No, polling isn’t the prob­lem

Lodi News-Sentinel - - OPINION - R.J. May III is founder and pres­i­dent of Ivory Tusk Con­sult­ing, a po­lit­i­cal con­sult­ing firm in Lex­ing­ton, South Carolina. He wrote this for In­sid­

Ispent a frus­trat­ing evening on the phone with a client re­cently, a can­di­date run­ning for of­fice in South Carolina. I ex­plained why his cam­paign needs to be­gin polling now for a pri­mary elec­tion that's still eight months off. All I got in re­turn was an ear­ful of the usual gruff: Polling is un­re­li­able. Polling can't be trusted. Why waste money on polling? And so on and so on, a full hour's worth.

We po­lit­i­cal strate­gists hear that a lot th­ese days. It's the lin­ger­ing after-ef­fect of 2016 when Don­ald Trump pulled off the big­gest up­set since Harry Tru­man won in a squeaker nearly 70 years ear­lier.

Trump's un­ex­pected vic­tory se­verely rat­tled con­fi­dence in polling. But here's the thing: The polls them­selves weren't to blame. It was how they were in­ter­preted. Or, more ac­cu­rately, mis­in­ter­preted.

Con­trary to the way it ap­pears, a pres­i­den­tial con­test isn't one big elec­tion where the can­di­date who re­ceives the most votes wins. It's ac­tu­ally 50 sep­a­rate state elec­tions held on the same day. Poll re­sults must be in­ter­preted within that frame­work rather than on a sin­gle na­tional scale.

Then there's the mat­ter of method­ol­ogy, of qual­ity over quan­tity. Does the poll ac­cu­rately re­flect and sam­ple the pop­u­la­tion the poll­ster is seek­ing to rep­re­sent? That's es­sen­tial be­cause, much like pol­i­tics as a whole, polling is both an art and a science. Only polls that uti­lize the ap­pro­pri­ate com­bi­na­tion of the two can be trusted.

And even when un­re­li­able polls are re­moved from the equa­tion, there's still the es­sen­tial mat­ter of cor­rectly in­ter­pret­ing what the data ac­tu­ally tell us. Re­mem­ber, polls aren't al­ways seek­ing to pre­dict the win­ner of an elec­tion. And they're cer­tainly not mag­i­cal crys­tal balls.

Rather, a poll is a snap­shot cap­tur­ing a spe­cific mo­ment in time. It shows what peo­ple were think­ing at the pre­cise mo­ment the poll was taken. Even then, it con­tains sev­eral wild cards with the po­ten­tial to se­ri­ously al­ter the cam­paign land­scape in the blink of an eye. For ex­am­ple, a sig­nif­i­cant num­ber of un­de­cided vot­ers could make up their minds a few days after polling is con­cluded, thus in­val­i­dat­ing the ear­lier find­ings. Some­times that even hap­pens in a mat­ter of hours rather than days.

Per­haps most im­por­tant, it's cru­cial to know at the out­set what you want the poll to do. For in­stance, is its pur­pose to help pre­dict the out­come of an elec­tion, or to guide mes­sag­ing and track voter at­ti­tudes through­out the cam­paign? That's an im­por­tant dis­tinc­tion be­cause while the two ap­pear in­ter­con­nected on the sur­face, they're quite dif­fer­ent in prac­tice. I'd much rather my­self, and our can­di­dates, be armed with data over a pre­dic­tion.

Fi­nally, the gen­eral pub­lic loves ob­sess­ing about what are gener­i­cally la­beled "the polls" at the na­tional level. Ca­ble net­works such as CNN, Fox, MSNBC and oth­ers breath­lessly an­nounce, "Look at this! Trump is trail­ing all lead­ing Demo­cratic con­tenders in a newly re­leased na­tional poll." That's grip­ping ma­te­rial for keep­ing view­ers glued to the screen. But it doesn't re­flect what's go­ing on be­neath the sur­face.

It be­hooves those of us who view pol­i­tics un­der a mi­cro­scope, such as po­lit­i­cal strate­gists and can­di­dates on down-bal­lot races, to care more about what moves num­bers rather than the daily horse race at the top of the ticket. Be­cause that's where polling is truly in­valu­able.

The right ques­tions, asked the right way to the right tar­geted group of peo­ple, can pro­vide a trea­sure trove of vi­tal in­for­ma­tion. They re­veal which spe­cific is­sues the elec­torate truly cares most about. (And be­lieve me, some­times those re­sults sur­prise even vet­eran cam­paign op­er­a­tives.) They in­di­cate how the elec­torate will re­spond to spe­cific mes­sag­ing, such as a cam­paign that's driven by the can­di­date's per­sonal bi­og­ra­phy, or whether is­sues are more up­front and cen­ter in the cur­rent elec­tion cy­cle. How ef­fec­tively a cam­paign ap­plies that in­for­ma­tion then af­fects the head-to-head or "horse race" sce­nario at the top of the ticket.

But the chal­leng­ing as­pect of all this for both those of us who work in the po­lit­i­cal arena and for the gen­eral pub­lic that fol­lows pol­i­tics is 2016's stain still clings to all polling, ev­ery as­pect and at ev­ery level.

I say those who use that ex­cuse to dis­re­gard the re­sults of re­li­able polls do so at their own peril. And I didn't need to con­sult a poll on it, ei­ther.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.