We can know if Rus­sia gave the elec­tion to Trump

The Washington Post Sunday - - SUNDAY OPINION - CHRIS­TIAN CARYL Ex­cerpted from wash­ing­ton­post.com/peo­ple/chris­tian-caryl

By now, it’s a well-es­tab­lished fact that Rus­sia tried to in­ter­fere in the 2016 vote. Yet we still don’t know whether those ef­forts made the dif­fer­ence. Given just how nar­row Trump’s mar­gin of vic­tory was — fewer than 80,000 votes in three key swing states — it stands to rea­son that any help he re­ceived from Moscow could have helped him to win. But we can’t be sure — be­cause we don’t know how many U.S. vot­ers were per­suaded by Rus­sian in­for­ma­tion op­er­a­tions to change their votes (or to stay away from the polls al­to­gether). Isn’t it time we tried to find out? Many com­men­ta­tors seem to as­sume that we’ll never be able to know. But Si­nan Aral, a pro­fes­sor at the Mas­sachusetts In­sti­tute of Tech­nol­ogy, says that’s mis­guided. “It is pos­si­ble to know,” he told me, “with a cer­tain de­gree of sta­tis­ti­cal con­fi­dence, the like­li­hood that Rus­sian in­ter­fer­ence changed the re­sults.”

It’s ex­tremely hard to do, he warns. But if we can mar­shal the will, we can get much closer to the truth.

Take, for ex­am­ple, those three cru­cial swing states — Penn­syl­va­nia, Wis­con­sin and Michi­gan. Aral says he and his col­leagues want to study the Rus­sian in­flu­ence cam­paign in pre­cisely this geo­graph­i­cal con­text. The MIT scholars have de­vel­oped a ro­bust method­ol­ogy for as­sess­ing how so­cial me­dia cam­paigns in­flu­ence the be­hav­ior of their tar­gets — and now they want to bring it to bear on the Rus­sian med­dling in 2016.

Aral and his MIT re­search part­ner Dean Eck­les sent me what they call a blue­print for such a study. They pro­pose ze­ro­ing in on the is­sue of causal­ity by an­a­lyz­ing how dif­fer­ent lev­els of dis­in­for­ma­tion changed be­hav­ior and opin­ions. “For ex­am­ple, Face­book and Twit­ter con­stantly test new vari­a­tions on their feed rank­ing al­go­rithms, which cause peo­ple to be ex­posed to vary­ing lev­els of dif­fer­ent types of con­tent,” they write. Given ac­cess to ad­e­quate data, the re­searchers claim they can es­ti­mate the im­pact of the Rus­sian in­flu­ence cam­paign in Michi­gan, Wis­con­sin, Penn­syl­va­nia and Flor­ida “with 95% to 99% con­fi­dence.”

To con­duct such a study prop­erly, we’d prob­a­bly need far more in­for­ma­tion from the so­cial me­dia plat­forms than they’ve been will­ing to re­lease so far. (We still don’t know ev­ery­thing that Face­book and Twit­ter know, for ex­am­ple.) And it cer­tainly wouldn’t hurt to know more about how the Rus­sians did their tar­get­ing and any of the help they re­ceived on that front from out­siders.

To be clear, we don’t need to do this to de­ter­mine whether Trump col­luded; the Mueller investigation has al­ready re­vealed plenty on that score, and there’s sure to be more to come. The point is to get a more pre­cise un­der­stand­ing of how on­line cam­paigns af­fect our real-world be­hav­ior — some­thing we’re only just be­gin­ning to con­front. We need to know for the sake of the fu­ture of U.S. democ­racy.

Newspapers in English

Newspapers from USA

© PressReader. All rights reserved.