Duration of Representatives Incumbency in the U. S. House

Transcription

1 Duration of Representatives Incumbency in the U. S. House By Congress and by State First through 108 th Congress Average Aggregate Tenure in the U. S. House First through 108th Congresses Tenure in years by Congress Quantitative Historical Analysis #9 September 5, Contact: Quidam@thirty-thousand.org

2 Table of Contents Section 1 Overview... 1 Explanation of analysis and data sources Section 2 Illustrations... 4 Graphical illustrations of the data Section 3 Average Tenure per Congress Average aggregate tenure in the U. S. House Section 4 Conclusions Appendices Appendix 1 TTO s Congressional Database Appendix 2 Methodology for Calculating Tenure Appendix 3 Average Tenure by Congress thirty-thousand.org Personal use of this material is permitted as long as it is attributed to thirty-thousand.org. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from thirty-thousand.org. Thirty-thousand.org believes the data contained herein to be correct; however, we would greatly appreciate any errors being brought to our attention.

3 Section 1 Overview 1. OVERVIEW 1.1. Purpose This report provides the average duration of Representatives service in the U. S. House, by Congress, from the first to the 108 th Congress. The purpose is to provide an alternate metric for evaluating questions and issues related to elected representation as a long-term vocation. Throughout this report, the term tenure is used to describe the duration of the Representative s incumbency in office. Compared to reelection rates, tenure data provides a more precise measure of the incumbents tendency to persist in office. For example, of the 108 th Congress, 87.6% of the Representatives were reelected from the 107 th Congress. At the same time, the average tenure increased from 9.8 to 10.2 years (from the 107 th to the 108 th, respectively). The inherent limitation of reelection rates is that they evaluate each election as an isolated event rather than a sequence of related events (related because they are comprised of a recurring set of candidates). What is not apparent from the reelection data is how the outcome of one election may influence the next. What the tenure data measures that the reelection rates do not is what could be called tenure momentum: the longer an incumbent is in office, the more likely is he to seek and win reelection. Incumbency persistence in office is a function of both the incumbents desire to remain in office and their ability to win reelection. With respect to the latter, the high rate at which incumbents win reelection is generally attributed to a number of facilitating factors which taken together comprise what is known as the incumbency advantage. The incumbency advantage, and the factors which may comprise it, are outside the scope of this report Data Source In order to conduct these analyses, thirty-thousand.org (TTO) has developed a database which includes the actual dates of service for each Representative elected to the House since the first Congress. Typically, Representatives dates of service coincide with the beginning and end of the terms to which they were elected. However, there are many instances where a Representative s initial or final date of service actually occurs during a Congressional term. Whichever applies, TTO s congressional database calculates the total number of days served consecutively by each Representative.

4 Section 1 Overview A more detailed description of TTO s congressional database is provided by the Biographical Compilation of U. S. House Representatives or Compilation Methodology For this analysis, tenure refers to the uninterrupted duration of the Representative s service in office beginning with his or her initial date of service which, in many cases, precedes the beginning of any given Congress. This section provides a brief overview of the analysis methodology used; a more detailed explanation is provided in Appendix 2. As illustrated in the chart below, a high proportion of incumbents have been reelected from one Congress to the next ever since the first Congress. Consequently, a Representative s tenure in office often lasts two or more Congressional terms. 500 Reelected Incumbents as a Percentage of the Current Congress Second through 109th Congresses % Reelected Incumbent Prevalence (RIP %) (line graph) % % The adjusted size of the House (column graph) and the number of Representatives reelected from the previous Congress (shaded area) US-E Chart A Reelected Incumbents in the House thirty-thousand.org 40% 20% 0% The analysis of reelection rates summarized in the chart above was provided by the previous report in this series. 2 1 The Compilation provides biographical data on all Representatives elected to the House since March of For additional information about the Compilation see Appendix 1. 2 This chart is from the Historical Prevalence of Reelected Representatives in the U. S. House ( 07.htm), which reports the reelection rates of the U. S. House of Representatives from 1791 to 2005 (by Congress and by state). The Reelected Incumbent Prevalence (or RIP ) is the percentage of Representatives, in any given Congress, who also served in the preceding Congress.

5 Section 1 Overview Incumbency tenure can be easily explained by example: a Representative elected to his fifth consecutive term begins that Congress with 8 years of tenure and, assuming he completes the fifth Congress, ends it with 10 years of tenure. Therefore, during the Representative s fifth Congress his average tenure is 9 years (the average of 8 and 10). The aggregate average tenure for any given Congress is the average of all tenures of all Representatives who served during that Congress, regardless of whether their tenure in office was two weeks, two years or two decades. The aggregate average tenure was determined for a cross section of dates (for the entire population of Representatives serving on those dates) throughout each Congress. The resulting set of tenure values was then averaged to produce a single value for the entire Congress. This analysis was done for each Congress by state and for the aggregate U. S. Because the average is derived from a cross section of dates during each Congress, it approximates the aggregate tenure as of the mid-point of the Congress, rather than at its conclusion. For the purposes of making relative comparisons across Congresses over time this mid-point average is as reliable as using an end-point value. Moreover, for reasons explained in Appendix 2, the mid-point value is more statistically reliable than simply taking the end-point (or any other single point) value. However, the mid-point value will tend to be approximately one year shorter than a similar tenure value calculated at the end of the same Congress (since the term is two years long). The resulting data is illustrated by the charts in the following section. A summary of the data is provided in Section 3, with greater detail provided in Appendix 3.

6 Section 2 Illustrations 2. ILLUSTRATIONS. Tenure of Representatives in the House As explained above, tenure refers to the Representatives total duration in office beginning with their initial dates of service which, in most cases, precedes the beginning of any given Congress. Chart B (below) illustrates the average aggregate tenure, in years, for all Representatives in each Congress. Average Aggregate Tenure in the U. S. House First through 108th Congresses Tenure in years by Congress Chart B Tenure in the U. S. House Each vertical bar in the chart above represents the average actual tenure for all Representatives who served during in that particular Congress. For example, the average tenure in the 93 rd Congress was nearly ten years (9.9 years to be exact). Because this is the average for the entire Congress, it actually represents the midpoint. 3 As required by the Constitution, every ten years a new apportionment is imposed upon Congress based on the results of the decennial population census. The reapportionments effectively reset the allocation of House seats and, as shown in the chart above, often results in a new ten-year tenure profile which may resemble a rank of pipe organs. Identifying these apportionment regimes is critical to evaluating the tenure data, especially with respect to understanding the drop in tenure which sometimes occurs at the beginning of a new regime. For example, the 15 th apportionment defined how the 73 rd through 77 th Congresses were comprised. This reapportionment resulted in an average tenure decline from 7.8 to 5.9 years (for the 21 3 Adding a year (raising it to 10.9 years) would provide an approximation of the tenure as of the end of the 93 rd Congress. As explained in Appendix 2, taking the average over the duration of the Congress avoids wide variations that may occur within shorter timeframes.

7 Section 2 Illustrations 72 nd and 73 rd Congresses, respectively). 4 In this case, the tenure decline is largely due to the transfer of 27 House seats from one set of states to another. 5 The chart below is identical to the one above except that, in order to provide historical context, the prevailing presidential administrations are also identified. George Washington John Adams Thomas Jefferson James Madison James Monroe John Quincy Adams William H. Harrison 1841 John Tyler Andrew James Jackson Polk Martin Van Buren James Buchanan Abraham Lincoln Andrew Johnson Ulysses S. Grant James Garfield 1881 Chester Aruthur Grover Cleveland Benjamin Harrison William McKinley Theodore Roosevelt William H. Taft Woodrow Wilson Warren Harding Calvin Coolidge Herbert Hoover Franklin D. Roosevelt Harry S. Truman John F. Kennedy Lyndon Johnson Dwight Eisenhower Richard Nixon Gerald Ford Ronald Reagan George H. W. Bush William J. Clinton George W. Bush Zachary Taylor Millard Fillmore Franklin Pierce Rutherford B. Hayes Grover Cleveland Jimmy Carter Chart C Tenure in the U. S. House with Presidencies Identified 4 Section 3 provides the aggregate U. S. tenure values. 5 Historically, because reapportionments have usually resulted in relatively small net increases in House seats, their primary impact has been to effect a zero-sum reallocation of existing House seats. To better understand how reapportionments reallocate seats, see Section 2.5 (Chart E) of The Size of the U. S. House of Representatives and its Constituent State Delegations which is available at (

8 Section 2 Illustrations. Tenure and Reelection Rates The chart below is identical to chart B (from the preceding section) except that it also includes a line graph illustrating the percentage of incumbents in the House who had also served in the previous Congress (RIP %). 6 Tenure and Incumbent Prevalence in the U. S. House First through 108th Congresses Reelected Incumbent Prevalence in the U. S. House 100% 80% 60% Tenure in years by Congress 40% 20% % Chart D Tenure vs. Reelected Incumbent Prevalence Note the apparent similarity between these two measures with respect to their underlying trends over the long run. As expected, tenure tends to increase along with the reelection rates. However, though they are both biennial measures, a reelection is simply a binary (yes/no) event while the underlying tenure data better resembles a time continuum. As a result, note how much more erratic the reelection values are in comparison to tenure. The relative stability of the tenure data results from tenure momentum; i.e., the longer an incumbent remains in office the less likely is he or she to be involuntarily removed. Because non-reelected incumbents are more likely to be those with shorter tenure, their elimination has a relatively smaller impact on the aggregate tenure average. Consequently, during periods of declining reelection rates, the tenure values often remain relatively stable (or even increase). This can be seen during the 20 th apportionment (Congresses 98 through 102). The effect of tenure momentum is most evident when examining the data associated with the first Congress of each new apportionment regime when, 6 The RIP% graph is also illustrated in Chart A (Section 1) of this report. For a complete description of the Reelected Incumbent Prevalence (RIP) and how it was calculated see the Historical Prevalence of Reelected Representatives in the U. S. House at

9 Section 2 Illustrations depending on the number of House seats eliminated, the decennial reapportionment usually forces a reduction in the reelection rates. However, the corresponding impact on tenure is not as predictable. Presumably this is due to tenure momentum since, in those states where seats are eliminated, the long-tenured alpha incumbents are likely to displace incumbents with shorter tenure. Not surprisingly, the efficacy of the incumbency advantage appears to be a function of tenure in office. As a result, major changes in the tenure continuum usually only occur in response to significant external events (e.g., a material reapportionment or a seismic political shift). Conversely, there are specific instances where the average tenure declines even though the percentage of reelected incumbents increases. Though this may appear contradictory, these instances can be understood when the underlying data is examined. 7 Not withstanding the foregoing, it is still apparent that there is a correlation between these two data sets. In fact, as explained below, the coefficient of correlation is.91 so, as would be expected, these two measures fully corroborate one another. Given the high correlation between the RIP% and the average tenure, these two measures may appear to be a comparable basis for statistical analyses of the incumbency advantage. However, as is shown below, tenure values provide a somewhat more reliable metric than do reelection rates. 7 For example, though the RIP % increases from Congress #94 to #95 (from 79% to 85% respectively), the corresponding average tenure declines (from 9.2 to 8.7 years, respectively). In this case, a relatively large number of high-tenure Representatives did not seek reelection to the 95 th or resigned (or passed away) early in the 95 th Congress.

10 Section 2 Illustrations. Correlation between Tenure vs. Reelection Rates Like the chart above, the chart below graphs: the average tenure in years (left Y axis); and, the percentage of incumbents in the House who had also served in the previous Congress (right Y axis). 12 Tenure and Incumbent Prevalence in the U. S. House Second through 108th Congresses % Reelected Incumbent Prevalence in the U. S. House Aggregate Average Tenure in years by Congress 100% 80% 60% 40% 20% 0 0% Chart E Tenure vs. Reelected Incumbent Prevalence The coefficient of correlation between these two sets of data is.91 (i.e..9188). Despite this high correlation, the tenure value appears to be a somewhat more reliable metric than the RIP%. This point is illustrated by the chart below Tenure and Incumbent Prevalence in the U. S. House Second through 108th Congresses Reelected Incumbent Prevalence in the U. S. House y = -3E-11x 6 + 1E-08x 5-2E-06x x x x R 2 = % 100% 80% 60% 40% y = -1E-10x 6 + 6E-08x 5-1E-05x x x x R 2 = Chart F Tenure vs. Reelected Incumbent Prevalence with trend lines Aggregate Average Tenure in years by Congress 20% 0%

11 Section 2 Illustrations In the second chart, a trend line has been added to each graph (represented by a dashed line). For each of the two data series, the smaller the dispersion of data points around the trend line, the more predictable are the data values based on the regression equation describing the trend line. Based on a visual inspection of the second chart, it would appear that the Reelected Incumbent Prevalence (RIP%) has a wider dispersion around its trend line than does the tenure graph. Fortunately, the actual extent of this dispersion around the trend line can be determined mathematically. This value, called R 2, is provided in the chart for each of the data sets. 8 The closer R 2 is to one (), the more predictive is the trend (regression) equation likely to be. In this case, the R 2 values for tenure and RIP% values is.95 and.84, respectively. 9 This result is consistent with the fact that the reelection rate data appears relatively inerratic compared to the tenure data and, moreover, indicates that the tenure values should provide somewhat better statistical predictability than the reelected incumbent prevalence. 8 R2, or R squared, is a statistical measure of how well a regression line approximates real data points; an R 2 of (100%) indicates a perfect fit. By definition, R 2, is the fraction of the total squared error that is explained by the model, or regression equation. In this case, each of the data sets was (independently) regressed over time (against Congresses 2 through 108). 9 For this analysis, the trend line was formulated as an Order 5 polynomial equation, the equation for which is shown in the chart for each data set. There are other methods which could be used to derive the trend lines and, depending on which method is used, the resulting R 2 value will change. For both of these data sets, trend lines were computed using several different methods and, in all cases tested, the R 2 value for the tenure values was always higher than the R 2 for the RIP% values.

12 Section 3 Average Tenure per Congress 3. AVERAGE TENURE PER CONGRESS The table below provides the average tenure in years for each Congress (for the aggregate U. S.). This data is also provided for each state in Appendix 3. Average Tenure by Congress (years) First through 108th Congresses Apportionment Regimes (1-22) C Congress (year) 1 (1789) 2 (1791) District Population 1 Congress (year) 3 (1793) 4 (1795) 5 (1797) 6 (1799) 7 (1801) District Population Congress (year) 8 (1803) 9 (1805) 10 (1807) 11 (1809) 12 (1811) District Population Congress (year) 13 (1813) 14 (1815) 15 (1817) 16 (1819) 17 (1821) District Population Congress (year) 18 (1823) 19 (1825) 20 (1827) 21 (1829) 22 (1831) District Population Congress (year) 23 (1833) 24 (1835) 25 (1837) 26 (1839) 27 (1841) District Population Congress (year) 28 (1843) 29 (1845) 30 (1847) 31 (1849) 32 (1851) District Population Congress (year) 33 (1853) 34 (1855) 35 (1857) 36 (1859) 37 (1861) District Population Congress (year) 38 (1863) 39 (1865) 40 (1867) 41 (1869) 42 (1871) District Population 9 Congress (year) 43 (1873) 44 (1875) 45 (1877) 46 (1879) 47 (1881) District Population Congress (year) 48 (1883) 49 (1885) 50 (1887) 51 (1889) 52 (1891) District Population Congress (year) 53 (1893) 54 (1895) 55 (1897) 56 (1899) 57 (1901) District Population Congress (year) 58 (1903) 59 (1905) 60 (1907) 61 (1909) 62 (1911) District Population Congress (year) 63 (1913) 64 (1915) 65 (1917) 66 (1919) 67 (1921) District Population Congress (year) 68 (1923) 69 (1925) 70 (1927) 71 (1929) 72 (1931) District Population Congress (year) 73 (1933) 74 (1935) 75 (1937) 76 (1939) 77 (1941) District Population Congress (year) 78 (1943) 79 (1945) 80 (1947) 81 (1949) 82 (1951) District Population Congress (year) 83 (1953) 84 (1955) 85 (1957) 86 (1959) 87 (1961) District Population Congress (year) 88 (1963) 89 (1965) 90 (1967) 91 (1969) 92 (1971) District Population Congress (year) 93 (1973) 94 (1975) 95 (1977) 96 (1979) 97 (1981) District Population Congress (year) 98 (1983) 99 (1985) 100 (1987) 101 (1989) 102 (1991) District Population Congress (year) 103 (1993) 104 (1995) 105 (1997) 106 (1999) 107 (2001) District Population Congress (year) 108 (2003) 109 (2005) 110 (2007) 111 (2009) 112 (2011) District Population 10.2

13 Section 4 Conclusions 4. CONCLUSIONS Incumbency tenure in office is ultimately a function of a multitude of unpredictable events and human decisions that defy modeling. Nonetheless, it is still possible to evaluate the historical data in search of trends and causal relationships. As shown in the charts above, the profile produced by the tenure analysis closely resembles that of reelection rates, so in recent years average tenure in office has been climbing along with the reelection rates. As would be expected, there is a high coefficient of correlation between incumbency tenure and reelection rates; however, they are not redundant measures. Whereas the reelection rates treat elections as a series of unrelated events, the tenure data shows the net cumulative result of a sequence of elections where incumbency entrenchment is a factor. Though both measures are frequently reduced by the decennial reapportionment (depending on the number of House seats reassigned among the states), the resulting decremental impact on reelection rates is usually much greater then that on tenure. This can be attributed to the fact that the longer an incumbent is in office, the more likely is he to seek and win reelection (due to tenure momentum); consequently, whenever the number of House seats has been reduced, senior incumbents usually displace junior incumbents. As with reelection rates, there is a natural upper boundary on tenure in office. In the case of reelection rates, the upper boundary is that no more than 100% of incumbents can be reelected. The upper boundary for tenure is a function of the human lifespan so, while it cannot be specifically quantified, it is just as real. Moreover, tenure in the House of Representatives is ultimately limited by the number of productive years which may be available after the age of 25. However, other limiting factors frequently intervene prior to infirmity or mortality. If not for that, the Representatives tenure in office might resemble that of Supreme Court justices. Therefore, assuming the continuation of high reelection rates, it can be imagined that tenure s trend line asymptotically approaches some natural upper limit. With respect to reelection rates and tenure, the importance of these asymptotical limitations is twofold. First, it provides context for evaluating what appears to be relatively smaller changes as the data approaches its natural limits. Second, it reminds us that projections can not be reliably extrapolated from the historical data as they may extend beyond these natural limitations.

14 Appendices

15 Appendix 1 TTO s Congressional Database Appendix 1 TTO s Congressional Database TTO s congressional database contains every instance of a Representative being elected to a Congress. This data can be aggregated to produce various statistical analyses such as the ones done for this report. In the current data set, there are 10,508 Representatives across 36,223 instances of a Representative being elected to Congress. The individual tenure durations range from zero to 53 years. 10 The content of TTO s congressional database is provided by the Biographical Compilation of U. S. House Representatives, 11 or Compilation. The Compilation is comprised of all persons who were elected to the U. S. House of Representatives between March of 1789 (the First Congress) and December 31, 2005 (mid way through the 109 th Congress). For each House Representative, the Compilation provides a brief biography and the following data: The state from which each was elected The Congress(es) to which each was elected The political party affiliation at the time of election (if known) Dates of service for each Representative, also the reason for early termination of service (e.g., due to death or resignation) when applicable When elected by a special election (in order to fill a prematurely vacated seat), the name of the predecessor is provided Nominal Tenure in office the number of consecutive Congresses to which a Representative has been elected Actual Tenure in office the cumulative number of days served consecutively without interruption 10 Zero tenure occurs when an elected Representative subsequently does not actually serve his or her term. The longest tenure was achieved by Jamie Lloyd Whitten (Mississippi) who served over 53 years between November 4, 1941, and January 3, Additional information on the Biographical Compilation of U. S. House Representatives can be found at:

16 Appendix 2 Methodology for Calculating Tenure Appendix 2 Analysis Methodology A) Tenure Calculation For this analysis, tenure refers to the Representative s uninterrupted term in office beginning with his or her initial date of service which, in many cases, precedes the beginning of any given Congress. Normally, the Representative s initial and final dates of service are determined by the start and completion dates of those Congresses to which they were elected. However, there are numerous cases where a Representative entered or left office during the two-year congressional period. Whatever the case, the tenure values used for this analysis were calculated from the Representatives actual beginning and ending dates of service. The figure to the right illustrates some of the various scenarios which may occur. The aggregate average tenure for any given Congress is the average of all tenures of all Representatives who served during that Congress, regardless of whether they were in office for two weeks, two years or two decades. The tenure computation excludes those instances of null tenure which occur when a newly-elected person does not subsequently take their seat in the House. (In these cases, some intervening decision or event has precluded them from occupying office.) In order to determine the aggregate average tenure for each Congress, it was first calculated for a set of dates throughout each Congress. For the period between March 4, 1789 and December 31, 2005, the average tenure (of all of the Representatives) was determined as of the 1 st and 16 th of every month. 12 For each Congress, this results in 48 sampling dates 13 (plus a partial sampling for the 109 th 12 This computational methodology would cause to be excluded those instances, if any, where a Representative s initial and termination dates both occur within any single interval between two successive sampling dates. 13 The only exception being the 73 rd Congress which had a shorter duration due to the new schedule imposed by the 20 th Amendment.

17 Appendix 2 Methodology for Calculating Tenure Congress). This analysis was done for each state and for the aggregate U. S. (by Congress). The illustration below provides, for a portion of the 108 th Congress, a sample view of the resulting data. Tenure is shown in days. Examples from the table above: The average tenure of the Alaska delegation in office as of 4/16/2003 was 10,998 days The average tenure of the Florida delegation in office as of 4/16/2003 was 3,119 days. In the chart below, each Congress is represented by its own tenure graph because aggregate tenure in the House of Representatives is actually a series of discrete (i.e., discontinuous) events. Each election results in a partial turnover in the House which is reflected by a new graph with an initial tenure lower than the tenure at the end of the preceding Congress. Because of the large number of Representatives who are reelected from the preceding Congresses, the initial tenure is always greater than zero. Over the course of each two-year term the average tenure increases nearly linearly; however, perfectly linear plots are prevented by intervening events such as the resignations or deaths of incumbents.

18 Appendix 2 Methodology for Calculating Tenure Tenure in the U. S. House First through 108th Congresses , , ,653 3,287 2,922 2,557 2,192 1,826 1,461 1, Chart G Tenure in the U. S. House by Congress The set of values for all the sampling dates (from each Congress) is then averaged to produce a representative aggregate value (for each Congress). This value is converted to years. 14 As further illustrated in the legend below, for each Congress in the chart, the average is indicated on each graph by a yellow diamond. For example, the average tenure for the 103 rd Congress was 9.1 years. As can be seen in the chart, this tenure value approximates a mid-point value. For the same Congress, the aggregate tenure is nearly 10 years at term end. Consequently, this tenure value is approximately one year shorter than a similar tenure value calculated at the end of that Congress (since the term is two years long). Because this average is computed from a cross section of dates across the duration of each Congress, it approximates the aggregate tenure as of the midpoint of the Congress rather than at its conclusion. For the purposes of making relative comparisons across Congresses over time the 14 For the purposes of this analysis, a year is equal to days. This is based on a four-year average of 3 years at 365 days plus a leap year.

19 Appendix 2 Methodology for Calculating Tenure average value is as reliable as using any other. However, the mid-point value will tend to be approximately one year shorter than a similar tenure value calculated at the end of a Congress (since the term is two years long). The average (or midpoint) value is more statistically reliable than simply taking the end-point (or any other single point) value. If the aggregate average tenure were determined for a single point of time (e.g., at the end of a Congress), problems can result due to the inherent variability of data that describes a small number of persons. For example, a Representative who has been in office for many years may have resigned, or passed away, near the end of their final congress. Such variations can produce a less reliable data. Consequently, in order to produce values which are reliably comparable across multiple Congresses, the tenure average is determined for the entire two-year Congressional term. The chart below is similar to Chart G (above) except that it only shows the average value (rather than the entire graph) for each Congress. It is this sequence of values that is provided by this report. 12 Tenure in the U. S. House First through 108th Congresses Aggregate Average Tenure in years by Congress Chart H Average Tenure in the U. S. House B) Disregarding the Dynastic Effect The dynastic effect refers to instances where someone is elected to succeed a relative (usually a father or a husband). This appears to occur most frequently as a result of the death or resignation of the predecessor prior to the completion of the session to which he or she had been elected. If intra-family successions were treated as a continuous tenure event (as with a dynasty), it would increase the overall average tenures of service.

20 Appendix 2 Methodology for Calculating Tenure Based on the data provided by the Compilation, approximately 1,500 House Representatives appear to have had some type of familial connection to another Representative (since the first Congress). This represents approximately 10% of all persons ever elected to the House. The total number of instances in which an elected Representative was directly related to his or her immediate predecessor was not determined as part of this analysis. 15 However, since the first Congress, at least 58 persons have been elected to the House to replace an immediate relative as a result of a special election. A few examples are shown in the table below: Representative State Congresses Successor Congresses George Joseph Bates MA Son Chester Castle Bolton OH Wife John David Dingell MI Son For the purposes of this report, the dynastic effect was not taken into account. 15 It was also not apparent whether or not there is a trend (over time) with respect to the frequency of dynastic successions or if this is a relatively random event.

21 Appendix 3 Average Tenure by Congress Appendix 3 Average Tenure by Congress The following tables provide the average tenure values in years for each state and for the total U. S., by Congress. As explained in the Methodology section, tenure refers to the Representatives uninterrupted term in office based on their actual dates of service. For additional information on how these values were computed, be sure to read the methodology section of this report.

22 Congress Start & End Dates Appendix 3 Average Tenure by Congress (years) Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Congressional Year Apportionment CONSTITUTION 1 2 Total U. S State: Admission Alabama Dec. 14, 1819 Alaska Jan. 3, 1959 Arizona Feb. 14, 1912 Arkansas June 15, 1836 California Sept. 9, 1850 Colorado Aug. 1, 1876 Connecticut Jan. 9, Delaware Dec. 7, Florida Mar. 3, 1845 Georgia Jan. 2, Hawaii Aug. 21, 1959 Idaho July 3, 1890 Illinois Dec. 3, 1818 Indiana Dec. 11, 1816 Iowa Dec. 28, 1846 Kansas Jan. 29, 1861 Kentucky June 1, Louisiana Apr. 30, Maine Mar. 15, 1820 Maryland Apr. 28, Massachusetts Feb. 6, Michigan Jan. 26, 1837 Minnesota May 11, 1858 Mississippi Dec. 10, 1817 Missouri Aug. 10, 1821 Montana Nov. 8, 1889 Nebraska Mar. 1, 1867 Nevada Oct. 31, 1864 New Hampshire June 21, New Jersey Dec. 18, New Mexico Jan. 6, 1912 New York July 26, North Carolina Nov. 21, North Dakota Nov. 2, 1889 Ohio Mar. 1, Oklahoma Nov. 16, 1907 Oregon Feb. 14, 1859 Pennsylvania Dec. 12, Rhode Island May 29, South Carolina May 23, South Dakota Nov. 2, 1889 Tennessee June 1, Texas Dec. 29, 1845 Utah Jan. 4, 1896 Vermont Mar. 4, Virginia June 25, Washington Nov. 11, 1889 West Virginia June 20, 1863 Wisconsin May 29, 1848 Wyoming July 10, 1890 thirty-thousand.org September 5, QHA #9 Appendix 3

23 Appendix 3 Average Tenure by Congress (years) Congress Start & End Dates Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Congressional Yea Apportionmen Total U. S State: Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming thirty-thousand.org September 5, 2006 QHA #9 Appendix 3

24 Appendix 3 Average Tenure by Congress (years) Congress Start & End Dates Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Congressional Yea Apportionmen Total U. S State: Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware 0.6 Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas 1.1 Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nebraska Nevada 0.2 New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming thirty-thousand.org September 5, 2006 QHA #9 Appendix 3

25 Appendix 3 Average Tenure by Congress (years) Congress Start & End Dates Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Congressional Yea Apportionmen Total U. S State: Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware 0.6 Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana 0.7 Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington 0.7 West Virginia Wisconsin Wyoming thirty-thousand.org September 5, 2006 QHA #9 Appendix 3

26 Appendix 3 Average Tenure by Congress (years) Congress Start & End Dates Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Mar-04 Mar-03 Congressional Yea Apportionmen Total U. S State: Alabama Alaska Arizona Arkansas California Colorado Connecticut Delaware 7.0 Florida Georgia Hawaii Idaho Illinois Indiana Iowa Kansas Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nebraska Nevada New Hampshire New Jersey New Mexico New York North Carolina North Dakota Ohio Oklahoma Oregon Pennsylvania Rhode Island South Carolina South Dakota Tennessee Texas Utah Vermont Virginia Washington West Virginia Wisconsin Wyoming thirty-thousand.org September 5, 2006 QHA #9 Appendix 3