All polling is suspect, and polling that occurs four months before election day is especially suspect. In addition to the normal vagaries of life that beset pollsters (especially in the last three election cycles), the passage of time and the increased focus of the general election campaign can change the dynamics of a race in ways that pollsters cannot really predict.
Hillary and Trump have only recently disposed of their intra-party demons, and it’s difficult to know how the public will react to the dynamics of their head-on confrontation in things like nationally televised debates.
This Bloomberg Poll, for instance, is less interesting for its top line result (which shows Trump trailing Hillary by 12) than for what it says about which attacks are likely to be effective against the other candidate. The poll “message tested” the most common attacks used by each candidate against the other, and asked voters whether the accusation (if true) bothered them “a lot,” “a little,” or “not at all.”
Here, for example, are the “message testing” results for Hillary’s most common attacks against Trump:
As you can see, Trump has several huge, gaping vulnerabilities that Hillary can exploit with even modest political skill. Trump’s comments about women, in particular, are a huge opening for Hillary to repeatedly pound between now and election day. Anything that scores “bothers me a lot” for over 60% of likely voters is fertile ground for basically infinite attack ads.
Now look at the “message testing” for Trump’s attacks on Hillary:
As you can see, Trump has some openings, but they are not nearly as wide or as easy for Trump himself to pitch. Part of message testing evaluates not only the attack, but the person who is making it. The attack “Donald Trump is sexist” is per se more effective coming from a woman. I don’t make the rules about how stupid people respond to political messaging, so don’t take it up with me. On the other hand, “Clinton is in bed with Wall Street” is significantly less effective coming from someone whose entire public persona is tied up in being one of the wealthiest New Yorkers in the world (whether that is true or not).
Conversely, some attacks that would be uniquely effective coming from Trump – e.g., “Clinton is a creature of the Washington establishment” are message testing duds. A full 46% per cent of respondents care exactly zilch about this. Similarly, Trump’s attacks on Clinton’s record as secretary of State – which have been some of the most disciplined attacks he has made – are similarly not landing. And none has flopped worse than the attack, which he maintained for three solid critical weeks, that Hillary Clinton is an enabler of Bill Clinton’s abuse of women.
And here’s what’s especially problematic for Team Trump: Trump seems to have no other response to the completely accurate charge that he has serially used offensive language towards women and treated them in a derogatory fashion. Over and over again throughout this campaign, whenever Clinton has hit Trump on this, Trump has fired back by calling Hillary an enabler of the Sexual Harasser-in-Chief. The problem is, the voters absolutely do not care and are moreover sick of hearing about it. Meanwhile, they care very much about Trump’s own problems.
Although we can’t say for sure, the overall picture that’s before us is that Clinton’s attacks on Trump are likely to be more effective than Trump’s attacks on Clinton. This points to an ever-widening gulf between the candidates as the race progresses. If Trump can’t find some new angles of attack, this race is likely to get worse, not better.