Some Statistics on Bigotry in Voting
Though the outcome of the 1960 Presidential election will be determined by a variety of factors, it is clearly the “religious issue” which most fascinates the majority of our political commentators. For if Kennedy wins, the bigotry against Catholics (particularly Irish Catholics) that has flared up at different times in American history—in the Know-Nothing party of pre-Civil War days which secured over 25 per cent of the Presidential vote in 1856, in the notorious American Protective Association (APA) of the 1880’s and 90’s, and in the Ku Klux Klan in the 1920’s—will finally have been laid to rest, at least as a potent political force. More significantly, the highest office in the land will for the first time have been bestowed on someone who does not belong to the inner circle of Protestants of Nordic and Anglo-Saxon descent.
But if religion seems to be of major interest to the political analysts and to the churchmen, particularly evangelical Protestants, there is surprisingly little actual evidence that the “religious issue” will be decisive for the voters themselves. In many states (except, of course, the Southern) Americans have repeatedly elected Catholics and Jews to high office. Thus Catholics have held senatorships or governorships in—to give a partial listing—Alaska, California, Ohio, New York, Pennsylvania, Minnesota, Montana, Wisconsin, Maine, Washington, and Rhode Island. And since the 1930’s, Jews have been elected to the office of governor and senator in Oregon, New York, Alaska, Connecticut, and Illinois. Granted that such office is not so important as the Presidency of the United States, one still would expect that people who feel strongly enough to vote against a Presidential candidate simply because of his religion would not knowingly vote for a Catholic or a Jew running for any major state office.
Two kinds of evidence are frequently cited to support the belief that the 1960 election may turn on the religious issue: the defeat of Al Smith in 1928 and the findings of various recent public opinion polls. Smith’s defeat, of course, has usually been attributed to the strength of anti-Catholic feeling in the country. Yet, as Richard Hofstadter (among others) has pointed out, Smith received a larger vote in 1928 than the Democrats had been able to muster for twelve years—since the wartime re-election of Woodrow Wilson in 1916. Moreover, the average Democratic percentage of the national vote for the first seven Presidential elections in the 20th century was 39.4, or over one per cent less than the 40.8 that Smith secured in 1928. One should also keep in mind that the 1928 election was held in a period of relatively full employment, which permitted Herbert Hoover to claim Republican credit for the country’s continued prosperity and personal credit for himself as the incumbent Secretary of Commerce (the Cabinet officer then most closely identified with economic well-being). Considering, thus, the factor of prosperity and the size of Smith’s vote, it is possible to argue that Smith did as well or even better than one would have expected a Democrat to do in 1928, and perhaps also to argue that his religious affiliation, far from being responsible for his defeat, may have been an asset at the polls.
But what of the other evidence which has been cited to support the contention that religious bigotry defeated the Democrats in 1928—the fact, for example, that Smith lost a number of states in the presumably solid South? On this point, it should be noted that Smith actually lost border states, not (except for North Carolina) Southern ones, the same states which in later years were to vote against Adlai Stevenson—Florida, Oklahoma, Tennessee, Texas, and Virginia. It is true that these states, as part of the solid South, had all voted Democratic in 1924, and that all, except Oklahoma and Tennessee, had also been Democratic in 1920, a year in which the Democrats, in a two-party race, secured only one-third of the national vote; obviously, Smith, in the Southern and border states, won much less than the normal Democratic vote. But this is only half the picture, for Smith made up for the loss in other parts of the country where he got, on the average, more votes than most of his 20th-century Democratic predecessors had. The question concerning Smith’s Catholicism, therefore, is not how many votes it lost him in the border states, but how the electorate as a whole behaved in 1928.
A detailed—and unfortunately largely ignored—ecological study of the 1928 voting patterns by two sociologists, William Ogburn and Nell Talbot (Social Forces, December 1929), tells part of the story. Relating the 1928 voting returns of 123 randomly selected counties to various attributes of the populations—the proportions of foreign-born, of urban residents, of regular Democratic voters, of Catholics, and of “wet” voters—Ogburn and Talbot found that the most decisive correlate of the Presidential vote was not religion but attitude toward prohibition. (Smith, it will be recalled, was a wet.) The wet influence was measured by analyzing previous votes of counties on the prohibition issue in various referendums, and the influence of religion was measured by relating the vote to the proportions of Catholics in the counties studied. Though a correlation was discovered between Catholicism and votes for Smith, it was much lower than the correlation between the opposition to prohibition and voting for Al Smith.
The counties investigated by Ogburn and Talbot which yielded these particular results were counties of eight non-Southern states, but the same findings may help account for the Democratic defections in the South as well. The South was not only a center of nativist anti-Catholicism, it was also the stronghold of prohibition (Mississippi to this very day is dry). Given the fact that the major issue of 1928 was prohibition, with Smith advocating repeal and Hoover supporting retention, it is not surprising that Smith should have lost heavily in prohibitionist centers.
Smith’s gains over his Democratic predecessors, on the other hand, came from the opponents of prohibition, people who were largely urban working-class and of recent immigrant background. But there is somewhat more to be said of the enthusiasm of these voters for the wet Smith. Their support of him was based not only on his opposition to prohibition and on his religion (since most Catholics were included among the wets), but also—as Samuel Lubell has pointed out—on his identification with economic reform and the further fact of his non-Anglo-Saxon ethnic origin. In other words, this group probably voted for Smith more heavily than it would have voted for a wet Democrat who was also an Anglo-Saxon Protestant. “What Smith really embodied,” Lubell says, “was the revolt of the underdog, urban immigrant against the top dog of ‘old American’ stock. His Catholicism was an essential element in that revolt.” In any case, there can be no doubt that the Republicans would have won against any Democratic candidate. The country in 1928 still preferred prosperity and the “noble experiment,” and there is considerable room for doubt that anti-Catholicism was the determining factor in Smith’s defeat.
Three decades after Smith’s defeat, the revived possibility of a Catholic nominee for President led the pollsters to investigate the connection between religious feeling and vote preference. As far back as the fall of 1958, the Gallup Poll asked a sample of the electorate: “If your party nominated a generally well-qualified person for President and he happened to be a Catholic, would you vote for him?” The question was also asked with respect to other groups—Jews, Baptists, Negroes, atheists, and women. Of all the categories asked about, including the Catholic, the atheists received the largest “No” proportion (75 per cent), and the Baptists the lowest (3 per cent). Over half the sample (53 per cent) said they would vote against a Negro, and antagonism to the idea of a female President (43 per cent) turned out to be much greater than to a Jew (28 per cent) or, most significantly for our purposes, to a Catholic (25 per cent). But of the Protestants interviewed—and this is the more crucial figure—34 per cent said they would not vote for a Catholic (or a Jew); of the Jews interviewed, only 6 per cent opposed the possibility of a Catholic President.
If these data told the whole story, those who fear that Kennedy will be defeated because of his religion would probably be proved right. However, a number of other variables must be considered. First, survey data suggest that Kennedy will secure the votes of at least three-quarters of the Catholic electorate—a group which now comprises over one-quarter of the total electorate, more than twice their proportion in Al Smith’s day. Although in 1956 (according to the Survey Research Center of the University of Michigan1) a majority of Catholics voted for Eisenhower, both party registration and response to polls questioning normal party allegiance show that the Catholic population during the Eisenhower era still retained its traditional identification with the Democratic party,2 Without a powerful nonpartisan figure like Eisenhower to lure them into the Republican column, normally Democratic Catholics will have no reason to change their usual party allegiance in 1960. Catholic voters, moreover, were particularly concerned with the Communist issue during the 1950’s, and many supported Eisenhower because they felt that his resistance to the Communist threat would be stronger than Stevenson’s. But it is safe to say that Catholics who are still concerned about the Communist issue will trust Kennedy far more than they did Stevenson, and Nixon far less than they did Eisenhower.
Apart from all this, many Catholics have reported to the pollsters that they vote for Catholics regardless of party affiliation. When asked by Gallup, early in 1960, “If the party other than the one you normally vote for nominated a man for President in 1960 who happened to be a Catholic, do you think you might vote for him, or not?”—52 per cent of the Catholics answering said they might shift parties, and 37 per cent said they would not. But evidence that even more concretely indicates the propensity of Catholies to support Catholic candidates, regardless of party, comes from the previously cited Michigan study of the 1956 elections; it found that in Congressional and Senatorial races “Catholic voters are quite willing to cross party lines to support a candidate of the same creed.” Catholic voters were also interviewed before the 1956 election in a private survey (reportedly sponsored by John F. Kennedy) to determine whether a Catholic running as President or Vice-President would influence the vote one way or the other. According to Arthur Krock in the New York Times, this survey concluded that the numbers of the Democratic Catholics who had defected in 1952 but would return to a 1956 ticket which included a Catholic would be enough to shift 132 electoral votes to the Democratic side of the ledger. A recently published study of The Catholic Vote3 by a well-known political scientist, John H. Felton, based on a detailed investigation of six elections in different states from 1922 to 1960, comes to similar conclusions.
Today approximately 26 per cent of the American electorate is made up of Catholics and 4 per cent of Jews. The latter are normally even more heavily committed to the Democrats than are the Catholics. (Early in 1960, 66 per cent described themselves as Democrats, while only 9 per cent were willing to own up to a Republican identification. In the 1956 elections over three-quarters of the Jews voted for Stevenson.) Although there is considerable dislike for Kennedy’s father among many Jewish voters, all the available post-convention survey data suggest that the Democratic proportion of the Jewish electorate should be at least as high as it was when Eisenhower headed the Republican ticket. Thus, the two major non-Protestant religious groups who together comprise 30 per cent of the voters should give between 75 and 80 per cent of their vote to Senator Kennedy. And this non-Protestant group is even more important than the percentages indicate, for it is concentrated in densely populated states which by themselves constitute close to a majority of the electoral votes.
As Peter Odegard has pointed out in his Religion and Politics,4 “over 80 per cent of American Roman Catholics are to be found in less than a dozen states.” Considerably more than 30 per cent of the population of eight politically potent states—New York, New Jersey, California, Illinois, Massachusetts, Michigan, Pennsylvania, and Connecticut—is made up of Catholics and Jews. Furthermore, the Northern Negroes who gave the Democratic ticket a large majority (about 80 per cent)—even against Eisenhower—are also concentrated in most of the same cities. Although Nixon is making valiant efforts to break this majority down, and some prominent Negroes such as Jackie Robinson have endorsed him, the organizations and leaders most involved in the struggle for civil rights are still behind the Democrats, and therefore, as surveys indicate, the Negroes are sure to remain a bulwark of the Democratic vote in the North.
If, then—as seems likely from the available data—Kennedy should receive majorities of from 75 to 80 per cent among Catholics, Jews, and Negroes, he could win the election with as little as 35—40 per cent of the white Protestant vote. This means that he could afford the defection (largely in the South) of 3—8 per cent of the 43 per cent of the white Protestants who call themselves Democrats, even in the impossible event that he did not receive a single vote from the 21 per cent who call themselves Independent, or the 36 per cent who identify with the Republicans.
But can we estimate how much of this white Protestant vote Kennedy will lose? As I have noted, considerable antagonism to a Catholic candidate for President was expressed by one-third of the Protestants interviewed in various Gallup surveys—and undoubtedly others were inhibited from admitting a religious prejudice to an interviewer. On the other hand, Gallup also found that Protestants are less interested than are Catholics in the outcome of the 1960 race (a fact which may reflect the lesser emotional involvement of the majority group), while the Michigan study indicates that in general Catholics as a group are much more prone to vote than Protestants. The investigations by Felton mentioned earlier also show that in election contests between Protestants and Catholics outside of the South, Protestants “are not as inclined [as Catholics] to discriminate with their ballots on the basis of religion.” He reports further that his statistical findings jibe with “more impressionistic observations derived from political leaders in such states as Ohio, Michigan, Kentucky, Maryland, and Massachusetts.”
A further point ought to be raised in connection with the Protestant vote. As has already been indicated, more Protestants think of themselves as Democrats than as Republicans (though the committed Democrats among them—43 per cent—are in a minority of the total group, which includes 36 per cent Republicans and 21 per cent Independents). Class background for Protestants, in addition, is a more important factor in party allegiance than for Catholics or Jews—the majority of the Protestant workers are Democrats, the middle-class and well-to-do are preponderantly Republican. Thus, since poorer and less educated Protestants are more likely to be vigorously anti-Catholic, the election may turn on the relative strength of their religious and class prejudices, for many of these poorer Protestants will have to choose between two sets of attitudes—their opposition to the Republicans as the party of the well-to-do, and their antagonism to John Kennedy as a Catholic. When entering the ballot box, many voters are faced—wittingly or not—with the need to choose among conflicting prejudices, and it often turns out that they prefer a Democratic Catholic to a Protestant Republican, or the other way around. Since a man’s ethnic origin is only one of the many factors which determine how the voters will react to him, in the case of the poorer Protestants, a goodly number are likely to find the idea of voting for a Republican more repugnant than voting for a Catholic.
In general, then, one can safely predict that Kennedy will not lose the election because he is a Catholic, that indeed, if the religious issue has any effect at all, it will be to add to his chances of winning an Electoral College majority.
1 Angus Campbell et al., The American Voter (John Wiley, 1960).
2 According to Gallup, for example, early in 1960, 57 per cent of American Catholics called themselves Democrats, as against 18 per cent Republican and 25 per cent Independent.
3 The Hauser Press (1960).
4 Oceana Publications (1960).