The Decline and Fall of Political Polling

                With all the chaos surrounding the presidential primary season, experts, journalists, and regular voters alike are relying on political polling to gauge the ebb and flow of the candidates’ fortunes.  This season, the polls are more important than any time in the past- the Republican debates have screened entry solely by poll numbers.  Yet even as their influence increases, both practical experience and experts show that polling accuracy, and thus relevance, is on the decline.

                Reputable polling institutes rely on reaching voters via landline telephones to ensure voter geography and demographics.  As cell phone usage rises rapidly, many people are unreachable by landline as they either don’t answer their phones or they simply eliminate their phone service altogether.  Many people consider the landline phone an antiquated technology; even in the Stevens Institute dorms, most of the landlines still present are no longer operational.  Palmer hall even has a “Telephone room” which now serves as a jantors’ closet.  According to a recent Politico article, almost half of American adults are not accessible by landline.  Traditional telephone, the enabling factor of reliable political polling, is dying.  As technology marches inexorably forward, it changes the landscape of almost every field, including politics and campaigning. 

                This rapid decline is accompanied by a severe increase in the importance of early-cycle polling.  The debates’ sponsors for the sprawling Republican field are determining who participates by national polling numbers; being relegated to an undercard event has crippled the campaigns of several candidates.  Cliff Zukin, past president of the American Association for Public Opinion Research, recently likened this early-cycle polling to “asking a scale that can only tell pounds to measure ounces.”  The ridiculously high margin of error on these polls, caused by the small (telephone contacted) sample sizes and limited demographic reach, severely limits the grip of polling on political reality, especially at such a nebulous stage in the election cycle.  The size of the Republican field, especially, doesn’t help matters; three or four points are seen as a significant jump for a candidate to make, but the margin of error in these polls is often around five points (as in the latest poll from CBS at the time of this writing).  To top of all of these mathematically obvious weaknesses, history shows the inaccuracy of early polls: this stage in the 2012 cycle, a Gallup poll showed Herman Cain, a failed candidate whose campaign never took root, within two points of Mitt Romney, the eventual nominee.  Pollsters have repeatedly warned the debate sponsors that polling at this early stage is an imprecise endeavor at best, but with the overflow of Republican candidates, polls are the only metric the sponsors have to narrow the field.

                Polling relies on landline telephone usage, a technology that is quickly becoming obsolete.  The rise of the consumer grade cell phone has decimated the usage numbers of landlines, stilting the demographics reachable and reducing the number who can be contacted via that method at all.  And yet as they lose their grip on real voters’ leanings, companies are relying on them more this cycle than ever before.  Polling needs to evolve drastically before its inaccuracies override any residual value this system still has left.

Sources:
http://www.politico.com/story/2015/10/polls-2016-gop-accuracy-214373
http://www.gallup.com/poll/149990/cain-surges-nearly-ties-romney-lead-gop-preferences.aspx