With a presumptive black presidential nominee, there will be plenty of speculation about the “Wilder Effect“: when (presumably white) survey respondents tell interviewers that they support a black candidate but really intend to support the opposing white candidate. The effect is named for Douglas Wilder, whose margin of victory in the 1989 Virginia gubernatorial race was closer than the polls suggested. In this election, Hillary Clinton’s surprise victory in the New Hampshire primary raised the possibility that the Wilder Effect may apply to Obama. I discounted this possibility here with some quickie data analysis.
Now comes political scientist Dan Hopkins with a more definitive answer:
bq. The 2008 election has renewed interest in the Wilder effect, the gap between the share of survey respondents expressing support for a candidate and the candidate’s vote share. Using new data from 133 gubernatorial and Senate elections from 1989 to 2006, this paper presents the first large-sample test of the Wilder effect. It demonstrates a significant Wilder effect only through the early 1990s, when Wilder himself was Governor of Virginia. Although the same mechanisms could impact female candidates, this paper finds no such effect at any point in time. It also shows how polls’ over-estimation of front-runners’ support can exaggerate estimates of the Wilder effect. Together, these results accord with theories emphasizing how short-term changes in the political context influence the role of race in statewide elections. The Wilder effect is the product of racial attitudes in specific political contexts, not a more general response to under-represented groups.
Below I reprint part of one of his graphs. It shows the gap between the candidate’s expected performance in the polls and their actual performance on Election Day. The blue trend line shows that while in the early 1990s black candidates did better _in_ polls than _at_ the polls, this “gap” disappeared soon thereafter.
Hopkins investigates the 2008 polls and actually finds the opposite of a Wilder Effect: on average, Obama does a little better than the polls predicted.
bq. The mean polling-performance gap was 1.4 percentage points for Senator Clinton, and the reverse for Senator Obama. This estimated mean is not at all sensitive to particular polls or states: if we remove the observations for five randomly chosen states at a time, we still observe that Senator Obama’s election-day performance was better than his polling on average in every one of the 10,000 simulations. This is yet more evidence that the Wilder effect, strong in the early 1990s, is strong no longer.
There is much more in the paper. Find it here.