledge, attitudes, and skills needed to interact positively with people from diverse groups. 1 Other larger community institutions have had to adapt to this rapidly changing demographic as well. As a result of more immigration large institutions such as hospitals and health care networks must to adapt by further transforming their services to be more accessible to immigrants and their families, as well as by culturally appropriating to their unique preferences and needs. Finally, because the number of immigrants entering the United States has been increasing it can only follow that in the long term they will grow to become a sizable fraction of the population. This is perhaps the biggest reason why the topics pertaining to immigration are important to study. The Pew Research Center estimates that if current demographic trends continue, future immigrants and their descendants will be an even bigger source of population growth. Between 2015 and 2065, they are projected to account for 88 percent of the U.S. population increase, or 103 million people. 2 Immigrants are therefore a demographic that cannot be ignored by future public policy. And future public policy will hinge on the attitudes of American voters today. So what will shifts in attitudes toward immigration alone then mean for policy? The bottom line is that as these attitudes change, the policy landscape will change in the same direction. For example, currently American universities policies towards fostering and encouraging diversity in their classrooms and on campus play a major role in exposing Americans to people of di erent backgrounds and cultures; however, in the future policies like these will most likely shift in ways that follow the country s attitude towards immigrants and immigration. On a much broader scale, depending on where the country s opinions fall towards immigrants the coming years could see the United States visa and naturalization policies take on substantial inclusive or exclusive type reforms. Therefore, because of these and many other implications which stem from immigration and attitudes toward immigrants, the study of these attitudes is a topic of great significance. This paper will attempt analyze attitudes towards immigration over the past decade, and study how and why these attitudes have changed and how these changes could a ect policy across the nation. 1 Banks, James A. et al.. Diversity Within Unity: Essential Principles for Teaching and Learning in a Multicultural Society. The Phi Delta Kappan 83.3 (2001): 196 203. Web. 2 http://www.pewhispanic.org/2015/09/28/modern-immigration-wave-brings-59-millionto-u-s-driving-population-growth-and-change-through-2065/ accessed October 8, 2015 2

The analytic aspect of this paper will use General Social Survey (GSS) data to explore opinions on the question of whether immigrants improve American society. The GSS is the nation s top social indicator survey. The survey, which is sponsored by the National Science Foundation, is conducted every other year and contains a number of questions pertaining to various social issues. The primary outcome variable utilized in this paper is LETIN1, which has been asked biennially since 1972; however, this paper will only study data from the years 2004, 2006, 2008, 2010, 2012, and 2014. LETIN1 data comes from the question: Do you think the number of immigrants to America nowadays should be increased a lot, increased a little, remain the same, decreased a little, or decreased a lot?. 3 Those who wish to see immigration increased will most likely support immigrant friendly policies at home while the opposite will hold true for those who wish to see immigration decreased. The national figures for LETIN1 are shown in table 1 below. 4 Table 1: Frequency Distribution of LETIN1 (2004-2014) Response N (weighted) Column Percentage Increased a lot 371 3.9 Increased a little 820.2 8.6 Remain the same 3,509.3 36.9 Decreased a little 2,341.1 24.6 Decreased a lot 2,467.8 26 Total 9,509.4 100 Table 1 shows that a majority of respondents wish that the number of immigrants coming to the United States is decreased. A little more than one third of the sample feels that the number of immigrants coming to the United States should remain the same, while about one eighth of respondents felt that that number should be increased. Given the backlash against immigrants and immigration in our country for the past decade, the fact that a majority of respondents in the sample wish to decrease the number of immigrants coming to the Unites States is not very surprising. On the other hand, the large number of those who believe that the number of immigrants coming into the country should remain the same could indicate that a large number of Americans may in fact be satisfied with the current state of immigration as is. The comparatively small group of Americans who wish to see the number of immigrants coming to the country increased could possibly grow in number over the coming years. This potential growth depends entirely on what factors drive these positive attitudes toward immigrants and if these factors will surge in the coming years. The breakdown of these numbers sheds light on the fact that most of our 3 https://gssdataexplorer.norc.org/projects/2550/variables/4863/vshow accessed September 30, 2015. 4 http://sda.berkeley.edu/sdaweb/analysis/?dataset=gss14 accessed October 2, 2015. 3

nation is not quite as open to immigrants as many might have once thought, but that there is still plenty of room for the story to change. The body of this paper will delve deeper into these observations by further analyzing Americans opinions towards immigration, and how these opinions have changed over time. Ultimately, this paper will come to conclusions which could a ect our nation s outlook and action on immigration and immigration related policies. 3 Attitudes Toward Immigration Over the Past Decade The previous section provided a table which outlined Americans attitudes towards how the number of immigrants coming into the country should be changed. The objective of this section is to answer how these attitudes have changed over the past decade. Tables 2 and 3 below present the numbers for every other year from 2004 to 2014. Response Table 2: Frequency Distribution of LETIN1 From 2004 to 2014 Year 04 06 08 10 12 14 N (weighted) Increased a lot 68.9 76.4 39.9 59.4 57.4 68.9 Increased a little 129.4 161.6 112.0 131.8 121.2 164.2 Remain the same 696.2 687.0 455.0 501.1 519.0 651.1 Decreased a little 560.1 462.6 316.3 342.3 289.5 370.3 Decreased a lot 524.8 558.8 382.8 361.3 285.4 354.8 Total 1,979.3 1,946.4 1,306.1 1,396.0 1,272.4 1,609.3 Table 3: Frequency Distribution of LETIN1 From 2004 to 2014 Response Year 04 06 08 10 12 14 Column Percent Increased a lot 3.5 3.9 3.1 4.3 4.5 4.3 Increased a little 6.5 8.3 8.6 9.4 9.5 10.2 Remain the same 35.2 35.3 34.8 35.9 40.8 40.5 Decreased a little 28.3 23.8 24.2 24.5 22.7 23.0 Decreased a lot 26.5 28.7 29.3 25.9 22.4 22.0 Total 100 100 100 100 100 100 4

Discernible patterns arise when we take a look at what has happened to opinions toward immigration over the past decade. The percentage of those who supported immigration being increased by a lot ultimately increased over the past decade, peaking in 2012 at 4.5 percent and falling back down to 4.3 percent in 2014. There also appears to be a dip in support for increasing immigration by a lot between the years 2006 and 2008; however, from the year 2008 (when the number of those who supported increasing immigration by a lot was its lowest) to the year 2010, we see the largest jump in the percentage of people who held this stance. A jump from 3.1 percent to 4.3 percent. The story is a bit di erent for those who only want to increase immigration by a little. This group did not fluctuate much, in fact it steadily increased throughout the past decade, reaching its peak in 2014 when 10.2 percent of those sampled held this opinion. Starting in 2004 both these groups combined stood at 10 percent. In 2014 this number stood at 14.5 percent, indicating that over the past decade the number of Americans who support increasing immigration has increased. The largest group throughout the past decade has been the group of respondents who want the amount of immigrants coming to the country to remain the same. After remaining relatively stagnant from the years 2004 to 2010 the percentage of Americans who felt this way leaped five percentage points to 40.8 percent and 2012 and essentially stayed this way in 2014. As far as those Americans who wish to see the number of immigrants coming to the U.S. decreased their numbers have dwindled. The percentage of people who wish to see immigration decreased by a little has decreased by roughly five percentage points over the past decade from 28.3 percent to 23 percent. The drop was 4 percent for the group who wishes to see immigration decreased by a lot. Their numbers dropped from 26.5 percent to 22 percent. The big picture tells us that over the past decade, although numbers have changed, the majority of Americans still want to see immigration decreased. On the other hand, their majority is waning as those who want immigration to increase or stay the same is on the rise. The combination of these three patterns could point to the fact that over the past decade the American public believed that immigration in fact decreased, leading the people who wanted to see it decreased move to the stay the same column, and those who wanted it to stay the same moved to the increase column. This would explain the drop in those who want immigration decreased and the rise in those who want it to stay the same and who want it increased. This is supported by the fact that immigration did in fact decrease over this time period, making this explanation very plausible. 5 Apart from this explanation potential factors shaping attitudes toward immigration, which will be investigated in the next section, could have shifted over the past decade and could have been the cause of these changes as well. In terms of policy implications these changes could signal an opening for pro-immigrant legislators to begin to push through immigration related policy which would allow for increased immigration, or at least accommodation of immigrants within our borders. This opportunity 5 http://www.pewresearch.org/fact-tank/2015/11/19/5-facts-about-illegal-immigration-inthe-u-s/ accessed January 3, 2015 5

would arise not only at a national level but as a local level as well where local governments can implement policies a ecting immigrants living within their jurisdiction. Aside from these policy implications the bottom line is that over the past decade America has become somewhat more open and accepting towards immigration. Why this has happened will be investigated in the rest of the paper. 4 Main E ects The following section will analyze opinions towards immigration across various racial, ethnic, religious, and social groups and determine which of these demographics is a main e ect on opinions towards immigration. The cross tabulation of a number of predictor variables with the outcome variable: attitudes towards how the number of immigrants coming into the country should change, leads to insights on factors which may be related to feelings towards immigrants. A chi-squared test ( 2 ) was used to test the significance of the predictor variables while phi ( ) in order to measure the e ect size of a significant variable. In order to analyze ordinal or interval variables t-tests were used to find the di erence of means while Cohen s D was used to measure the e ect size for those variables that were significant. Because I wanted to include a large number of e ects in the analysis, I set the alpha level at.10 instead of the customary.05. The table below represents a summary of the main effects analyzed. The following section will study in more detail the main e ects of the variables chosen while ruling out those variables which did not carry a statistically significant e ect. Table 4: Significant and Non-Significant Main E ects by E ect Size (weighted) Variable Test Higher percent favorable toward immigration if... P value E ect size Years of education t higher grade level 0 0.291 Age t younger 0 0.283 Socioeconomic status t higher socioeconomic status 0 0.276 Highest degree earned 2 more educated 0 0.211 Race 2 Non-white other 0 0.211 Religion 2 practicing Eastern religion 0 0.196 Political a liation 2 Democrat or third party 0 0.173 Region 2 From Mid Atlantic region 0 0.142 Class 2 upper class 0 0.115 Sex 2 Male.9 0.093 The table above indicates that a person s education, age, socio-economic status, race, religion, political a liation, and region of residence all hold considerable weight when it comes to forming opinions towards immigration. A 6

person s class and gender were also factors which influenced one s opinions towards immigration although these factors hold a relatively smaller e ect on the formation of those opinions. Looking at the significant main e ects, immigration is likely a class based issue for most people. A higher proportion of the more educated hold favorable opinions towards immigration, while younger upper class Americans tend to fall in the same boat. Favorable opinions towards immigration also seem to be concentrated in the Atlantic region of the United States. As far as religion goes, Eastern religions such as Hinduism, Buddhism, and Islam are religions in which a large percentage of people who practice them hold a positive opinion toward immigration. Other races besides black or white also seem to hold a larger percentage of favorable opinions toward immigration. Unsurprisingly Democrats are also a group which stands out in holding a larger percentage of responses seeking an increase in immigration. Finally, males compared to females are much more satisfied with the amount of immigration to the country. In terms of legislation these results are important for policy makers in areas where these demographics are prevalent. Policy makers in these areas could most likely push for further inclusive immigration reform, without expecting much public opposition. On a national scale, as more and more Americans are getting higher levels of education, and the amount of minorities is increasing, lawmakers should not be surprised to see more immigration related policy arise on their dockets. Given these projected demographic changes the nation could also start see policy rhetoric shift from being exclusionary to inclusionary. These main e ects could also have repercussions for non-legislative entities. Private institutions which advocate for inclusive immigration policies and look to further the country s knowledge on the issue should focus on appealing and reaching out to those who are less educated, older, and who live in lower class areas of the country. Community institutions such as schools in areas that hold a favorable view towards immigrants and immigration should prepare for more diversity within their systems and should contemplate the implementation of methods which focus on the teaching of inclusion, tolerance, and open mindedness towards di erent peoples and their cultures. The discussion that follows will go into a deeper analysis of each main e ect in order of e ect size. Education had the largest e ect size: There is a statistically significant di erence of roughly one year of schooling between those who held a favorable attitude towards immigration and those did not. Those who held a favorable opinion toward immigration had a mean of about 13 years of schooling while those who held an unfavorable opinion had a mean of roughly 14 years of schooling (t= -5.772, unequal variances, df=1566.3, p=.000). The di erence was significant with a 95% confidence interval from.88 to.45 years and a substantial e ect size of.291. When the education variable was changed to shift the focus to the degree one has earned a similar result with a smaller e ect size was discovered. 7

The degree earned could provide additional insights for policy analysis than the total number of years of schooling someone has attained. Below is a table examining a cross-tabulation between degree and opinion toward immigration: The number of immigrants to America nowadays should be... Table 5: Attitude Toward Immigration and Degree Degree High school College (2yr.) College (4yr.) Graduate Row Total Reduced a lot 1819 (29.9%) 241 (29.9%) 282 (16.3%) 126 (14.2%) 2468 Reduced a little 1530 (25.1%) 194 (24.0%) 434 (25.1%) 184 (20.7%) 2342 Remain the same 2062 (33.9%) 307 (38.0%) 765 (44.3%) 376 (42.4%) 3510 Increased a little 422 (6.9%) 44 (5.5%) 190 (11.0%) 165 (18.6%) 820 Increased a lot 258 (4.2%) 21 (2.6%) 56 (3.2%) 36 (4.1%) 371 Column Total 6091 807 1727 887 9512 A higher proportion of favorable opinions towards immigration is associated with higher levels of education. Comparing those with a high school education to those who had at least attended college, there is actually a drop in those who wish to see immigration increased; however, there is a five percentage point increase in those who are satisfied with immigration. When we look at the jump from high school to a four-year college education the story becomes a bit clearer. Not only is there a jump in the percentage of people who wish to see immigration increased, there is a leap of eleven percentage points for those who are satisfied with immigration. This trend continues for those with graduate education where the percentage of those who are satisfied with immigration stays roughly the same as the percentage for those with a college degree, but for those who wish to see immigration increased the graduate education group holds the highest percentage of respondents who wish to see this change. The pattern in the table holds major importance for people wondering why some people are still intolerant towards immigration. We could attribute the association between higher education levels and a higher proportion of favorable opinions toward immigration to the fact that these people simply hold more favorable opinions toward immigrants as well. Through college and postgraduate institutions, students become more exposed to cultural diversity while learning to respect other peoples and their ideas all the while seeing the value that these cultural disparities hold. This finding only further stresses the importance of higher education; however, it is not so much the academic aspect in which the importance lies but the environment which one is exposed to when attending these institutions. This rich environment is crucial for exposing people to di erent cultures and ideas. Although policy makers cannot force anybody to go to these 8

institutions, the institutions themselves must recognize the importance of their power to expose their students to diversity and their ability to help form a more accepting and tolerant citizenry. Another potential reason why this pattern exists could be that more educated people have more encounters with immigrants which could contribute to their favorable opinions, whereas those who are less educated could live in communities where they might interact with a di erent group of immigrants that lead them to forming their unfavorable opinions. As far as policy goes, because of this education e ect potential policies to increase the proportion of Americans who hold favorable opinions towards immigration should be aimed towards educational institutions. Seeing that higher education institutions are associated with a greater proportion of favorable opinions towards immigration other educational institutions such as high schools could try to replicate universities goals of increasing diversity within their institutions. One way this could be done is through the implementation and augmentation of programs which bring foreign students to high school campuses. Outside the realm of education, the private sector could a ect attitudes towards immigrants and immigration by implementing positive discrimination hiring practices (ex. A rmative Action) in order to diversify their work space. Aside from these two recommendations, perhaps some of the most influential policies that can be implemented today are policies that would incentivize communities to build low income housing in more a uent areas. Since a large number of immigrants are in fact living in poverty most of them end up in low income housing situations. But what exactly would this clash of socioeconomic classes do for cultural tolerance? The Pew Research Center, when studying relations between blacks and Hispanics found that Hispanics and blacks living in counties with relatively high concentrations of African Americans are somewhat more likely to say that blacks and Latinos get along well (65 percent Hispanics, 72 percent blacks) than are Hispanics and blacks living in low-density black counties (50 percent and 57 percent respectively), suggesting proximity is associated with greater acceptance. 6 Thus, by attempting to break down cultural boundaries and by making more communities a sort of a melting pot, policy could act as a catalyst in forming more favorable opinions towards new cultures and ideas and thus towards immigrants and immigration. Self-identified Class had the next largest e ect size: A person s socioeconomic status, which is an indicator of their class, was shown to have a statistically significant di erence of four on the socioeconomic status scale. A higher number on the socioeconomic status scale corresponds to a higher class placement. The people who held a favorable opinion toward immigration had a mean value of 51.82 on the socioeconomic scale while those who held an unfavorable opinion had a mean value of 48.09 (t= -4.3673, unequal variances, df=999.88, p=.000). The di erence was significant with a 95% confidence interval from roughly 5 to 2 points on the scale and yielded e ect size of.28. Since socioeconomic status indicates class, we can examine another variable to focus 6 http://www.pewsocialtrends.org/2008/01/31/do-blacks-and-hispanics-get-along/ 9

exclusively on a person s self-identified class. The number of immigrants to America nowadays should be... Table 6: Attitude Toward Immigration and Class Class Lower class Working Class Middle Class Upper class Row Total Reduced a lot 243 (36.5%) 1132 (27.1%) 1036 (23.8%) 47 (17.0%) 2458 Reduced a little 133 (20.0%) 1014 (24.3%) 1122 (25.8%) 65 (23.6%) 2334 Remain the same 205 (30.8%) 1526 (36.8%) 1658 (38.1%) 100 (36.2%) 3499 Increased a little 47 (7.1%) 316 (7.6%) 414 (9.5%) 39 (14.1%) 816 Increased a lot 38 (5.7%) 180 (4.3%) 121 (2.8%) 25 (9.1%) 364 Column Total 666 4178 4351 276 9471 As self-identified class status goes up the favorability towards immigration goes up as well. As far as increasing immigration goes, the percent of people in the lower and working class wish to see immigration increased is about half of the upper class percentage. This alone clearly indicates a class e ect in determining attitudes toward immigration. Even more convincing; however, is the fact that as socioeconomic status goes up from the lower class to the upper class the percentage of those wish to see immigration reduced by a lot decreases by a little more than half. Although these two observances are evidence of the class e ect there are some interesting exceptions. looking at the group who wishes to see immigration reduced by a little, the percentages stay stagnant across every class. A similar but less defined occurrence of this occurs for those who wish for immigration to stay the same. Overall, viewing socio-economic standing as a factor which forms opinions towards immigration hints that the e ects immigration and immigrants impose upon society a ect di erent classes in di erent ways, more specifically, a ecting the lower classes more than the middle/upper class. One possibility could be that people in the lower class prefer being around people of the same cultural background simply because they feel more comfortable in this environment. Another possibility could be that the middle/upper class favors immigrants more than the lower classes because the middle/upper class is not really threatened by immigrant presence in the labor market. Perhaps the low favorability of the lower classes towards immigration stems from the fact that they compete against immigrants for work. This question will be investigated further in the linear regression section. If immigrants, by entering the labor market, do in fact make those in lower classes worse o, this finding could be of use for groups that wish to limit immigration. On the other hand, if immigrants do not harm the lower class then this would serve to advance the pro-immigrant agenda. Another reason why this pattern is important is 10

because as time passes depending on how the economy grows if more people find themselves in a higher class the socio-economic e ect on opinions towards immigration will follow and more people should want more immigration; however, the opposite could occur as well. Table 7: Attitude Toward Immigration and Race The number of immigrants to America nowadays should be... Race White Black Other Row Total Reduced a lot 2077 (28.8%) 292 (23.3%) 99 (9.5%) 2468 Reduced a little 1879 (26.0%) 296 (23.6%) 166 (16.0%) 2341 Remain the same 2527 (35.0%) 461 (36.8%) 521 (50.2%) 3509 Increased a little 550 (7.6%) 119 (9.5%) 151 (14.6%) 820 Increased a lot 186 (2.6%) 85 (6.8%) 100 (9.6%) 371 Column Total 7219 1253 1037 9509 Blacks and whites both fell behind other races when comparing which race had the most favorable attitudes toward immigration. Other races had an overwhelmingly smaller percentage of respondents who wished to see immigration reduced, standing at roughly twenty-five percent while this same number was fifty-five percent for whites and forty-seven percent for blacks. When observing the category for those satisfied with the amount of immigration coming to the country blacks and whites stand at around thirty-five percent while other races stand much higher at fifty percent. Finally, when observing the categories for those who wish to see immigration increased whites stand at around ten percent, blacks at fifteen, and other races at twenty-three percent. The race e ect then as we can see is fairly robust. A potential explanation for the underwhelming support blacks and whites have for immigration could be attributed to negative day to day experiences these groups might experience with immigrants. Negative experiences with others leave lasting impressions on one s mind and this could be the situation for blacks and whites within communities that house large immigrant populations. Blacks more specifically are disadvantaged in American society and could view immigrants as an additional threat to their already tenuous position. The fact that other races besides white or black have such a large percentage of support for immigration might be because many immigrants find themselves in this race category and would most certainly recognize how they and their peers positively contribute to American society. More often than not these culture clashes perpetuate themselves due to lack of communication between the di erent parties. In a policy oriented sense, communities could foster strong relationships between immigrants and various race groups in order to increase support for immigration by implementing community programs which 11

attempt to include a diverse range of participants. This proposal is similar to what Robert D. Putnam puts forth in his paper E Pluribus Unum: Diversity and Community in the Twenty-first Century: To strengthen shared identities, we need more opportunities for meaningful interaction across ethnic lines where Americans (new and old) work, learn, recreate, and live. Community centers, athletic fields, and schools were among the most e cacious instruments for incorporating new immigrants a century ago, and we need to reinvest in such places and activities once again, enabling us all to become comfortable with diversity. 7 Religion had the next largest e ect size: Table 8: Religious Di erences in Opinions Towards Immigration (weighted) Religion Christian Jewish E. religion or Islam None Christian N/A.184*.428* 0.095* Jewish -0.184* N/A.244* -0.089 E. religion or Islam -0.428* -0.244* N/A -.333* None -0.095* 0.089 0.333* N/A * significant at a p-level of.05 In the table above only results significant at.05 have an asterisk placed beside them. An example in narrative form to interpret the significant di erences in the table: About 18.4 percent more Jewish people hold favorable opinions towards immigration than Christians, while this di erence widens to 42.8 percent when comparing Christians and those who practice a Islam or another Eastern religion. The pattern that arises consists of Eastern religions and Islam haveing the greatest amount of favorable opinions toward immigration, followed by Jews then those who do not practice a religion then finally Christians, holding the lowest amount of favorable opinions toward immigration compared to other religions. These di erences and the other significant di erences in the table were found using an analysis of variance to test for di erences of means across the eight di erent categories pertaining to a respondents religious preference (F = 50.6, df=3 and 5923, p= 0.00, eta = 0.02424 ). The Levene test of homoscedasticity showed that we must reject the hypothesis of equal variances, so we used the post hoc Tamhane method for interpreting results. Below is the cross tabulation between religion and attitudes toward immigation: 7 http://macaulay.cuny.edu/eportfolios/benediktsson2013/files/2013/04/putnam.pdf 12

The number of immigrants to America nowadays should be... Table 9: Attitude Toward Immigration and Religion Religion Christian Jewish E. religions & Islam None Row Total Reduced a lot 2035 (27.7%) 27 (16.6%) 14 (7.6%) 339 (20.7%) 2415 Reduced a little 1873 (25.5%) 40 (24.5%) 20 (10.9%) 369 (22.6%) 2302 Remain the same 2603 (35.5%) 63 (38.7%) 97 (52.7%) 679 (41.6%) 3442 Increased a little 552 (7.5%) 25 (15.3%) 39 (21.2%) 188 (11.5%) 804 Increased a lot 279 (3.8%) 8 (4.9%) 14 (7.6%) 59 (3.6%) 360 Column Total 7342 163 184 1634 9323 Religion has a relatively large e ect size compared to other variables. Almost all of those who practice Islam or an Eastern religion were satisfied with the current amount of immigration or wanted to see it increased. For those who do not practice any religion, around forty-one percent thought that the current amount of immigration should stay the same while roughly fifteen percent wanted to see it increased. Christians had a smaller percentage of those who wish immigration to increase or stay the same compared to the Jews who had twenty-one percent of respondents wanting immigration to be increased while thirty-eight percent wanted immigration to stay the same. The percentages for those who practice Islam or Eastern religions could be explained by the fact that most of the people that practice these religions in the U.S. are probably immigrants themselves and would logically favor more immigration. On the other hand, since many other immigrants come from many di erent religions, Americans belonging to certain religions, specifically those which we have seen are not so keen on immigration, might not like or want to associate with immigrants simply because belong to a di erent religion. This could be what is causing certain religious groups to have relatively low proportions of favorable opinions towards immigration. Further thinking leads to a new viewpoint on the issue of immigration. Most people would say that religions in general tend to preach tolerance and acceptance of others; consequently, one would expect higher percentages in the categories for increasing immigration. This could potentially point to the fact that many American s do not bring their religious beliefs into account when forming opinions on issues such as immigration. This could lead us to ask the question of which opinions are formed in part by taking into account one s religious beliefs. All this could mean that peoples stance on this issue might be more fundamental and foundational than previously thought. As far as policy 13

within religious institutions goes the truth seems to be that nothing much can be done if immigration is as much of a foundation issue as the numbers paint it to be. Table 10: Attitude Toward Immigration and Political A The number of immigrants to America nowadays should be... Political A liation liation Left Center Right Third Party Row Total Reduced a lot 922 (21.4%) 415 (25.0%) 1066 (32.0%) 45 (26.3%) 2488 Reduced a little 1019 (23.7%) 368 (22.2%) 914 (27.4%) 35 (20.5%) 2336 Remain the same 1713 (39.8%) 641 (38.7%) 1085 (32.5%) 55 (32.2%) 3494 Increased a little 447 (10.4%) 151 (9.1%) 194 (5.8%) 27 (15.8%) 819 Increased a lot 200 (4.7%) 82 (4.9%) 77 (2.3%) 9 (5.3%) 368 Column Total 4301 1657 3336 171 9465 liation and opinions toward im- Below is the cross tabulation for political a migration: Immigration without a doubt is a highly politicized issue. The left, namely the Democrats, tend to be more in favor of increased immigration while the right, namely republicans, would be on the opposite side of the issue. The data back up this assertion, with the left holding a considerably larger percentage of those who wish to see immigration increased. 15.1 percent of those who identify as on the left of the political spectrum wish to see immigration increased while for the right the number stands at roughly eight percent, a little less than half of what the left s percentage. The right also overwhelmingly supports the reduction of immigration. Exactly 59.4 percent of those who identify with the right in the sample believe that immigration should be reduced. As for those in the center of the aisle, the numbers are very similar to those on the left. The left; however, is not the group which is the most open to increases in immigration. Looking to the third party category we can observe that the support for increasing immigration is overwhelming compared to the other political a liations. This could be explained by the number of leftist third parties which have relatively sizable followings. Roughly six percentage points higher than the same figure for those on the left, 21.1 percent of third party respondents wished to see the amount of immigration to the U.S. increased. This gives rise to an important policy implication, especially for those on the left. If Democratic leaders wish to see future inclusive immigration reforms, appeals to independents and those belonging to third parties would be a solid strategy. Furthermore, the decade trends analyzed in a previous section point to the fact that attitudes which wish 14

to see an increase in immigration are on the rise. Therefore, the public should expect to see inclusive type reforms in the future, rather than exclusive ones such as those proposed by Republican candidate Donald Trump. Region had the next largest e ect size: Table 11: Regional Di erences in Opinions Towards Immigration (weighted) Region E. Coast N. Central S. Central W. Coast E. Coast N/A -.044* -.014.023 N. Central 0.044* N/A.030*.068 S. Central 0.014-0.030 N/A.038* W. Coast -0.023-0.068* -0.038 N/A * significant at a p-level of.05 In the table above only results significant at.05 are shown. An example in narrative form to interpret the table: About 4.4 percent less North central residents hold favorable opinions towards immigration compared to residents of the East coast, while about 2.3 percent more West coast residents hold favorable opinions towards immigration compared to their East coast counterparts. The rest of the table indicates that most regions are in fact similar when it comes to opinions towards immigration. The largest di erence arises between the West coast and the North central region of the country. The West coast houses 6.8 percent more favorable opinions compared to the North central region. These significant di erences were found using an analysis of variance to test for di erences of means across the four regions, (F =6.999, df=3 and 6040, p=0.0001072, eta =0.002969). The Levene test of homoscedasticity showed that we must reject the hypothesis of equal variances, so we used the post hoc Tamhane method for interpreting results once more. Below is the cross tabulation of region and attitudes toward immigration. By looking at the table it is clear that the West coast houses a much larger percentage of those who are approve of the current amount of immigration to the U.S. compared with the rest of the country. Furthermore, the West coast also contains the smallest percentage of those who wish to reduce immigration. On the other side of the country, the East coast seems to have the largest percentage of those who wish to see immigration increased. This percentage is 13.3 percent while the West coast does not fall far behind at 13 percent. Around thirty-five percent of inhabitants in the North and South central regions approve of the current state of immigration; however, when we compare these regions overall numbers to those who reside on the West or East coast, middle America tends to hold a less favorable view of immigration. The introduction of this paper demonstrated that most of this country would rather keep immigration 15

The number of immigrants to America nowadays should be... Table 12: Attitude Toward Immigration and Region Region East Coast North Central South Central West Coast Row Total Reduced a lot 996 (27.7%) 601 (27.6%) 451 (29.2%) 419 (19.1%) 2467 Reduced a little 867 (24.1%) 554 (25.4%) 368 (23.8%) 552 (25.1%) 2341 Remain the same 1251 (34.8%) 799 (36.7%) 521 (33.8%) 938 (42.7%) 3509 Increased a little 326 (9.1%) 171 (7.8%) 118 (7.6%) 205 (9.3%) 820 Increased a lot 151 (4.2%) 54 (2.5%) 85 (5.5%) 81 (3.7%) 371 Column Total 3591 2179 1543 2195 9508 the same or reduce it. The data shown above can only reinforce this claim. Since there seems to be clearly dominant side with respect to this topic, the question remains of why the country remains so divided on immigration related issues. So now we turn to why these regional patterns arise. Starting o with the West coast s dominance in being pro-immigration this could be potentially explained by the fact that the West coast houses a very large immigrant population. This explanation reinforces the claim made earlier in the paper which postulates that when people are more exposed to diversity and the e ects it disperses throughout society people tend to generally form more favorable opinions towards people from other cultures and backgrounds. This increased tolerance towards immigrants themselves could only lead to an increase in support for more immigration. This potential phenomenon is of great importance because since we do not know how immigrant populations will spread across di erent parts of the country in the future, we might soon be able to see more parts of the country begin to develop pro-immigrant and pro-immigration attitudes simply because immigrant populations have spread into their communities On the other hand the country could also become even more divided on the issue as immigrant populations might only concentrate in certain areas of the U.S. and not in others. On the next page is the cross tabulation for sex and attitudes toward immigration: 16

Table 13: Attitude Toward Immigration and Sex The number of immigrants to America nowadays should be... Sex Male Female Row Total Reduced a lot 1135 (26.0%) 1333 (26.0%)) 2468 Reduced a little 1062 (24.3%) 1279 (24.9%) 2341 Remain the same 1617 (37.0%) 1892 (36.8%) 3509 Increased a little 384 (8.8%) 436 (8.5%) 820 Increased a lot 175 (4.0%) 196 (3.8%) 371 Column Total 4373 5136 9509 When looking at sex we can observe that there are more men compared to women who hold favorable opinions towards immigration. 12.8 percent of men wish to see immigration increased while that number is 11.8 percent for women. For those who believe that immigration should stay the same, men stand at 37 percent while women fall just behind at 36.8 percent. Despite men holding a larger percentage of favorable opinions toward immigration compared to women, the di erences are not that large and thus the gender e ect, as shown by prior statistical tests, is not as strong as the e ect of other factors. The section above has described the main e ects pertaining to peoples attitudes towards how the number of immigrants entering the country should change and the unique importance of each main e ect. Next, this paper will analyze the interactions between the variables above in order to gain insight into the underlying e ects that they could hold for our investigation. By doing this we can avoid coming to misleading interpretations that could be harmful for our conclusions. 5 Interaction E ects Interactions between the variables listed above can occur, making some of the main e ects spurious. The table below lists interaction e ects between the main e ects discovered in the previous section and the variables education, race, age, and sex. We are controlling for education, race, age, and sex because these variables tend to be the ones which interact with other variables the most. 17

Table 14: Summary of Main and Interaction E ects (weighted) E ect persists when we control for... Variable E ect size Race Age Sex Degree Highest degree earned 0.211 persists for all persists for all persists for all N/A Race 0.211 N/A persists for all persists for all persists for all Religion 0.196 whites & blacks persists for all persists for all persists for all Political a liation 0.173 persists for all persists for all persists for all persists for all Region 0.142 other & white except for 65+ persists for all persists for all Class 0.115 persists for all persists for all persists for all persists for all Sex 0.093 does not persist persists for all N/A does not persist Controlling for race, age, education, and sex weakened the main e ect for religion, region, and sex. While the education, race, political a liation, and class e ects remained robust, the sex e ect became mostly spurious as a main e ect because it did not persist when race and education were controlled for. Thus, the finding that a lower proportion of females want to see immigration increased or stay the same compared to males is misleading. Looking down the control variable columns we can identify a couple of patterns. In terms of race almost every main e ect except for gender was significant for whites. This indicates that our main e ects interacted with race and also persist for whites. As far as controlling for gender and education the almost every main e ect persisted when controlling for these factors (with the exception of sex not persisting when controlling for education). When controlling for race religion persisted for whites only while regional e ects did not persist for blacks. Regional e ects also did not persists for those above age 65. These patterns can allow us to focus in on the actual groups for which the main e ects actually shape views towards immigration. As a whole, however; most all of the main e ects remained robust which signals that the patterns studied in the main e ects section still hold considerable weight in our study of what shapes Americans attitudes toward immigration. When looking at religion, its e ect, which implied that a lower percentage of Christians, Jews and those who do not practice any religion held favorable opinions towards immigration compared to those who practice an Easter religion or Islam, is partly spurious. For every age, gender, and education level this e ect persisted, however; when it comes to race the religion e ect did not persist for those who are of a non-black non-white race. That implies that, unlike the main e ect suggests, those in the other race category who do not practice Islam or an Eastern religion hold favorable opinions towards immigration at about the same proportion as those who are Christian, Jewish, or who do not practice any religion. There was a significant association however between religion and opinions towards immigration for white individuals, ( 2 )= 114.86, p=.000, for blacks, ( 2 ) = 36.6, p=.000. Let us examine the interaction between race and religion a little further. When constrained to whites, those who practice an Eastern religion or Islam had a roughly twenty-five percent 18

higher proportion of individuals who were in favor of increasing immigration or want it to stay the same compared to those who are Christian or Jewish. For blacks this di erence was only around ten percent. The most plausible explanation why the religion e ect persists for only blacks and whites stems from the interaction with the race e ect. The race e ect in our analysis posits that those in the other race category hold a much higher proportion of favorable and inclusive attitudes toward immigration compared to blacks and whites. Since this is true, no matter what religion someone in the other category belongs to, these Americans will always hold that immigration should be increased or remain the same. But why do they necessarily think this? Those who are in the other race category are most likely Hispanic or Asian, and according to the Immigration Research Group at The University of California Riverside the latter part of the 20th century saw... a heavy influx of undocumented immigrants from Latin America and Asia; therefore, most of these individuals or their grandparents or parents could be immigrants themselves and would most certainly agree that immigrants improve American society and would hold favorable attitudes toward immigration. 8 The cross-tabulation for race shows that this race category holds the highest proportion of favorable opinions towards immigration. This then is an explanation as to why those who are under the race classification of other have roughly the same proportion of people who hold favorable opinions towards immigration despite religion. This only further proves my earlier proposal that immigration is most likely a fundamental issue along with being a class based issue. The sex e ect, which shows that men on balance tend to be more open to increased immigration, should be taken with caution as this e ect did not persist when controlling for race or education. Thus, the sex e ect is mainly spurious. Now turning to region, we can see that the main e ect is not entirely robust. Although the e ect persists for both genders and for every level of education the e ect does not persist entirely for race or for age. The region e ect is not significant when we control for those who are black and for those who are 65 or older. Perhaps for those 65 or older immigration remains a fundamental issue; therefore, region should not a ect their opinions toward this issue. At a p-value of.5, the interaction between race and region when controlling for race reveals that the e ect does not persist for blacks. The percentages between regions when looking at only blacks seem to be very similar all across the board, regardless of whether someone wanted to increase or decrease immigration. There is no clear reason as to why the region e ect does not persist for blacks; however, one potential explanation could be that blacks as a whole have rigid opinions toward immigration that aren t a ected by the surroundings region that they live in. As far as policy goes, prior policy implications and recommendations made above should shift in a way that mainly focuses on blacks and whites instead of other races. If policy makers are trying to increase support for immigration 8 http://www.pewresearch.org/fact-tank/2015/11/19/5-facts-about-illegal-immigration-inthe-u-s/ 19