MY VIEW

Trolls and polls

BY JOHN PACE
Posted 10/27/22

In recent times and with more urgency, the veracity of polls has come into question. Famously, in the 2015 presidential election and in the face of what certain polls then seemed to say, how did Tr**p surprisingly outperform what was predicted?

This item is available in full to subscribers.

Please log in to continue

Log in
MY VIEW

Trolls and polls

Posted

In recent times and with more urgency, the veracity of polls has come into question. Famously, in the 2015 presidential election and in the face of what certain polls then seemed to say, how did Tr**p surprisingly outperform what was predicted? Further, why do more recent polls seem to vary somewhat inconsistently with each other? The exact responses may be elusive, but there are some definite things that we can say. 

Polls work on the principle that at any given moment in time, a (randomly chosen) small sample of a larger population will accurately reflect a truth about that population. Not unlike when you taste a spoonful of gravy in the pot and conclude something about the general taste of that gravy, without eating the whole potful.  But the accuracy of your conclusion about the gravy depends on it having been properly stirred, so that every spoonful is now pretty much the same as every other, and all are equally likely to be chosen. 

Statisticians are all very good with the well-known scientific details of polling, and so any errors in the technical aspects of sampling, generally, have been and continue to be exceedingly rare. 

But, consider a different aspect. Polls work on another principle, namely, that respondents are telling the truth in their responses. Suppose the level of cynicism and distrust in our political discourse has gotten so caustic that a particular population stratum has decided to include pollsters on the “enemy list” and thus refuses even to cooperate or perhaps purposely engages in deceptive responses.  As such, a significant uncooperating stratum of poll respondents might skew the results to the extent that any notion of “within the three-percent margin of error” would no longer apply. Even when the various polls are aggregated (averaged) this skewed poll result might invalidate at least some of the conclusions. Not unlike a very precise clock that is set to the wrong time, the readings would be very scientifically fine but, nonetheless, give the incorrect time.

None of this presupposes a conscious conspiracy out there. It may just be that “like minds think alike.” Perhaps a negative view of polls and pollsters is more likely among those who are inclined to see experts and their work as part of some kind of noxious plot to be resisted, attributed to part of that “damn pencil-neck bureaucracy” that runs the government—even though the pollsters are generally private companies, and only rarely have anything at all directly to do with the government.  

Accurate or not, political campaigns depend on the narrative that polling helps to provide, and clearly, as polling has increasingly unfolded as a big, big and expanding business, it will not go away. Perhaps pollsters might want to reconsider both their assumptions about respondents and their methods of outreach to them. Currently, we should likely all reconsider what polls can definitively and factually tell us. 

John Pace lives in Honesdale, PA. He is a retired math professor.

polls, voter fraud, elections

Comments

No comments on this item Please log in to comment by clicking here