Win­ning isn’t ev­ery­thing in pri­maries

The Denver Post - - NEWS - By Jonathan Bern­stein

Win­ning isn’t all that mat­ters in the early con­tests on the pres­i­den­tial nom­i­na­tion cal­en­dar. Can­di­dates seek to beat ex­pec­ta­tions, and the press and the par­ties judge the re­sults against what they thought would hap­pen.

In fact, beat­ing ex­pec­ta­tions can mat­ter even more than the raw re­sults be­cause it af­fects the amount and tone of the cov­er­age can­di­dates get. Those, in turn, can af­fect vot­ers in the next state on the cal­en­dar.

A clas­sic ex­am­ple: In 1984, Gary Hart was per­ceived to have “won” Iowa with 16 per­cent— no one ex­pected him to fin­ish that strongly— andWal­terMon­dale was thought to have “lost” with 49 per­cent, be­cause he failed to reach 50 per­cent. Hart then re­ceived so much pos­i­tive pub­lic­ity that he pulled ama­jor up­set in the New Hamp­shire pri­mary and went on to be­come a se­ri­ous com­peti­tor for the nom­i­na­tion.

Al­most ev­ery smart ob­server de­plores this state of affairs: Shouldn’t we just ac­cept the re­sults, rather than in­ter­pret­ing them through the bias of “ex­pec­ta­tions”?

Here’s why the ex­pec­ta­tions game can make sense— and how it can go wrong. For good, bad and ugly:

The Good. Done well, the ex­pec­ta­tions game is an im­por­tant cor­rec­tive to the big­gest flaw of the se­quen­tial nom­i­na­tion process, which is that early states are more im­por­tant than later states, but that no state is rep­re­sen­ta­tive of the na­tion as a whole. By tak­ing into ac­count the par­tic­u­lar de­mo­graphic fea­tures or re­gional place­ment of Iowa, New Hamp­shire, South Carolina and Ne­vada as a way to dis­count suc­cess and fail­ure, both the me­dia and party ac­tors are able to ad­just the dis­tor­tions of those states to a na­tional scale.

The Bad. Un­for­tu­nately, the ex­pec­ta­tions game is tricky, and it’s easy to get caught up in the can­di­dates’ spin about what they should be “ex­pected” to do in any state, or pre­vi­ous me­dia nar­ra­tives.

Ex­pec­ta­tions are rel­a­tively easy in gen­eral elec­tions, be­cause we know that most Demo­cratic vot­ers will sup­port the Demo­crat and most Repub­li­can vot­ers will choose the Repub­li­can. So it’s easy to con­clude that Democrats should do well in Rhode Is­land and that Repub­li­cans should win eas­ily in Wy­oming.

It’s a lot harder, how­ever, to fig­ure out which Repub­li­cans should do well in Iowa on Feb. 1. Yes, the cau­cuses usu­ally turn out more Chris­tian con­ser­va­tives than can be found in Repub­li­can pri­maries in most states. But what does that mean ( for ex­am­ple) for Marco Ru­bio’s proper ex­pec­ta­tion level? There’s no ob­vi­ous an­swer. And given that ev­ery can­di­date will at­tempt to set ex­pec­ta­tions low— so that even a poor re­sult can be spun as a win— it’s es­pe­cially hard for neu­tral ob­servers to get it right.

The Ugly. At its worst, the ex­pec­ta­tions game can wind up as non­sen­si­cal as its crit­ics fear. That hap­pens when ex­pec­ta­tions are set not by ob­jec­tive fac­tors, but by polls— so that “beat­ing ex­pec­ta­tions” just be­comes a ques­tion of do­ing bet­ter than the last round of polls.

In that case, the best that beat­ing ex­pec­ta­tions might in­di­cate is which cam­paign has the best get- out- the- vote op­er­a­tion. Most likely, it just ran­domly re­wards can­di­dates who hap­pen to have had bad luck in polling. There’s no rea­son to ad­just for those things, the way there is a rea­son to ad­just for lo­cal de­mo­graphic ad­van­tages.

Done cor­rectly, judg­ing can­di­dates by whether or not they beat ex­pec­ta­tions can be very pos­i­tive. It’s just very easy to get it wrong.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.