Why the polls were so wrong — 2020 US election

There are sev­er­al pos­si­ble rea­sons for the sur­veys to fail:

The turnout this year was unusu­al­ly high, the high­est it has been in more than 100 years. In the past, high turnout has most­ly result­ed in good Demo­c­ra­t­ic results. This is due to the fact that in pre­vi­ous elec­tions the “blue” can­di­dates — that is, the Democ­rats — were able to attract dis­pro­por­tion­ate­ly large num­bers of mem­bers of minori­ties to the bal­lot box in these cas­es. In the elec­tion this year, how­ev­er, Trump in par­tic­u­lar was able to mobi­lize his clien­tele. The pro­por­tion of white vot­ers with­out a col­lege degree was unex­pect­ed­ly high. The US polit­i­cal sci­en­tist Sal­va­tore Babones describes anoth­er rea­son in a guest arti­cle in the “Syd­ney Morn­ing Her­ald” as the “dirty secret of pub­lic opin­ion polls”. Accord­ing­ly, the response rates to tele­phone sur­veys in the USA have down­right col­lapsed. Accord­ing to the polling insti­tute Pew, 20 years ago 36 per­cent of those called took part in sur­veys by tele­phone. This fig­ure is now six per­cent. Babones also writes that there are rumors in the indus­try that only around three per­cent will par­tic­i­pate. His guess as to what could be behind this is: In an age in which almost every­one is equipped with a smart­phone, hard­ly any­one takes calls with a hid­den num­ber. How­ev­er, what speaks against this the­sis is that many online sur­veys were also not more precise.

An effect that also exists in oth­er coun­tries and that is also known in Ger­many could also have played a role. It is about the con­cept of social desir­abil­i­ty devel­oped in 1972 by soci­ol­o­gists Derek Philipps and Kevin Clan­cy. Accord­ing to this, some peo­ple refuse to take sur­veys or know­ing­ly give wrong answers because they fear that pro­vid­ing truth­ful infor­ma­tion vio­lates social norms. As a result, pre­dic­tions about opin­ions out­side of the social main­stream are much less reli­able. In elec­tions in Ger­many, this effect could be observed in the sur­pris­ing suc­cess of the Repub­li­cans in the state elec­tions in Baden-Würt­tem­berg in 1992 and in the state elec­tions in Sax­ony-Anhalt in 1998, where the right-wing extrem­ist DVU got 12.9 per­cent of the votes from a stand­still — last polls saw the par­ty at only six per­cent. And the elec­tion results of the AfD are tra­di­tion­al­ly char­ac­ter­ized by this effect — in some cas­es con­sid­er­ably. In the last state elec­tions in Baden-Würt­tem­berg, Rhineland-Palati­nate and Berlin, the polls on elec­tion day (exit polls) showed a devi­a­tion of 25 per­cent. In the 2013 fed­er­al elec­tion it was almost 50 percent.

In the US, this effect is also explained with “Shy Trump Vot­ers”, with “shy” Trump vot­ers. How big the effect real­ly is can­not yet be said. The Repub­li­can camp has also tried in the past to devel­op a polit­i­cal nar­ra­tive out of it — that of the “silent major­i­ty”, the silent major­i­ty that sup­pos­ed­ly stands behind Trump. This is sup­port­ed by the fact that a poll in Flori­da, which was con­duct­ed by com­put­er votes, was much clos­er to the actu­al result. Because in front of a com­put­er nobody has to be ashamed of their pref­er­ences. Against this, as men­tioned above, the fact that even anony­mous online sur­veys were some­times unable to pro­vide exact results. It is pos­si­ble that Trump was able to con­vince many Amer­i­cans in the last few meters in the elec­tion cam­paign. In con­trast to his com­peti­tor, the pres­i­dent held mass events in the impor­tant swing states despite the coro­na risk. The incum­bent may have been able to mobi­lize many vot­ers in the last few days before the elec­tion — and it was already too late for this devel­op­ment to be reflect­ed in the polls.

(Vis­it­ed 1 times, 1 vis­its today)



Tags: design TT Mod­ell­bahn TT H0 N schal­ten mod­elleisen­bahn bahn spiele­max preise 

Ein Reichsmarschall von Adolf Hitler hatte auch Märklin Modelleisenbahn Modelle > read more

Schreibe einen Kommentar