Problematic Prognostications

While of course the upcoming presidential election teems with uncertainties, Faigy Grunfeld’s review in Hamodia Prime of the predictive shortcomings of polls leaves a few points unconsidered (“Polling the Polls,” Hamodia Prime, July 22, 2020).

For one thing, third-party candidacies in 2016 accounted for over seven million votes, six million of which were cast for the Libertarian and Green parties. Both were seen by the Clinton camp as diversionary, attracting voters who might otherwise have sided with the Democrats. Indeed, the Clinton campaign exhorted these dissidents to turn to her instead. In any event, the third-party variable has been largely excised from this year’s equation. Moreover, according to the Cooperative Congressional Election Study, 12% of Bernie Sanders’ supporters actually voted for Donald Trump; this time, however, the senator from Vermont stands alongside Joe Biden.

In addition, the piece ignores the fact that Hillary Clinton won nearly three million more popular votes than Donald Trump; indeed, the latter’s 46.09% proportion in 2016 falls notably below John Kennedy’s 49.72%, the latter figure attained in what Grunfeld terms “one of the tightest elections in the 20th century.” In fact, Trump’s electoral college percentage of 57.3 actually fell short of Kennedy’s 58%.

And at least one of my eyebrows was hoisted upward by Joshua Sandman’s characterization of Biden as “inarticulate.” While it is true that the Democratic designee-apparent is no master of the spoken word, the fact that much of what Donald Trump says is not printable in family magazines makes comparisons elusive. Nevertheless, it is fair to say that, measured against the stable genius’ expletive-deleted conversational bearing, Joe Biden’s rhetorical style sounds downright Shakespearean.


Avraham Katz, Edgware

Mrs. Grunfeld responds:

Thanks for your letter which highlights interesting points.

Dr. Norpoth’s primary model is one-dimensional, or perhaps two-dimensional. It only considers the following: how well each candidate did in the primaries, and what the country is “due for” in terms of political party (the pendulum swing is fairly predictable). He considers no other variable — all the other details do not factor in, and in 2020, there are a host of new factors analysts would want to consider. The article was focusing on Norpoth’s model because it goes counter to the dominant narrative out there, and Hamodia wanted to give readers another perspective.

Interestingly, while researching another piece, I stumbled upon some Pew research which seems to agree with Dr. Norpoth’s work: The candidate who does better in the primaries usually goes on to win the general.

However, polling, modeling and analysis on this subject are in abundance. It’s hard to take any of it too seriously.