In the last week, a slew of national polls looking at a general-election matchup between Hillary Clinton and Donald Trump have been released. Despite varying sample sizes and methodologies, the results were strikingly consistent. In aggregate, according to the RealClearPolitics average, they show Trump narrowing Clinton’s lead to a single percentage point.
A closer look at two of the more well established polls used in the RealClearPolitics average reveals that movement toward Trump does not appear to be an artifact of unrealistic estimates of the composition of the electorate—the factor that plagued all those polls predicting a President Mitt Romney in 2012. It also shows the tightening of the race is not completely due to Republicans coming home to their presumptive nominee. Trump’s increased strength boils down to a shift in support among independents, and that should be of concern to Brooklyn.
The demographics of the polls are virtually identical, with both assuming the electorate to be about 70 percent white and with both showing Democrats benefiting from a 7-point advantage in party identification among the registered voters sampled. While we don’t know for sure what the composition of the electorate will be this fall, given long-term demographic trends and the historical nature of presidential-election voters, those proportions are reasonable and favorable to Democrats.
In 2012, the number of white voters in a survey could often explain differences in polls. In this instance, the difference (Trump narrowly ahead in ABC/Post and narrowly trailing in NBC/Journal) is entirely due to differences in the preferences of white voters in each survey, not the number of white voters. Trump wins whites by 24 points in the ABC/Post poll while he wins whites by 16 points in the NBC/Journal poll.
Overall, comparing the two most recent ABC/Post polls on the likely matchup between Trump and Clinton, Trump moved from a 9-point deficit in March to a 2-point advantage in May. Similarly, he closed his gap with Clinton by 8 points in the most recent NBC/Journal poll.
In the ABC/Post poll, among Republicans, Clinton does 6 points worse in the new poll than the last poll and Trump does 10 points better. This yields a net 16-point improvement among Republicans for Trump on the difference between the two candidates. That improvement for Trump’s margin over Clinton is 22 percentage points among independents. (In the March ABC/Post poll, Clinton led Trump by 9 points and in the most recent study, she trails him by 13.)
Multiplying the movement toward Trump among each party group by their share in the electorate shows how much of Trump's total movement can be attributed to that group.
As the table shows, yes, he gains from Republicans. But, of Trump’s gain from last month in the horse race, the largest comes from independents. To be sure, Clinton will get back a few Democrats, but she already has high loyalty (86 percent support) among members of her party and is not likely to increase an already large Democratic advantage (+7) in party ID among the electorate. At some point, she is going to have to win a greater share of independents. She can do that by winning a sort of independent with a particular set of attitudes.
With that in mind, one data point to pay attention to is President Barack Obama’s job approval number (remember him?). It is now over 50 percent, according to Gallup and a variety of other surveys, and that should be good news for Clinton. Even though Obama is not on the ticket, the approval rating of the incumbent president is an excellent predictor of the vote share of his party’s nominee. Clinton’s chances are greatly buoyed by an improving view of Obama’s job performance. That said, for Clinton to win, at some point, her vote share and favorability rating will need to come more into sync with the president’s.
That suggests this campaign will be about Clinton trying to earn—and turn out—the votes of those who approve of the job Obama is doing, but who currently have an unfavorable view of her. After she puts away Bernie Sanders, that is.
Ken Goldstein is professor of politics at the University of San Francisco and is Bloomberg Politics’ polling and political advertising analyst.