In addition to voting intention, one important question to ask respondents to our polls is who they themselves think will win the election. Back in May, we at Redfield & Wilton Strategies asked respondents in six US swing states the following question: Taking aside whether you want him to win or to lose, how likely do you think President Donald Trump is to win or lose his re-election bid? In May, we found that in all states, a plurality of respondents believed that Trump was more likely to win than to lose.

However, the phrasing of this question, which we had initially incorporated when the Democratic nominee was still to be determined, is somewhat problematic. It presents itself as a question regarding Donald Trump, specifically, rather than the actual contest: Donald Trump versus Joe Biden.   Earlier this month, we re-phrased this question, asking respondents in a national poll who they thought was more likely to win. Nationally, a plurality of respondents (37% to 30%) thought Biden was more likely to win.  

A week later, in mid-June, responses to this question were more split in the swing states. In Pennsylvania (39%), Arizona (36%), North Carolina (36%) and Florida (38%), belief that Trump will emerge victorious has dropped considerably from a level of over 40% in these states in May (if we are to allow for comparisons between the two questions). Nevertheless, despite a drop-in support, in all four of these states, a larger proportion of the public believe that Trump will win than think Biden will win, which differs from our national poll.

In all six states, belief in who is going to win correlates strongly with partisan lines. For example, belief among 2016 Trump voters that their preferred candidate will be re-elected in November lies within a range of 62% to 74%. It is also interesting to note that a significant minority of former Hillary supporters suspect that Trump will have another four years at the White House: in Pennsylvania, for instance, 11% of those who voted for Clinton in 2016 believe that Donald Trump is set to win in November.

For context, respondents to the same poll in Pennsylvania largely stated that they intend to vote for Biden, a state that had voted Blue in the six elections prior to 2016. As such, likely Biden supporters appear careful to predict a Biden win, perhaps given that pre-election polls in 2016 had led many to believe that Hillary Clinton would win.

In Michigan (36%) and Wisconsin (34%), a plurality of respondents believe that Joe Biden will win in November. Notably, in these two states, 11% and 12% of those who voted for Donald Trump in 2016 forecast a Biden victory.

While this question can serve as a measure of confidence and belief among likely voters in their respective candidates, it remains to be seen whether the public’s predictions on who will win the Presidency will impact voter turnout in November. Strong belief that a certain candidate will win may motivate some voters to turnout, while convincing others that their vote will not matter. For example, if voters in these swing-states believe that their least favoured candidate will win, a fear of losing may well drive up turnout, yet voters may also be resigned to defeat and not make the effort to cast their ballot. Conversely, if members of the public think that their candidate will win, their positive outlook may motivate them to vote, yet they may also act with complacency and not turn-out in high numbers.

These results should furthermore be interpreted with caution, as a significant portion of the population said that they believe the major party candidates are equally likely to win or answered that they do not know. Therefore, although Biden commands a significant lead in many polls, many Americans clearly believe that the unique circumstances in the build-up to this election, and the clear weaknesses both candidates have, mean that the result remains too difficult to predict.  

Regardless, this question can be useful. Not only would it measure confidence among a candidate’s supporters, but it can also highlight whether an election result would come as a surprise to the public or as something expected. At this moment, given the roughly even split among the swing states, and the high proportion of respondents saying ‘equally likely’ or ‘don’t know,’ it is clear that swing states voters recognize that it is only June and there is a long way to go until November.