The Jewish vote in Lexington, Kentucky, for the 2000 presidential election

By Ben Gerdts and Josh Adland

Disclaimer:

This study was done in an AP Statistics course with relatively small sample sizes. The validity of such studies must always be questioned. Please keep this in mind if you use or report the results of this study.

Abstract:

Since Jews started to immigrate to this country en masse in the early 1900’s, they have traditionally voted overwhelmingly for Democratic candidates, especially in presidential elections.  Since 1972, no presidential candidate other than Jimmy Carter has received less than 64% of the Jewish vote, while no Republican candidate received more than 40%^[1].  Though Jews make up only 1-2% of the American population, their vote is extremely important because it usually constitutes about 4% of the votes cast in any election.  In addition, because Jews are among the more successful (both financially and educationally) ethnic groups in this country, their campaign contributions are also vigorously sought.

            Recently, however, there has been speculation that the general Jewish political standing in this country is beginning to shift more towards the center, increasing the numbers of Jews choosing to vote for Republican candidates.  These changes are not taking place all at once, and different Jewish communities in different parts of the country are shifting towards the right at different rates.

            This following study is designed to test the political standing of the Jewish community in Lexington, Kentucky, based on how individuals voted in the 2000 presidential election.  It aims to answer the following two-pronged question: Is there a statistically significant difference between the way the members of the two synagogues in Lexington voted in the 2000 presidential election, and did Jews in Lexington vote similar to Jews all over the country in that election?  The study aims to examine whether a difference exists between the populations or not. 

The null hypothesis is expected to hold in the first test because in this city in particular, there is not a large difference in the religious observance, which can often influence political standing, between the two synagogues.  Also, the reasons people join each of these synagogues does not always reflect their religious observance.  Social and economic factors help determine which synagogue people will join.  The null hypothesis is expected to hold in the second outcome, that Jews in Lexington vote the same at Jews on the national level, because this community has Jews with a wide range of influences, and is hardly homogeneous.

            The four populations included in this study are Temple Adath Israel, Ohavey Zion Synagogue, the combination of the previous two, and all of American Jewry.  For the first experiment, a random sample will be taken from each of the synagogues.  Subjects will be asked for whom they voted in the election of 2000 (Bush, Gore, other).  The “other” category will further be investigated.  People whose votes went to Ralph Nader will be counted in the “Gore” category, and people who voted for Buchanan will be counted in the “Bush” category.  This grouping is based on those independent candidates place in the political spectrum.  Nader is more liberal as Buchanan is more conservative, as a Democrats and Republicans respectfully.  Any other subjects will be discarded.  The results will then be compared in a Chi-Squared test to determine if there is a significant difference.  The combined totals for each candidate will then be compared to the proportion of Jews voting for those candidates on the national level (based on a CNN/Voter News Service poll).  Another Chi-Squared test will be performed to determine whether there is a significant difference between those two samples.

Methodology:

            This study aims to answer the question: Is there a significant difference between the political standing of the members of the two synagogues in Lexington, and do Jews in Lexington vote similar to Jews all over the country? 

Sampling

The population of Temple Adath Israel, a Reform Jewish organization, is 330 families.  Based on an alphabetical list, each household will receive a number.  A random number generator will designate a list of 30 households for this sample (included).  Each of the 30 households will be called and the sampler will ask to speak to a registered voter.  The voter will be asked if they voted in the 2000 election.  If they did, they will be asked if they voted for Gore, Bush, or other (a further question will be asked in order to define “other”).  If not, the sampler will go on to the next subject.  Because the study seeks a sample of thirty, subjects will be chosen by the random number generator until the total votes for Gore, Bush, and other total 30.

The population of Ohavey Zion Synagogue, a Conservative Jewish organization, is 170 families.  The same procedure will be applied to this synagogue as was applied to TAI, but only 15 samples will be sought.

PHONE CALL SCRIPT:

SUBJECT: Hello?

CALLER: May I please speak to Mr. or Mrs. Goldfarb?

SUBJECT: This is them.

IF NO

CALLER: Thank you.   Have a nice day.

IF YES

CALLER:  Thank you. Did you vote in the 2000 presidential election?

IF NO

CALLER: Thank you, have a nice day.

IF YES

CALLER: Did you vote for George W. Bush, Al Gore, or other?

SUBJECT answers, record result

Reducing Bias

            Because one’s past voting record is something most people are not afraid to talk about, no real bias can be introduced into this study.  The fact that the data collector and the subject may know each other will not in any way affect the results of the study.  Also, the question was asked in such a way by the caller as not to allude to a certain candidate.  These measures were the main steps taken to reduce bias.

Exploratory Data Analysis:

Note: The only “other” votes were for Nader and will be counted in the Gore category.  More than 95% of all subjects voted for Bush or Gore.

Data Tables

Table 1 – Comparison of voting record of Jews at both synagogues in Lexington for 2000 presidential election.

Graph 1

Table 2 – Comparison of voting record of Lexington Jews versus American Jews in the 2000 presidential election.

Graph 2

Statistical Inference:

Test 1

            To analyze the data from Test 1, which compared the voting records of Jews at each of two synagogues in Lexington, Kentucky, a chi-squared test must be performed.  For this test to be completely valid, two assumptions must be met:

v     All expected counts are greater than 1 – This condition is met.

v     No more than 20% of expected counts are less then 5. – This assumption is not met, but only one out of the four counts is less than 5, making it cautiously acceptable.

H₀: There is no difference between how Jews from each of the two synagogues in Lexington voted in the 2000 presidential election.

H_a:  There is a statistically significant difference between how Jews from each of the two synagogues in Lexington voted in the 2000 presidential election.

Below is the Chi-Squared significance test for this comparison:

Minitab Output for Temple Adath Israel vs. Ohavay Zion Synagogue   Expected counts are printed below observed counts             Bush  Gore/Other  Total   TAI        9       21       30           8.67    21.33     OZS        4       11       15           4.33    10.67   Total       13       32       45   Chi-Sq =  0.013 +  0.005 +           0.026 +  0.010 = 0.054   DF = 1, P-Value = 0.816   1 cells with expected counts less than 5.0

 

Test 2

            To analyze the data from Test 2, which compared the voting records of Jews all over the country to that of Jews in Lexington, Kentucky, a chi-squared test must be performed.  For this test to be completely valid, two assumptions must be met:

v     All expected counts are greater than 1 – This condition is met.

v     No more than 20% of expected counts are less then 5. – This condition is met.

H₀: There is no difference between how Jews from Lexington voted and how Jews across America voted in the 2000 presidential election.

H_a:  There is a statistically significant difference between how Jews from Lexington voted and how Jews across America voted in the 2000 presidential election.

Below is the Chi-Squared significance test for this comparison:

	Minitab Output for Lexington, KY vs. the USA   Expected counts are printed below observed counts             Bush  Gore/Other  Total   LEX       13       32       45           8.92    36.08     USA      100      425      525         104.08   420.92   Total      113      457      570   Chi-Sq =  1.865 +  0.461 +           0.160 +  0.040 = 2.526   DF = 1, P-Value = 0.112

 

Conclusions:

1.      There is not a statistically significant difference between Temple Adath Israel and Ohavay Zion Synagogue when you compare the way Jews voted at both synagogues in the 2002 presidential election.

2.      There is not a statistically significant difference between the way affiliated Jews in Lexington, Kentucky, and Jews across the USA voted, in the 2000 presidential election.

The data from the study, however, makes it possible to look much closer into the data.  It should be noted that the first test had a p-value of  .816.  This means that there is an 81.6% chance that that the data we collected differed from the null hypothesis by chance.  On the other hand, the second test had a p-value of .112.  While this is technically not statistically significant at the α = .05 level or α = .1 level, it should be noted that there is only a 11.2% chance that data indicating Lexington Jews and American Jews voted differently occurred merely by chance.  While the study must conclude that there is not a significant difference in the two voting records of Test 2, the study does note a possible shift in the conservative direction.

            A few factors could have negatively influenced this study.

1.      During the sampling procedure, it was often difficult to reach people.  On average, for every one person who successfully contributed to the study, there were at least two others who were not home or wouldn’t answer.  It is possible that this prevented data collectors from obtaining a completely random sample.  The data collectors called during concentrated periods of time, which also may have affected the outcome. 

For example, the TAI subjects were polled one evening.  It was later discovered that on the same evening, a TAI  Preschool end-of-the-year program was in session.  Many young parents from TAI attended this program and did not answer the phone, for obvious reasons, if and when they were called for this study.  Thus, a significant age/income sector of TAI was likely left out of this study. 

Another example may have inhibited a fair and random sample from being extracted from OZS.  The data collector in this instance was only able to call OZS members before 5:00 PM on a Tuesday.  Because many of the people who did not answer were probably at work, an entire group of voters were left out of the study.

2.      Another weakness in this study occurred in the chi-squared test.  In Test 1, the second assumption, no more than 20% of expected counts are less than 5, was violated, which may have invalidated the study.  However, because the p-value was so high, it is doubtful that a change in that value “less than 5” would have changed much in our data.

Thus, there is little evidence that Lexington Jews voted differently in the election based on their religious affiliation.  The study cannot also safely conclude that Jews in Lexington are shifting to be more Republican than American Jews as whole, but the strong evidence that was calculated might encourage further study into this part of the issue.  It would be wise for those using this study, to anticipate a statistically significant difference by the 2004 or 2008 presidential elections.