How to Dis­rupt Bias

Be­fore you can dis­rupt bias in your or­ga­ni­za­tion, you must first de­ter­mine pre­cisely where and how it is be­ing felt.

Rotman Management Magazine - - FRONT PAGE - By Sylvia Ann Hewlett, Ripa Rashid and Laura Sherbin

the Bos­ton Sym­phony Orches­tra fa­mously in­tro­duced IN 1952,

‘blind au­di­tions’ to test for gen­der bias in hir­ing, by hav­ing mu­si­cians au­di­tion be­hind a screen. At first, the screen made no dif­fer­ence in who made it past the first round of au­di­tions. Then, sus­pect­ing that a click-clack of high heels from fe­male mu­si­cians as they en­tered the room ren­dered the screen in­ef­fec­tive, the hir­ing com­mit­tee had the mu­si­cians re­move their shoes when they en­tered. In the ab­sence of gen­dered foot­falls, the judges trained their acute ears on the mu­sic it­self.

Other or­ches­tras be­gan adopt­ing the Bos­ton Sym­phony’s ap­proach in the 1970s, and by the 1990s, many saw in­creases in the num­ber of fe­male play­ers. The New York Phil­har­monic, for ex­am­ple, reached 35 per cent fe­male mu­si­cians by 1997 — a dra­matic in­crease over hav­ing had zero fe­male play­ers for decades. One study of 11 ma­jor or­ches­tras found that up to 55 per cent of their in­crease in new fe­male hires could be at­trib­uted to blind au­di­tions.

To­day, this case serves as a para­ble for bias-busting and has helped many busi­ness lead­ers un­der­stand the enor­mous role that bias can play in how they hire, eval­u­ate and pro­mote em­ploy­ees. Hir­ing de­ci­sions in par­tic­u­lar have at­tracted scru­tiny, with study af­ter study demon­strat­ing that the iden­tity im­plied by the name on a ré­sumé of­ten sways de­ci­sions.

Mean­while, com­pa­nies have also be­gun to rec­og­nize the pos­i­tive links be­tween di­verse lead­er­ship and busi­ness out­comes. As the work­force be­comes in­creas­ingly di­verse, busting bias is a busi­ness-crit­i­cal pri­or­ity.

In re­cent years, com­pa­nies have fo­cused a great deal on im­plicit or un­con­scious bias. In 1995, so­cial psy­chol­o­gists Mahzarin R. Banaji and An­thony Green­wald pi­o­neered the the­ory of ‘im­plicit so­cial cog­ni­tion’ and cre­ated the IAT (Im­plicit As­so­ci­a­tion Test) as a means of un­der­stand­ing bias. The IAT asks test tak­ers to quickly sort words and images into pos­i­tive and neg­a­tive cat­e­gories. Then, it gen­er­ates in­stant re­sults that tell test tak­ers whether they had more trou­ble as­so­ci­at­ing pos­i­tive char­ac­ter­is­tics with a given iden­tity — in other words, whether they are bi­ased against a cer­tain so­cial group.

Tal­ent spe­cial­ists across in­dus­tries have em­braced the IAT as a so­lu­tion to tack­ling the thorny is­sue of bias in the work­place. Each year, tens of thou­sands go through train­ing based on its the­o­ret­i­cal frame­work and learn about bias from Pro­fes­sor Banaji her­self. The IAT pro­vides a clear ex­plana­tory frame­work for in­ter­pret­ing a cor­po­rate re­al­ity: We all make rapid de­ci­sions in­formed by bi­ases we hold in our un­con­scious minds.

Anti-bias train­ing is par­tic­u­larly pop­u­lar in Sil­i­con Val­ley, where tech ti­tans like Face­book and Google man­date it, al­though a sim­i­lar trend is on the rise among For­tune 500 cor­po­ra­tions. Ac­cord­ing to one con­sul­tant, by 2019, half of the di­ver­sity strate­gies at large U.S. em­ploy­ers could be built around it. How­ever, de­spite all these ef­forts, gen­der rep­re­sen­ta­tion in the top ranks of large cor­po­ra­tions has re­mained roughly the same since 2000, and the num­bers for mi­nor­ity ex­ec­u­tives in the pri­vate sec­tor aren’t much bet­ter.

The Prob­lem: Bias Train­ing Has a Blind Spot

We have un­cov­ered a sig­nif­i­cant blind spot in im­plicit bias train­ing, which deeply af­fects com­pa­nies’ abil­ity to map, mea­sure and dis­rupt bias: The IAT and re­lated train­ing fo­cus on in­di­vid­u­als who har­bour bias. At first blush, it’s easy to see why, since man­agers and lead­ers are the ones mak­ing hir­ing, fir­ing and pro­mo­tion de­ci­sions — and who are in the po­si­tion to see po­ten­tial in and grant op­por­tu­ni­ties to the em­ploy­ees re­port­ing to them. Yet fo­cus­ing on those who hold un­con­scious bi­ases doesn’t al­low us to un­der­stand how bias is ac­tu­ally show­ing up in a com­pany from day to day.

There is scarce ev­i­dence to show how much in­di­vid­u­als ac­tu­ally act on their un­con­scious bi­ases when they make de­ci­sions. A man­ager who, ac­cord­ing to the IAT or any other such mea­sure, har­bours a host of bi­ases may or may not act upon these in the work­place — in other words, there is no proven cor­re­la­tion be­tween IAT scores and ac­tual de­ci­sion-mak­ing be­hav­iour. Re­ly­ing on the IAT, com­pa­nies are left with­out a clear un­der­stand­ing of how bias is ex­pe­ri­enced in their or­ga­ni­za­tion — or the cost it levies on their bot­tom line.

Com­pa­nies do have other ways of track­ing how em­ploy­ees ex­pe­ri­ence bias and dis­crim­i­na­tion. Many of­fer chan­nels — hu­man re­sources, hot­lines, om­buds­men—for em­ploy­ees to re­port such in­stances. Yet these chan­nels are dis­parate, scat­ter­shot and pri­mar­ily geared to­wards le­gal pro­tec­tion for the com­pany. They do not com­prise a ro­bust dataset from which to an­a­lyze bias’s true cost — or to sur­face ef­fec­tive so­lu­tions to re­me­di­ate it.

All too of­ten, the ‘so­lu­tion’ for mit­i­gat­ing bias is light-handed: Ask­ing in­di­vid­u­als who go through train­ing to try to re­sist their own bi­ases, or to call oth­ers out on their bi­ases in the mo­ment. This ig­nores the cul­ture and pro­cesses that gov­ern be­hav­iour — which might them­selves be bi­ased. Ask­ing in­di­vid­u­als to fix their own bi­ases fails to ad­dress sys­temic pres­sures that re­ward ‘fast think­ing’ or gut de­ci­sion-mak­ing—pres­sures that en­sure cur­rent lead­er­ship archetypes are repli­cated.

Decades of re­search af­firm that when sys­tems, pro­cesses and cul­ture are changed, in­di­vid­u­als are curbed from mak­ing re­flex­ive or bi­ased de­ci­sions. For ex­am­ple, say or­ches­tras had, in­stead of in­stalling screens, asked hir­ing com­mit­tee mem­bers to con­sider their own bi­ases when au­di­tion­ing mu­si­cians? The re­sults would not likely have been nearly as dra­matic.

A New Ap­proach to Bias

Armed with soft­ware that ‘blinds’ the vet­ting of job can­di­dates and with rig­or­ous pro­to­cols to cod­ify the in­ter­view process, some tal­ent spe­cial­ists have made sig­nif­i­cant progress in dis­rupt­ing bias in hir­ing de­ci­sions. But in de­ci­sions af­fect­ing tal­ent de­vel­op­ment and pro­gres­sion, bias is much harder to de­tect, let alone dis­rupt — and far less progress has been made in this area. That’s be­cause de­ter­min­ing who should ad­vance de­pends not just on assess­ing per­for­mance, but on assess­ing po­ten­tial as well—a no­to­ri­ously sub­jec­tive ex­er­cise.

Such as­sess­ments take place in the thou­sands of spot de­ci­sions lead­ing up to an an­nual per­for­mance re­view. Ev­ery day, week and month, man­agers make de­ci­sions about which tasks, clients, pro­jects and roles are as­signed to a given team mem­ber, and these seem­ingly mi­nor judg­ments in­form big­ger de­ter­mi­na­tions: Who will be awarded op­por­tu­ni­ties for learn­ing and de­vel­op­ment, in­ter­na­tional ro­ta­tions and stretch as­sign­ments? In­deed, life­time ca­reer out­comes are the cul­mi­na­tion of mo­men­tary as­sess­ments of in­di­vid­ual po­ten­tial — as­sess­ments that of­ten de­pend on sub­jec­tive as­sump­tions — a ripe breed­ing ground for bias.

To help root out bias in judg­ments of po­ten­tial, we have de­vel­oped ACE, a frame­work that cod­i­fies em­ployee po­ten­tial so that we might then use it to an­a­lyze bias in as­sess­ments of po­ten­tial. ACE is com­prised of six ar­eas com­monly used to as­sess an em­ployee’s po­ten­tial in both day-to-day and cycli­cal de­ci­sions — and all six ar­eas can be prone to deep sub­jec­tiv­ity.

1. ABIL­ITY: An em­ployee who went to a rep­utable school, for in­stance, might be per­ceived as more ca­pa­ble of tak­ing on new re­spon­si­bil­i­ties.

2. AM­BI­TION: An em­ployee who vol­un­teers for ex­tra work or has a lead role in spe­cial pro­jects is likely to be per­ceived as ‘driven to ad­vance.’

3. COM­MIT­MENT: An em­ployee who is seen put­ting in long days at the of­fice or mak­ing her­self avail­able nights and week­ends is likely to be per­ceived as ‘com­mit­ted to the job’.

4. CON­NEC­TIONS: An em­ployee in­volved in es­teemed or­ga­ni­za­tions or so­cial cir­cles out­side of work is likely to be per­ceived as hav­ing re­la­tion­ships that will ben­e­fit the busi­ness.

5. EMO­TIONAL IN­TEL­LI­GENCE: An em­ployee who senses what oth­ers need from him and mod­i­fies his in­ter­ac­tions ac­cord­ingly is likely to be per­ceived as hav­ing self-aware­ness, cul­tural aware­ness, and ap­pre­ci­a­tion for dif­fer­ence.

6. EX­EC­U­TIVE PRES­ENCE: An em­ployee who looks, sounds and acts like top com­pany lead­ers is likely to be per­ceived as hav­ing the abil­ity to com­mand a room.

To bet­ter un­der­stand the ex­pe­ri­ence and im­pact of bias against each ACE di­men­sion, we charted our sur­vey find­ings into a ‘heat

In de­ci­sions af­fect­ing tal­ent de­vel­op­ment and pro­gres­sion, bias is harder to de­tect.

map’. The Bias Heat Map (see Fig­ure One) shows neg­a­tive bias by tal­ent co­hort against each of the ACE frame­work’s six ar­eas of po­ten­tial. The or­der of the tal­ent co­horts re­flects the preva­lence of ACE bias among our sur­vey re­spon­dents: those least likely to re­port ACE bias are at the far left, on the yel­low end of the scale. Those most likely to re­port ACE bias are at the far right, on the red end of the scale. So, what does the Bias Heat Map tell us?

Ex­pected Find­ings: In some key ways, the heat map con­firms tra­di­tional hy­pothe­ses around bias and di­ver­sity. For in­stance, among our sur­vey re­spon­dents at large com­pa­nies, black, Latino, and Asian co­horts are more likely to re­port ACE bias than white co­horts are. Em­ploy­ees with dis­abil­i­ties — a group for which ex­clu­sion is well-doc­u­mented — are on the far right of the Heat Map, along with em­ploy­ees born out­side of the U.S.

Sur­pris­ing Find­ings: Fur­ther scru­tiny of the Heat Map re­veals some dis­tinct sur­prises. In our sam­ple of em­ploy­ees at large com­pa­nies, men were more likely to re­port ACE bias than women, and flex workers (male and fe­male) were less likely to re­port ACE bias than men as a whole. When we dug into in­ter­sec­tion-al­i­ties, even more sur­prises emerged: Asian men are more likely than Asian women to per­ceive ACE bias, es­pe­cially on am­bi­tion and abil­ity; black women are less likely to per­ceive bias than black men; and gay or bi­sex­ual men are more likely than les­bian or bi­sex­ual women to per­ceive bias.

As any pro­fes­sional who has ex­pe­ri­enced bias knows well, its im­pact can be pro­found. Feel­ing mis­judged on one’s pro­fes­sional po­ten­tial dam­ages pro­duc­tiv­ity and com­mit­ment. Re­mem­ber, ACE bias is the em­ployee’s per­cep­tion that su­pe­ri­ors mis-as­sess their po­ten­tial. Per­cep­tion of bias can lead em­ploy­ees to be­have in ways that cost their com­pa­nies. In­deed, we found that em­ploy­ees who per­ceive ACE bias be­have in three very costly ways:

Em­ploy­ees who per­ceive ACE bias are more THEY BURN OUT. likely than those who don’t to burn out. They’re also more likely to feel alien­ated at work, with­hold their ideas and lack pride in their com­pa­nies. In other words, they’re more likely to be dis­en­gaged.

ACE bias is linked to costly de­creases in re­ten­tion. THEY BUST OUT.

Those who per­ceive ACE bias, de­mor­al­ized by limited prospects for ad­vance­ment due to neg­a­tive as­sess­ments of their po­ten­tial, are far more likely to start look­ing for an­other job, or to want to leave their com­pa­nies within the year.

The most ex­treme re­ac­tion em­ploy­ees can have THEY BLOW UP. when they per­ceive ACE bias is to en­gage in small (or large) acts of sab­o­tage. That can mean in­ten­tion­ally fail­ing to fol­low through on an im­por­tant as­sign­ment or speak­ing neg­a­tively about their com­pany in pub­lic — for ex­am­ple, on so­cial me­dia. Blow­ing up is rare, but it car­ries enor­mous costs. Just think of the dam­age whistle­blow­ers have in­flicted on their em­ploy­ers over the years when they find their com­pa­nies lack the sys­tems and pro­cesses to re­spond to their con­cerns. The same is true when bias goes unchecked.

Dis­rupt­ing Bias: A Three-part Strat­egy

How can we de­ter­mine which in­ter­ven­tions dis­rupt bias? By look­ing at which in­ter­ven­tions ‘cool down’ our Heat Map. We tested sev­eral po­ten­tial so­lu­tions, and dis­cov­ered three to be par­tic­u­larly ef­fec­tive in min­i­miz­ing ACE bias across tal­ent co­horts.

To mea­sure di­ver­sity in lead­er­ship, we

1. DI­VER­SIFY LEAD­ER­SHIP. asked sur­vey re­spon­dents, ‘Which kinds of di­ver­sity do you see in your com­pany’s lead­er­ship?’ Then, we asked them to check any of the fol­low­ing that ap­ply:

• Gen­der

• Race/eth­nic­ity

• Age

• Re­li­gious back­ground

We con­sider the lead­er­ship of a re­spon­dent’s com­pany to be in­her­ently di­verse if there are at least three types of di­ver­sity rep­re­sented. Di­ver­sity in lead­er­ship is cru­cial to dis­rupt­ing ACE bias. In­her­ently di­verse ex­ec­u­tives demon­strate that dif­fer­ence is val­ued at their com­pa­nies. They are in a po­si­tion to en­dorse ideas from di­verse em­ploy­ees. They also pro­vide a ‘counter-stereo­typ­i­cal’ ex­am­ple for all em­ploy­ees, ex­pand­ing their no­tions of who suc­cess­ful top lead­ers are, un­der­cut­ting deep-seated bi­ases.

In or­der to truly see 2. AD­VANCE AN IN­CLU­SIVE LEAD­ER­SHIP CUL­TURE. the po­ten­tial of di­verse em­ploy­ees on teams, lead­ers have to get be­yond gut-level as­sump­tions. They can do so by fos­ter­ing a ‘speak-up cul­ture’ where every­one feels wel­come and in­cluded, free to share their ideas and opin­ions, and con­fi­dent that their ideas will be heard and rec­og­nized. With that kind of cul­ture, in­di­vid­u­als of all back­grounds have the op­por­tu­nity to sub­vert oth­ers’ as­sump­tions or bi­ases about them. In other words, they are in­clu­sive lead­ers.

In­deed, our data shows that with in­clu­sive team lead­ers,

em­ploy­ees at large com­pa­nies are 87 per cent less likely to per­ceive ACE bias and 39 per cent more likely to be en­gaged. For ex­am­ple, ev­ery man­ager at Sodexo is ex­pected to pri­or­i­tize di­ver­sity and in­clu­sion (D&I) — and that ex­pec­ta­tion is backed by com­pen­sa­tion in­cen­tives. Fully 10 per cent of man­agers’ com­pen­sa­tion is de­ter­mined by their rat­ings on a D&I Score­card — based on both their abil­ity to meet di­ver­sity tar­gets in the com­po­si­tion of their teams and on their in­clu­sive ap­proach. “It’s ex­tremely im­por­tant to show we are will­ing to put our money where our mouth is on in­clu­sion,” says Stephen Dun­more, CEO of North Amer­ica Schools at Sodexo.

How can com­pa­nies CON­NECT DI­VERSE TAL­ENT TO SPON­SORS. di­ver­sify lead­er­ship if tal­ent that doesn’t fit nar­row lead­er­ship

archetypes con­tin­ues to ex­pe­ri­ence bias and stall on the way to the top? Five of our own stud­ies con­ducted since 2010 point to the so­lu­tion: Di­verse em­ploy­ees need spon­sors or se­nior-level ad­vo­cates to lever them into lead­er­ship, ef­fec­tively by­pass­ing or negat­ing the ef­fects of man­age­rial bias.

Top tal­ent isn’t spon­sored equally: Men are 46 per cent more likely than women to have a spon­sor, and Cau­casians are 63 per cent more likely than em­ploy­ees of colour to en­joy such ad­vo­cacy. Spon­sors ad­vo­cate for their pro­tégés’ ad­vance­ment — and they see fur­ther­ing a pro­tégé’s ca­reer as an im­por­tant in­vest­ment in their own. Like in­clu­sive lead­ers, spon­sors turn out to have a pro­foundly mit­i­gat­ing ef­fect on the ACE bias that em­ploy­ees per­ceive. Why? Be­cause spon­sors see their pro­tégé’s po­ten­tial clearly — and are will­ing to fight to make sure oth­ers see it as well. With spon­sors, em­ploy­ees at large com­pa­nies are 90 per cent less likely to per­ceive ACE bias and 21 per cent more likely to be en­gaged.

To en­sure more women and peo­ple of colour cir­cum­vent bias to break through to lead­er­ship, or­ga­ni­za­tions have crafted pro­grams that get se­nior lead­ers to feel com­fort­able ad­vo­cat­ing for di­verse in­di­vid­u­als. The re­sults are truly bias-busting: At one For­tune 100 fi­nan­cial ser­vices firm, 62 per cent of par­tic­i­pants in its women’s spon­sor­ship pro­gram have re­ported a change in their level of re­spon­si­bil­ity as a direct re­sult of the pro­gram.

Else­where, In­tel’s Ex­tend Our Reach pro­gram for women and peo­ple of colour has cul­mi­nated in mo­bil­ity for 40 per cent of par­tic­i­pants since its in­cep­tion in 2011, with 16 of them at­tain­ing vice-pres­i­dent level pro­mo­tions and one at­tain­ing the level of cor­po­rate vice pres­i­dent.

In clos­ing

For most or­ga­ni­za­tions, the so­lu­tion set for elim­i­nat­ing bias will be more com­plex than put­ting up a screen and ask­ing ap­pli­cants to re­move their shoes. It will have to tar­get both man­ager be­hav­iour and the cul­ture that shapes the em­ployee ex­pe­ri­ence. But first, it re­quires in­di­vid­u­als to un­der­stand pre­cisely where bias is felt, so that or­ga­ni­za­tions can map, mea­sure and dis­rupt bias where costs to in­di­vid­u­als and the or­ga­ni­za­tion are the great­est.

Sylvia Ann Hewlett is the founder and CEO of the Cen­tre for Tal­ent In­no­va­tion, and codi­rec­tor of the Women’s Lead­er­ship Pro­gram at Columbia Busi­ness School. Ripa Rashid is co-pres­i­dent at the Cen­tre for Tal­ent In­no­va­tion. Laura Sherbin is Co-pres­i­dent, Chief Fi­nan­cial Of­fi­cer and Direc­tor of Re­search at the Cen­tre for Tal­ent In­no­va­tion and an Ad­junct Pro­fes­sor at Columbia Univer­sity. They are the co-au­thors of Dis­rupt Bias, Drive Value: A New Path To­ward Di­verse, En­gaged, and Ful­filled Tal­ent (Rare Bird Books, 2017).

Newspapers in English

Newspapers from Canada

© PressReader. All rights reserved.