QUES­TIONS FOR Iris Bohnet

A lead­ing be­havioural econ­o­mist talks about re­mov­ing work­place bi­ases with ‘be­havioural de­sign’.

Rotman Management Magazine - - CONTENTS - In­ter­view by Karen Chris­tensen

There is some dis­agree­ment about the ‘busi­ness case’ for gen­der equal­ity. What is your take on it?

The dis­agree­ment is jus­ti­fied. The fo­cus to date has largely been on the di­ver­sity of cor­po­rate boards and se­nior man­age­ment teams, and the prob­lem is, we don’t have the data re­quired to make solid con­clu­sions. Even when we find a cor­re­la­tion be­tween gen­der di­ver­sity on a board and a com­pany’s per­for­mance, we have no way of prov­ing that there is a causal re­la­tion­ship go­ing on.

Re­cently, a meta-analysis came out, sum­ma­riz­ing abou t 120 stud­ies, and it did find a small pos­i­tive cor­re­la­tion be­tween gen­der di­ver­sity and over­all firm per­for­mance. But again, this was a cor­re­la­tion, not cau­sa­tion. If we want to es­tab­lish causal­ity, we will have to cre­ate teams ran­domly and mea­sure whether the more di­verse teams out­per­form the ho­mo­ge­neous teams. Some of the best work in this area has been done in the realm of ‘col­lec­tive in­tel­li­gence’ (i.e. the in­tel­li­gence of groups). This re­search has found a strong causal re­la­tion­ship be­tween gen­der di­ver­sity and team per­for­mance across many dif­fer­ent tasks.

As a re­sult, I be­lieve we have enough ev­i­dence at the mi­cro level that a busi­ness case ex­ists. How­ever, I’d love to see us move this dis­cus­sion be­yond a num­bers game, and start to think more about fos­ter­ing in­clu­sive be­hav­iour.

How do you de­fine ‘be­havioural de­sign’ ?

The re­search shows that we can’t help but put peo­ple into cat­e­gories, and be­havioural de­sign builds upon this el­e­ment of how our minds work. Ba­si­cally, it uses be­havioural in­sights to de-bias or­ga­ni­za­tional prac­tices and pro­ce­dures, rather than fo­cus­ing on chang­ing mind­sets. Within an in­di­vid­ual mind, bi­ases tend to oc­cur au­to­mat­i­cally and un­con­sciously, and it’s re­ally hard to change that. It’s much eas­ier to take steps to de-bias an or­ga­ni­za­tion.

Do di­ver­sity train­ing pro­grams work?

We don’t re­ally know, be­cause most or­ga­ni­za­tions don’t mea­sure the re­sults — and the few that do have gen­er­ally found that they don’t work. We have some cor­re­la­tional data look­ing at whether or not a com­pany has a di­ver­sity train­ing pro­gram and the ac­tual di­ver­sity of its work­force, and in short, that cor­re­la­tion does not ex­ist. So the pic­ture is not op­ti­mistic.

A few com­pa­nies are try­ing in­no­va­tive ap­proaches — from im­plicit bias train­ing to pro­grams aimed at spe­cific in­equal­i­ties. Carnegie Mel­lon’s Linda Bab­cock and Ge­orge Loewen­stein have re­searched the ef­fec­tive­ness of var­i­ous de-bi­as­ing tech­niques. One in­ter­ven­tion they stud­ied is ‘per­spec­tive tak­ing’, which sim­ply means try­ing to walk in your coun­ter­part’s shoes, take their per­spec­tive and un­der­stand where they are com­ing from. For ex­am­ple, ‘walk­ing in an el­derly per­son’s shoes’ by writ­ing an es­say from their per­spec­tive was shown to re­duce stereo­types about the el­derly.

Bab­cock and Loewen­stein also ex­per­i­mented with a ‘con­sider the op­po­site’ strat­egy, which in­volves be­ing your own devil’s ad­vo­cate and ques­tion­ing your as­sump­tions — ac­tu­ally com­ing up with ar­gu­ments for why your think­ing might be wrong. This has been shown to work — but it re­quires a lot of ma­tu­rity and self-aware­ness to be able to ques­tion your­self. It’s eas­ier if some­one else does the ‘heavy lift­ing’ for you.

Given all the ev­i­dence, I would urge com­pa­nies to fo­cus their train­ing pro­grams on ca­pac­ity build­ing and adopt the ‘un­freeze-change-re­freeze’ frame­work — a method bor­rowed from my Har­vard col­league, Max Baz­er­man. Suc­cess­ful ‘un­freez­ing’ hap­pens when peo­ple start to ques­tion their cur­rent strate­gies and be­come cu­ri­ous about al­ter­na­tives. Once ‘un­frozen’, you spend some time on what your or­ga­ni­za­tion is cur­rently do­ing, and what could change. Fi­nally, you think of ways to ‘re­freeze’ the new in­sights gained and the new be­hav­iours learned. In the end, the path­way to be­havioural change may not be a change in in­di­vid­ual be­liefs, but in­stead a change in so­cially-shared def­i­ni­tions of ‘ap­pro­pri­ate be­hav­iour’.

One of the more re­cent ap­pli­ca­tions of Big Data in the work­place is ‘peo­ple an­a­lyt­ics’. Please de­scribe how it works.

This ba­si­cally en­tails bring­ing the rigour of your fi­nance or mar­ket­ing de­part­ment to HR, ar­gu­ing that data can help us bet­ter pre­dict, for ex­am­ple, the fu­ture per­for­mance of a par­tic­u­lar job can­di­date than the best in­ter­view ever could. It in­volves mov­ing away from in­tu­ition and build­ing on data.

The ques­tion is, What kind of data? Or­ga­ni­za­tions can use all sorts of data points, but one pow­er­ful ex­am­ple is ‘look­ing back­wards’: You can use data and ma­chine learn­ing to ba­si­cally learn from the past. For ex­am­ple, you could take a close look at the data points for ‘in­di­vid­u­als who have been highly suc­cess­ful’ in your or­ga­ni­za­tion: What are their shared char­ac­ter­is­tics? You might look at which uni­ver­si­ties they went to, and find that it’s a good thing not to come from an Ivy League school — or maybe that it’s bet­ter to have an Engi­neer­ing back­ground than a Math back­ground.

I’d love to see us move this dis­cus­sion be­yond a num­bers game, and start to think more about fos­ter­ing in­clu­sive be­hav­iour.

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.