Bernie Sanders, facing a virtually insurmountable climb to win the Democratic nomination through pledged delegates, has turned to trying to convince superdelegates that he is more electable. Sanders is doing this by pointing at current polling which does indeed show him performing stronger against Donald Trump or the other Republicans than Hillary Clinton fares against the Republicans. As a pollster, I’d like to make sure folks understand that this “theory” has more than a touch of wishful thinking in it.
Bernie does indeed poll better against Trump than Clinton does in head to head matchups. This is driven in no small part by strong positive sentiments towards Sanders — which we measure through favorability scores which are currently a solid 10 percentage points higher than Clinton. Sanders’ national favorability is a new-found thing. He started this election cycle unfamiliar to most people and over the last year has built up a positive reputation. However, he has also been exposed to far, far fewer negative attacks over this time than other candidates and certainly compared to attacks Clinton has experienced over the last 30 years.
If Sanders’ candidacy were to move forward, he would suddenly be subject to intense negative attacks from the Republican party (indeed, the likely reason he has not been the subject of such attacks is due to the assumption by Republicans that Clinton will be the nominee – and why waste ad dollars on someone who won’t win?). Regardless, we wanted to model what impact this could have on Sanders’ “electability”, as a good proxy for what a Sanders win might yield at a general election. The chart below illustrates Clinton and Sanders relative standing compared to Donald Trump in our recent Ipsos/Reuters poll. In the left, larger box is the current polling data, with the associated ‘favorability’ ratings above the vote shares. The right box shows a model indicating the likely impact of Sanders’ favorability ratings declining by 5 or 10 percentage points – which we anticipate would occur should he win the nomination and be exposed to negative campaigning and advertising attacks.
This modeling shows that if Sanders’ favorability comes down by only 5 percentage points, his “electability” will be worse than Hillary Clinton’s, even though Clinton’s favorability ratings are still lower. Now this stylized model data may not reflect what happens in real life, but this modeling shows the resilience of Clinton’s electability built on decades of work against the relatively new-found position of Sanders.