Donald Trump’s relentless messaging about a “rigged election” is having an effect on the confidence voters have that their votes will be counted accurately. But, it’s not the effect you think.
I came to this conclusion as I was considering yesterday’s Morning Consult poll results about confidence in the vote count. It so happens that I asked almost exactly the same question on a national poll during the pre-election period in 2012. (I can’t take all the credit. My colleague at Reed College, Paul Gronke, joined me in sponsoring a “double-wide” module on the 2012 Cooperative Congressional Election Study.) I decided to compare what Morning Consult found today with what we found almost exactly four years ago.
The results were surprising. The percentage of respondents who say that they are “very confident” that their own votes will be counted accurately is virtually unchanged from 2012. Confidence that votes nationwide will be counted accurately has, if anything, increased since 2012. Trump’s rhetoric appears not to have reduced Republican confidence in the accuracy of the vote count over the past four years. Rather, it has increased the confidence of Democrats. The degree of party polarization over the quality of the vote count has increased since 2012, but it is Democratic shifts in opinion, not Republican, that are leading to this greater polarization.
Let me sketch out the background here. In 2012, Gronke and I coordinated our modules in the CCES to ask a series of questions about election administration to a representative sample of 2,000 adults. Two of these questions were:
- How confident are you that your vote in the General Election will be counted as you intended?
- How confident are you that votes nationwide will be counted as voters intend?
The first question was asked of respondents who reported they intended to vote in 2012; the second question was asked of all respondents.
The response categories for both questions were (1) very confident, (2) somewhat confident, (3) not too confident, (4) not at all confident, and (5) I don’t know.
The corresponding Morning Consult questions were:
- How confident are you that your vote will be accurately counted in the upcoming election?
- How confident are you that votes across the country will be accurately counted in the upcoming election?
The response categories were identical to ours, with the exception of an additional “no opinion” option with Morning Consult.
So, while the questions are not 100% identical, they are close enough to allow some meaningful comparisons. (For those interested in a more systematic example of how similar survey research questions can be combined in this type of analysis, see the article I co-authored with Mike Sances, which appeared last year in Electoral Studies.) Both the 2012 and 2016 studies were conducted about three weeks ahead of the general election, so the timing couldn’t be better.
In the table below, I compare Morning Consult’s 2016 results with Gronke’s and my results in 2012. The numbers in the table are the percentages of the indicated respondents who gave the “very confident” response.
| Your own vote | Votes nationwide | |||
|---|---|---|---|---|
| 2012 (Gronke/Stewart) | 2016 (Morning Consult) | 2012 (Gronke/Stewart) | 2016 (Morning Consult) | |
| All registered voters | 41% | 45% | 16% | 28% |
| Democrats | 47% | 59% | 20% | 43% |
| Republicans | 42% | 41% | 13% | 18% |
The 2012 patterns were consistent with what my colleagues and I have regularly reported: the “winning” party tends to be more confident than the “losing” party and voters tend to be much more confidence of their own votes being counted accurately than votes nationwide.
The 2016 patterns are similar, with a couple of major differences. The most important similarity is that respondents in both 2012 and 2016 were more confident their own votes would be counted accurately than votes nationwide. In 2012, the local-nationwide gap was 25 percentage points (41% vs. 16%); in 2016, the local-nationwide gap dropped to 17 percentage points (45% vs. 28%).
The most important changes come as we look down the table, at the Democratic-Republican differences. Republican and Democratic opinions have changed in very different ways since 2012. At the local level, Republicans remain about as confident as they were in 2012, but Democratic confidence has grown. As a consequence, the Democratic-Republican gap in the confidence about local vote counting has grown from 5 percentage points to a much more substantial 18 percentage points.
In assessing the accuracy of the vote count nationwide, Republicans are actually a little more confident in 2016 than in 2012 (18% vs 13%), but this small change from 2016 is likely due to subtle differences between the two studies. On the other hand, Democrats have become a lot more confident. They are now a whopping 23 percentage points more confident than in 2012 that votes will be counted accurately nationwide (43% “very confident” vs 20%).
Much more work needs to be done on this issue, but a couple of tentative conclusions seem in order. The first is that Donald Trump’s complaints about a “rigged” electoral system most clearly reminded his strongest supporters of what they already believed. It is much less clear that Republicans who were not already convinced of the corruption of the election system have now had a change of heart.
The second conclusion is that Trump’s charges appear to have counter-mobilized Democratic opinion in novel ways. Democrats have come to the defense of vote counting, not only in their own back yards, but even in other people’s back yards.
Either way, summary judgements about the legitimacy of the electoral process have become more polarized in 2016 than they were in 2012. One possibility is that as time progresses, support for the electoral process as a whole will become associated with the Democratic Party in the public’s mind, with opposition associated with the Republican Party. I am hoping that this is not the case, because we have seen important bipartisan improvements in the world of election administration over the past four years, despite continued partisan differences over voter ID laws and amending the Voting Rights Act.
We certainly need to be concerned about undermining the legitimacy of elected officials, especially in circumstances where there is no hard evidence of election rigging going on. But, we also need to recall that once the November election is done and gone, elections will continue to be administered at the state and local levels. The danger for election administration with all this unsubstantiated talk about fraud is that it will undermine the comity that has often existed in handling the day-to-day details of running elections. In other words, the failure to institute improvements to local election administration will become collateral damage of this heightened polarization.