Wednesday, September 18, 2019 | ePaper
Study Shows No Influence On US Vote
The good news is because of the bad news.
The good news is that Russian disinformation and propaganda probably did not influence vote tallies in the 2016 election. The bad news is that this seems to be the case because no one listened to the propaganda any more than they listened to legitimate factual presentations.
Alan Abramowitz-a prominent scholar of elections and political polarisation-just published an important study of voting patterns in 2016. The gist of the new analysis is that Trump's narrow victory is accounted for by the known causes-partisanship, ideology, proportion of working-class whites in a state, economic conditions-with very little play left. If you look at Trump's vote share across states, the variations in these known causes explain almost all of the variations in the voting patterns. To a level rarely seen in statistical analyses, Trump's successes and failures are fully accountable. The unmeasurable X factor of Russian influence does not have an unexplained hole to fill.
The ability to account for the Trump vote totals may strike some readers as surprising. Wasn't the 2016 result the opposite of what pollsters and election scholars had predicted? Respected analysts gave Trump only a 15 percent chance (The New York Times) or 1% chance to win (Princeton Election Consortium). The highest evaluation of Trump's likelihood of winning was by Nate Silver of Five Thirty Eight at 29 percent. The reason Abramowitz can account for the vote totals in a way that earlier analyses could not is his inclusion of the variables we now recognize is at the heart of support for Trump.
The first is the deep influence of partisanship: the vote share gained by Romney in 2012 is strongly related to the 2016 vote. Regardless of the feeling that Trump is decidedly different in personality and rhetoric from previous Republican candidates, partisan attachment overcame those qualms much more than many observers expected. The reason is a stronger feeling: dislike for the opposition. The growth of emotional disgust at the opposing team-what scholars call affective polarisation-meant that even those who disliked Trump personally also disliked Clinton more.
The second variable that explains a great deal of the outcome is the proportion of working-class whites in a state. This accounts for a remarkable amount of the variation in Trump support, and it was simply missing from the mainstream pre-election analyses. We now know that a great deal of the misunderstanding of the 2016 dynamic came from ignoring or misunderstanding the psychology of working-class voters, especially white male ones who previously voted for Obama but felt neglected by Clinton.
The first variable is a means of gauging the Republicans who supported Trump even though they don't like him; the second is a gauge of the unanticipated voters who actually do like him. Together they account for Trump support to a truly remarkable degree, with no additional influence needed.
The most convincing part of Abramowitz' analysis is that in the crucial states-Wisconsin, Michigan, and Pennsylvania-Trump slightly underperformed the expectations of the model that includes the major causal variables. Any Russian bump would have caused him to overperform, which is not what happened. (One could hypothesize that there might be some other factors that deflated Trump votes in those three specific states, while the Russian interference raised them in a way that is invisible in the analysis; but I have no idea what those negative factors specific to those states might be and haven't heard any good arguments for them). So I am pretty convinced by this analysis, as are many of my colleagues.
The reason the Russians did not appear to move votes is likely the same reason we have been discussing on this blog for some time-citizens come to their factual perceptions grounded in their own values and social groups far more than from any outside influences. Voters do not change their minds from propaganda any more than they do in response to real information. The power of prior beliefs is too much for new appeals to alter, especially for impressions of events. So those who thought ill of Hillary Clinton continued to hold those views if they saw Russian material that reinforced their priors. But few minds were changed.
So should we not worry about Russian interference? That is not at all what the Abramowitz evidence or any other analysis suggests. The point of Russian meddling may never have been changing votes, but instead was to sow distrust and break down American institutions. By making news reports as distrusted as possible, they hope that future reports of Russian wrongdoing in Europe and elsewhere will not generate opposition. Most analysts believe the Russian goal was to divide us so we cannot fight them. The success of that goal is harder to measure.
So the bad news is that few are listening to any evidence (fake or real), which explains the good news that Russian appeals had little effect on votes. The other bad news is that distrust continues to grow.
(David C. Barker, Ph.D is Professor of Government and Director of the Center for Congressional and Presidential Studies (CCPS) at American University.
Morgan Marietta, Ph.D is Associate Professor of Political Science at the University of Massachusetts Lowell, where he studies and writes about the political consequences of belief).