15
Republican Voting among Democratic Presidential Contributors: A Test of Rival Hypotheses Author(s): Clyde Wilcox, Clifford W. Brown, Jr. and Lynda W. Powell Source: Polity, Vol. 26, No. 1 (Autumn, 1993), pp. 127-140 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/3234999 Accessed: 27-04-2016 06:50 UTC REFERENCES Linked references are available on JSTOR for this article: http://www.jstor.org/stable/3234999?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Polity This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTC All use subject to http://about.jstor.org/terms

Republican Voting among Democratic Presidential Contributors: A Test of Rival

Embed Size (px)

Citation preview

Republican Voting among Democratic Presidential Contributors: A Test of RivalHypothesesAuthor(s): Clyde Wilcox, Clifford W. Brown, Jr. and Lynda W. PowellSource: Polity, Vol. 26, No. 1 (Autumn, 1993), pp. 127-140Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/3234999Accessed: 27-04-2016 06:50 UTC

REFERENCES Linked references are available on JSTOR for this article:http://www.jstor.org/stable/3234999?seq=1&cid=pdf-reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

http://about.jstor.org/terms

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted

digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about

JSTOR, please contact [email protected].

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access toPolity

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note

Republican Voting Among Democratic Presidential Contributors: A Test of Rival Hypotheses*

Clyde Wilcox, Georgetown University Clifford W. Brown, Jr., Union College Lynda W. Powell, University of Rochester

Over the past decade, we have begun to develop a more detailed portrait of those who contribute to political candidates and organizations. We know that only a small minority of Americans contribute to candidates, and that those who give "serious money"-defined here as contributions of $200 or more-are a wealthy, well-educated, and well-connected polit- ical elite.2 It is usually assumed that contributors to candidates, like those who campaign actively on their behalf, are strong partisans who seek to elect the candidates of their party.3

This note attempts to explain a surprising result: among those who contributed more than $200 to a Democratic presidential candidate in 1988, nearly one in four ultimately voted for George Bush for President. Although substantial numbers of Democrats have defected to Repub- lican candidates over the past several presidential elections, such defec- tions are rare among strong partisans. Presidential contributors of serious money are more likely to be strong partisans than the general public: in our survey more than half of those who gave to presidential candidates strongly identified with a party, while less than a third of

*The authors thank the Washington Area Campaign Finance Discussion Group for its help in the preparation of this note. Funding was provided by the Society for the Scientific Study of Religion, Georgetown University, Union College, and the University of Rochester.

1. Frank Sorauf, Money in American Elections (San Francisco: Scott Foresman, 1988); Ruth Jones and Warren Miller, "Financing Campaigns: Macro Level Innovation and Micro Level Response," The Western Political Quarterly, 38 (June, 1985): 187-210.

2. John Green and James Guth, "Big Bucks and Petty Cash," in Interest Group Poli- tics, 2nd ed., ed. Allan Cigler and Burdette Loomis (Washington, DC: CQ Press, 1986); Clifford Brown, Lynda Powell and Clyde Wilcox, "National Networks in Presidential Fundraising" (paper presented at the Sixty-third Annual Meeting of the Southern Political Science Association, Tampa, FL, November 9, 1991).

3. Sorauf, Money; Bruce Jacobson, Money in Congressional Elections (New Haven: Yale University Press, 1980).

Polity Volume XX VI, Number) Fall 1993 Volume XXVI, Number I Fall 1993 Polity

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

128 Research Note

those who responded to the 1988 National Election Study (NES) were strong identifiers. We will explore several hypotheses for the Republican votes of these Democratic contributors.

I. Hypotheses

Hypothesis 1. Republican voting may be highest among Republicans who gave to Democratic candidates because they were solicited by per- sonal friends, because they gave for business or social reasons, or because (for whatever reason) they gave to the home-state candidate.

The earliest studies of voting identified partisanship as the principle source of vote decisions,4 and although recent studies suggest that the role of partisanship has declined, it remains an important vote cue.5 Although we expect a majority of those who gave to Democratic presi- dential candidates to be Democrats, it is likely that some Republicans may have also contributed.

Those who give to presidential candidates do so for many reasons. Following James Q. Wilson, we focus on three: purposive or ideological motives, solidary or social motives, and material benefits.6 Previous research has shown that many Democratic candidates built networks of contributors whose contributions were solicited by friends or business contacts, and who gave for solidary or material motives, rather than pur- posive ones.7 Such networks may have included a number of Repub- licans. Republicans may have given to Gephardt's campaign, for exam- ple, because a friend asked them to, or because they expected him to become a party leader in the House of Representatives, and therefore support for his presidential campaign might help gain access of impor- tance to their business. Some Republicans may also have given to a Democratic candidate because he comes from their home state. In-state

money is less ideological than out-state money, i.e., those who give to their home-state candidate are less likely to be ideologically similar to

4. Angus Campbell, Philip Converse, Warren Miller and Donald Stokes, The Ameri- can Voter (New York: Wiley, 1960).

5. Paul Abramson, John Aldrich and David Rohde, Change and Continuity in the 1988 Elections, rev. ed. (Washington, DC: CQ Press, 1991).

6. James Q. Wilson, Political Organizations (New York: 1973). 7. Clifford Brown, Lynda Powell and Clyde Wilcox, "The Constituency Base of Cam-

paign Contributions to the 1988 Presidential Nomination Candidates" (paper presented at the Annual Meeting of the American Political Science Association, San Francisco, CA, September 2, 1990); Clifford Brown, Lynda Powell and Clyde Wilcox, "National Net- works in Presidential Fundraising.'

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 129

that candidate than those who give to an out-of-state candidate.8 Hence, some Republicans give because friends or business associates ask them, and others give from home-state pride. Of course, when Republicans give to Democrats but ultimately vote for a Republican, this is not what we normally mean by vote defection.

Hypothesis 2. Republican voting may be highest among conservative Democrats who were ideologically closer to Bush than to Dukakis. Defection may be most closely linked to foreign policy conservatism and conservatism on social issues.

James Lengle has argued that the Democratic party frequently nom- inates candidates that are more liberal than its rank-and-file membership because of the ideological distinctiveness of primary voters.9 Although the thesis that Democratic primary voters pull the party to the left has been challenged in recent years,10 it is often argued that Democratic nominees lose votes because they are more liberal than the electorate, especially on foreign policy.1'

Hypothesis 3. Republican voting may be linked to symbolic and cultural politics. Those Democrats who feel closer to the Moral Majority and the National Rifle Association, and cooler toward feminists and the Ameri- can Civil Liberties Union will be more likely to vote for Bush. White evangelicals may be more likely to vote for Bush.

Many scholars have argued that political divisions in American politics are closely linked to symbolic issues. The symbolic politics approach fre- quently focuses on group affect. Ted Jelen, for example, has argued that the politics of the Christian Right is best understood by focusing on the social and political groups that conservative evangelicals dislike.'2 Those Democrats who approve of the Moral Majority and the NRA, but who disapprove of feminists and the ACLU might be more likely to vote for Bush, even after holding constant ideological beliefs.

One social group that has increasingly supported Republican can-

8. Brown, et al., "Constituency Base." 9. James Lengle, Representation and Presidential Primaries: The Democratic Party in

the Post-Reform Era (Westport, CT: Greenwood Press, 1981). 10. Barbara Norrander, "Ideological Representativeness of Presidential Primary

Voters," American Journal of Political Science, 33 (August, 1989): 570-87. 11. But see William Schneider, "Tough Liberals Win, Weak Liberals Lose," in The

Quest for National Office, ed. Stephen Wayne and Clyde Wilcox (New York: St. Martin's, 1991).

12. Ted Jelen, The Mobilization of Religious Belief (New York: Praeger, 1991).

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

130 Research Note

didates in the past several years is white evangelical Christians. Lyman Kellstedt, and Corwin Smidt and Paul Kellstedt have reported an increas- ing tendency for white Democratic evangelicals to defect to Republican presidential candidates, and for younger evangelicals to identify with the Republican party.13 The increasing Republicanism of white evangelicals has been hypothesized to be linked to symbolic politics.14

Hypothesis 4. Republican voting may be linked to racial attitudes and affect. Recent scholarship has suggested that racial issues have been an important source of changing party loyalties. Edward Carmines and James Stimson have argued that race has developed into the most signifi- cant difference between the parties, and that the race issue has caused substantial defections among traditional Southern Democrats.'s The authors report that party activists generally take quite different positions on racial issues. Yet it is possible that the presence of two white Southern candidates, both of whom sought to position themselves as moderates or even conservatives, may have attracted contributions from racially con- servative Democrats who later defected to Bush rather than vote for Dukakis.

Hypothesis 5. Vote defection may be highest among those who strongly preferred another Democratic nominee to Dukakis.

Although Richard Born's claim that divisive primaries hurt the party's nominee is controversial,16 there is good evidence that divisive primaries matter in presidential campaigns.17 The research on this question has largely focused on aggregate results, and many studies show that parties with divisive primaries are more likely to lose the general election. Lengle shows that in states with divisive primaries, those who preferred the los- ing candidate more frequently defect to the other party. Presumably the

13. Lyman Kellstedt, "Evangelicals and Political Realignment" (paper presented at the Annual Meeting of the American Political Science Association, Washington, DC, Sep- tember, 1986); Corwin Smidt and Paul Kellstedt, "Evangelicals and the Post-Reagan Era: An Analysis of Evangelicals in the 1988 Presidential Election" (paper presented at the Biannual Citadel Symposium on Southern Politics, Charleston, SC, 1990).

14. Pamela Conover, "The Mobilization of the New Right: A Test of Various Explana- tions," Western Political Quarterly, 36 (December, 1983): 632-49.

15. Edward Carmines and James Stimson, Issue Evolution: Race and the Transforma- tion of American Politics (Princeton, NJ: Princeton University Press, 1989).

16. Richard Born, "The Influence of House Primary Election Divisiveness on General Election Margins," Journal of Politics, 43 (August 1981): 640-61.

17. Lengle, Representation; Patrick Kenney and Tom Rice, "The Relationship between Divisive Primaries and General Election Outcomes," American Journal of Political Sci- ence, 31 (February, 1987): 31-44.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 131

mechanism for this defection is complex. Undecided and/or independent voters may be influenced by the negative advertising of the intra-mural contest. Those who strongly preferred another candidate may sit out the election, or may actually defect to the nominee of the other party. There- fore we suggest that those contributors who strongly preferred a losing candidate may be more likely to defect to Bush in 1988.

II. The Data

The data for this study come from a national mail survey of individuals whose contributions to a single presidential candidate totalled at least $200. We sampled pages of the detailed reports filed by the candidates at the Federal Election Commission. We conducted three mailings between December, 1988, and March, 1990. Our survey netted a response rate of approximately 37 percent, yielding 1248 respondents.s1 We have weighted the data to compensate for the uneven probabilities of selection of those who gave several separate contributions, and have further weighted the data to reflect the approximate number of contributors of $200 and above that gave to each candidate. The weighted number of Democratic contributors is 514, 482 of whom reported a vote choice in 1988.

As independent variables, we will include standard demographic varia- bles. To test Hypothesis 1, we include a seven-point measure of partisan- ship, a dummy variable to identify those whose contributions were per- sonally solicited by someone they knew, a dummy variable to identify those contributors in the home state of the candidate, and two scales con- structed to measure material and solidary motives (see the Appendix for details of scale construction).

To test Hypothesis 2, we have constructed several measures of ideol- ogy. First, we have constructed a measure of ideological proximity from items that asked contributors to place themselves and the various can- didates on a seven-point liberal/conservative scale. This item measures the difference in proximity from Bush and Dukakis. In addition, we have constructed three scales to measure attitudes on social issues, economic

18. This response rate is somewhat lower than what was once the standard, but response rates to phone and mail surveys have declined in recent years. Mail surveys now compete with many direct-mail solicitations that arrive in envelopes marked "SURVEY ENCLOSED: PLEASE OPEN AT ONCE." Our response rate is approximately that of some other recently published studies, e.g., John C. Green and James L. Guth, "The Christian Right in the Republican Party: The Case of Pat Robertson's Supporters," Jour- nal of Politics, 50 (1988): 150-64.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

132 Research Note

issues, and foreign policy issues. Finally, we use four ideological self- placement items, one for general ideology, one for economic issues, one for social issues, and one for defense issues.

To test Hypothesis 3, we have constructed a scale from feeling ther- mometers toward the Moral Majority, the NRA, feminists, and the ACLU. The scale is the difference in the mean evaluation of the two

culturally conservative groups and the two culturally liberal groups. In addition, we have identified evangelical Christians through two ques- tions: one on the authority of Scripture, a second on a born-again experi- ence. We define as evangelicals those who believe the Bible is the inerrant word of God and who report a born-again experience.19

To test Hypothesis 4, we include a measure of attitudes on affirmative action, and a feeling thermometer toward civil rights leaders.

To test Hypothesis 5, we have constructed an item that measures the maximum difference in affect between any other Democratic candidate and Dukakis. The item is constructed by first finding the maximum score on a feeling thermometer for all other Democratic candidates, then sub- tracting the feeling thermometer for Dukakis from that maximum affect measure. We have also constructed dummy variables to measure con- tributions to each candidate.

III. Bivariate Results

In Table I, we show the percentage of contributors of serious money to each of the Democratic candidates who ultimately voted for Bush. Respondents were invited to pull the identifying sticker from their envelope, and some did. We are therefore unable to determine the can- didates to whom a few of the contributors gave. Although the number of contributors to some candidates is small, it is evident from this table that some contributors to all Democratic candidates voted for Bush, and that this rate of Republican voting was larger for those who gave to Gephardt or Gore. Fully 20 percent of those who gave $200 or more to Dukakis during the primary elections did not vote for him in the general election.

In Table II, we summarize the bivariate evidence in support of each hypothesis. There is strong support for a partisanship explanation. Fully 20 percent of those who gave to Democrats considered themselves Republicans, and more than 80 percent of these Republican contributors

19. Our Bible item allowed respondents to indicate that they thought the Bible was inerrant (i.e., contains no errors), and the even more orthodox position that it is literally true. We found that respondents were able to distinguish between these two positions. We have chosen to allow either of these orthodox responses to be coded as evangelical.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 133

Table I. Republican Voting among Democratic Contributors in 1988

PERCENT

CANDIDATE VOTING FOR BUSH N

Babbitt 29 22 Dukakis 20 189

Gephardt 42 53 Gore 37 77

Jackson 14 36 Simon 17 43 Not Ascertained 14 62

Total 24 482

Table II. Bivariate Results: Correlations with Vote Choice

CORRELATION

Partisanship Party Identification .69b Asked to give by Acquaintance .12a In-State Contributor .13a Material Benefits .20b

Ideology Proximity .75b Self-Classification

Overall .71 b Economic .60b Social .55b Defense .60b

Issue Positions

Economic .57b Social .49b Foreign Policy .56b

Symbolic Politics Polarized Affect .57b White Evangelical .17b

Racial Politics (All Respondents) Affirmative Action .49b Affect, Civil Rights Leaders -.45b White Southerner .17a

Candidate Preference Preference for Other Democrat -.03

Values represent Pearson's correlations between item or scale and voting for Bush among Democratic contributors.

ap < .05.

bp c .01.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

134 Research Note

voted for Bush. Republican contributors to Democratic candidates were more likely to live in the home state of the candidate to which they gave, to be asked to give by a personal acquaintance, and to say they were motivated by material incentives. Each of these variables was associated with a higher rate of Bush voting.

There is strong support for an ideological explanation as well. Each of our ideology measures is strongly correlated with Republican voting, and with each other. This elite sample had strongly constrained attitudes, so each of these measures of ideology is highly correlated with the others, and each does almost equally as well in predicting a Bush vote.

The symbolic/cultural politics explanation also receives some support. Polarized group affect is a strong predictor of voting for Bush, and white evangelicals were slightly more likely than others to cast Republican ballots. There is evidence as well for an explanation that focuses on racial politics. Attitudes on affirmative action and affect toward civil rights leaders were strong predictors of the vote.

Finally, there was no evidence that those contributors who personally preferred a different Democratic candidate to Dukakis were more likely to defect to Bush. Because the number of cases involved was small, we were unable to test this explanation separately for each Democratic can- didate. In the multivariate analysis below, we will look for differences in rates of Republican voting by those who gave to specific candidates.

IV. Multivariate Results

Multivariate analysis is needed to sort out the separate effects of each of our hypotheses. We estimated a logistic regression equation with vote choice as the dichotomous dependent variable, and some independent variables from each explanation (Table III). In addition, we included several demographic variables, including race, sex, age, education, income, region, and religion. In addition to white evangelicals, we iden- tified two other religious and ethnic groups who gave to Dukakis in dis- proportionate numbers: Greek Orthodox and Jewish contributors. Because these two groups were disproportionately likely to give to Dukakis, and because there were ethnic ties for each group to the can- didate or his wife, we expect a lower rate of Republican voting among them. A second equation (not shown) added dummy variables to identify those who gave to each other Democratic candidate. The excluded cate- gory is those who gave to Dukakis, so that in this second equation the effects of these dummy variables are assessed relative to Dukakis givers.

Because our ideology measures were highly correlated with each other, we use only one in the multivariate analysis. We have elected to use ideo-

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 135

Table III. Multivariate Results: Logistic Regression Analysis of Vote Choice

b -2 log LR

Demographic Variables Blacks -.71 .75 Women -1.23c 2.86 Age .05c 2.94 Education -.51a 6.43 Income .40 1.73 Southern .56 .68 Greek Orthodox -4.23b 11.43 Jewish -.33 .00

Other Variables

Partisanship .76b 18.80 Material Motives .06 .02 Solicited by Acquaintance 1.06C 2.76 In-State .78 1.83

Ideological Proximity .99b 33.60

Polarized Group Affect .02a 4.56 White Evangelical -2.17 2.11 Affirmative Action -.02 .01 Prefer Other Democrat -.83 1.15

Percentage predicted correctly: Bush 86

Dukakis 97 Overall 93

Coefficients are unstandardized.

ap < .05.

bp c .01. Cp ' .10.

logical proximity, since it was the most strongly correlated with Repub- lican voting. In addition, the two measures of racial politics were also highly correlated, so we include only attitudes toward affirmative action in the analysis below.20

The coefficients in Table III are unstandardized MLEs, so their inter-

20. We make this choice because affect toward civil rights leaders was highly correlated with the affect measure for the symbolic politics explanation. Of course, in this elite sam- ple, there are fairly sizable correlations between many of the independent variables. None of the correlations in the models estimated reach levels substantial enough to raise the ques- tion of multicollinearity, however.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

136 Research Note

pretation is not straightforward. We show in addition the -2 log LR, which is intuitively similar to the T value in OLS regression, although the values needed for statistical significance are somewhat higher. Higher values indicate greater confidence that the coefficient is significantly dif- ferent from 0.

Ideology is a powerful predictor of vote choice. The results suggest that conservatives voted for Bush even after controls for partisanship, so Dukakis lost votes among conservative Democratic contributors. We estimated a series of additional equations to determine whether any single aspect of ideology was most important in producing Republican voting. Because of the high correlation between the ideology variables, we entered a separate measure in each equation. The results suggest that conservatism on economic issues is a far better predictor than either foreign policy or social issue conservatism.21

Other significant predictors of vote choice were partisanship, Greek orthodox religious affiliation (Greeks remained more loyal to Dukakis), education (with better educated contributors supporting Dukakis), race (blacks were more loyal) and symbolic/cultural group affect (those who felt closest to culturally conservative groups and coolest to culturally liberal groups voted for Bush).

Although in-state contributors, those asked personally by an acquain- tance, and material motives were not significant predictors of a Bush vote, they are important indirect sources of Republican voting. Each of these three variables is a significant predictor of partisanship among Democratic contributors-that is, Republicans who gave to Democratic candidates were generally motivated by material benefits or personal friendship networks.

In the second equation, only one of the dummy variables that identi- fied specific candidates proved statistically significant. Those who gave to Jesse Jackson were significantly more likely to defect to Bush than those who gave to other Democrats after controls for race and ideology. In Table I, we see that in the bivariate case Jackson contributors were more loyal to Dukakis than those who gave to other Democratic candi- dates. We interpret this coefficient to indicate that Jackson supporters defected more than would be expected given their Democratic partisan- ship and their strong liberalism. Vote defection was much larger among white Jackson contributors than among blacks, although the number of cases is quite small.

21. Social issue conservatism was correlated with symbolic group affect at .60, therefore reducing its overall effects. Neither concrete issue positions on social issues nor on foreign policy were significant predictors of a Bush vote, but economic issue positions were the second strongest predictor in the equation in which they were included.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 137

V. Conclusions

Our results suggest that many of the Democratic contributors who ultimately voted for Bush were Republicans. Indeed, nearly one in five of those who gave to Gephardt had voted for Republican candidates in every presidential election since 1975. These cross-party contributions were made primarily for non-ideological reasons such as business con- tacts, home-state pride, or because a personal acquaintance solicited the money.

Yet even after controls for partisanship, a powerful predictor of a Bush vote from those who gave money to Democratic candidates was ideological proximity. Indeed, a separate analysis conducted among only those with Democratic partisanship showed an even stronger relationship between vote choice and ideology. This result fits the claims by some scholars that Democratic candidates frequently lose votes because they are too liberal. Yet it was not on foreign policy or social issues that ideol- ogy seemed to matter most, but on economic matters. Democratic and Republican contributors were a wealthy lot, and although there were substantial differences between Democratic and Republican contributors on economic issues, a core of Democrats who were fairly moderate on economic issues voted for Bush.

Symbolic and cultural politics also played a role in Republican voting. Those who were relatively warm to the Moral Majority and the National Rifle Association, and relatively cool toward feminists and the ACLU were more likely to vote for Bush. This suggests that analysts who have focused on the symbolic content of the Bush campaign, and the emphasis the campaign placed on values, have identified a source of Republican voting that affects this elite group. Interestingly, white evangelicals were not more likely to cast Republican ballots after multivariate controls.22

Two of the hypotheses received no support in these data. Racial poli- tics did not seem to play a significant role in Republican voting, although there was a significant bivariate correlation between racial attitudes and a Bush vote. After controls for partisanship and ideology, those who voted for Bush were not more likely to oppose affirmative action. In addition, generally speaking, there was little evidence that strong support for another Democratic candidate led to vote defection, although those who gave to Jackson were somewhat more likely to defect to Bush than would be predicted from their ideology and partisanship.

22. A large proportion of white evangelical contributors gave to Republican candidates, with fundamentalists and other evangelicals giving to Bush and pentecostals contributing to Robertson.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

138 Research Note

Appendix: Question Wording and Scale Reliabilities

Hypothesis 1.

Generally speaking, do you consider yourself a strong Democrat; a not very strong Democrat; an Independent, closer to Democrat; an Indepen- dent, closer to neither; an Independent, closer to Republican; a not very strong Republican; or a strong Republican?

Thinking back to when you made your first contribution to a Presiden- tial candidate in the recent election, how was that contribution solicited?

I was asked by someone I knew personally

There are a number of reasons people give for being active in politics. Please describe the importance of each of the following for you per- sonally:

MATERIAL MOTIVES:

for business or employment reasons it is expected of someone in my position

gamma = .62

Hypothesis 2.

Here is a seven-point scale on which political views that people might hold are arranged from extremely liberal to extremely conservative. Where would you place yourself and each of the others listed on this scale? (Choices are extremely liberal, liberal, slightly liberal, moderate or middle of the road, slightly conservative, conservative, or extremely conservative).

Generally how liberal or conservative are your views on economic issues, involving taxes and spending decisions? (same choices)

Generally how liberal or conservative are your views on defense issues, involving our relationship with Russia, aid to the Contras, Star Wars and spending on conventional and nuclear forces? (same choices)

Generally how liberal or conservative are your views on social issues, such as abortion, mandatory school prayer, the death penalty, pornogra- phy and so forth? (same options)

SOCIAL ISSUE SCALE: (Starred items are strongly agree to strongly disagree)

Mandatory death penalty for murder should be the law.* The government should not regulate what adults read, even if that in-

cludes pornography.*

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

Research Note 139

We should have mandatory prayer in the public schools.* The government should prohibit abortion.* Some people feel that women should have an equal role with men in run-

ning business, industry and government. Others feel that women's place is in the home. (7-point anchored scale)

Alpha = .80

ECONOMIC ISSUES SCALE

This country would be better off if it returned to the gold standard.* The government should provide a program of health insurance for those

not currently insured.* Free trade is important, even if it results in the loss of U.S. jobs.* Child care is the parent's responsibility; government should not be

involved.*

Some people feel that the government in Washington should see to it that every person has a job and a good standard of living. Others think the government should just let each get ahead on his own. (7-point anchored scale)

Alpha = .71

FOREIGN POLICY SCALE

Defense spending should be reduced.* U.S. military aid for the Contra rebels in Nicaragua should be con-

tinued.*

Some people think the government should provide much less money for defense. Other people feel that defense spending should be greatly increased. (7-point anchored scale)

Some people feel it is important for us to try to cooperate more with Russia, while others believe we should be much tougher in our dealings with Russia. (7-point anchored scale)

Alpha = .86

Hypothesis 3.

Would you rate the following groups in terms of how favorably or un- favorably you feel toward them. Use a scale between 0 and 100. Ratings over 50 mean that you feel favorable toward the group. Ratings under 50 mean that you feel unfavorable. A rating of 50 means that you don't feel favorable or unfavorable, but have some knowledge of the group. If you have no knowledge of the group, please place an X on the line.

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms

140 Research Note

National Rifle Association

Moral Majority Feminists ACLU

Which comes closest to your view of the Bible? the actual Word of God, to be taken literally word for word the inspired Word of God: it contains no errors but should not be

taken literally the inspired Word of God, but it contains human errors a good book but not the Word of God

Please check all the terms that accurately reflect your religious beliefs: born again Christian

gamma for evangelical Christian scale = .89

Hypothesis 4.

Feeling thermometer item toward Civil Rights Leaders

Some people feel that the government in Washington should make every effort to improve the social and economic position of blacks. Others feel that the government should not make any special effort to help blacks because they should help themselves. (7-point anchored scale)

Hypothesis 5.

A measure was constructed that equalled the warmest score on a feeling thermometer item for any Democratic candidate other than Dukakis. The feeling thermometer for Dukakis was subtracted from this measure. The resultant item ranged in theory from -100 (when Dukakis had a 0 and another Democratic candidate had 100) to 100 (when Dukakis had 100 and all other Democratic candidates had 0).

This content downloaded from 141.161.13.197 on Wed, 27 Apr 2016 06:50:20 UTCAll use subject to http://about.jstor.org/terms