NEW STUDY: Research Suggests the Arbitrariness of Facial Features Affects Jurors’ Sentencing Decisions in Death Penalty Cases

A new study from Columbia University researchers indi­cates that jurors’ per­cep­tion of facial fea­tures in white defen­dants affects their sen­tenc­ing deci­sions, much like the bias­es that affect every day social inter­ac­tions and deci­sion mak­ing. Through four exper­i­ments with 1,400 vol­un­teers, ​“the researchers found that when real-world defen­dants have facial fea­tures that appear untrust­wor­thy, they are more like­ly to be sen­tenced to death than life in prison.” Particular facial fea­tures, such as a heavy brow or down-turned lips, are known to cause oth­ers to view an indi­vid­ual as untrust­wor­thy, despite these fea­tures pro­vid­ing no actu­al indi­ca­tion of one’s char­ac­ter. The study also found that par­tic­i­pants ​“were more like­ly to rec­om­mend a rul­ing against hypo­thet­i­cal defen­dants with an untrust­wor­thy facial appear­ance.” Using mugshots of 400 white Florida pris­on­ers sen­tenced to either death or life in prison, researchers asked par­tic­i­pants to decide who seemed trust­wor­thy based sole­ly on their faces. In mea­sur­ing implic­it social cog­ni­tion, the research shows that par­tic­i­pants har­bored ​“uncon­scious bias that pre­dict­ed who was ulti­mate­ly sen­tenced to death.”

As part of this study, researchers also devel­oped an inter­ven­tion method with the inten­tion of break­ing par­tic­i­pants’ reliance on facial stereo­types. Those who under­went the train­ing uncon­scious­ly stopped rely­ing on facial stereo­types in their sen­tenc­ing deci­sions, ​“while par­tic­i­pants in a con­trol group who nev­er received train­ing remained strong­ly biased.” The train­ing inter­ven­tion employed by the research team, led by asso­ciate pro­fes­sor of psy­chol­o­gy Jon Freeman, ​“works by mak­ing the implic­it link in people’s minds between cer­tain facial fea­tures and an untrust­wor­thy reac­tion as no longer sta­ble or reli­able.” The train­ing con­di­tions par­tic­i­pants to asso­ciate untrust­wor­thy facial fea­tures with trust­wor­thy behav­ior, rid­ding par­tic­i­pants of the implic­it con­nec­tion between cer­tain facial fea­tures and untrust­wor­thi­ness. ​“These find­ings bol­ster pri­or work that facial stereo­types may have dis­as­trous effects in the real world, but, more impor­tant­ly, pro­vide a poten­tial inroad toward com­bat­ing these sorts of bias­es,” Prof. Freeman said.

The judg­ment of anoth­er person’s trust­wor­thi­ness or lack there­of is sig­nif­i­cant­ly influ­enced by both racial and gen­der bias­es. These bias­es exist along­side the facial-fea­ture bias­es exam­ined by researchers in this study. Prior research from pro­fes­sors at Stanford, UCLA, Yale, and Cornell focus­ing on the per­cep­tion of stereo­typ­i­cal­ly black facial char­ac­ter­is­tics indi­cates that these char­ac­ter­is­tics affect jurors’ deci­sions in death penal­ty sen­tenc­ing. Using data from a 1998 study in Pennsylvania by Professor David Baldus, the group, led by Professor Jennifer Eberhardt, found that among Black defen­dants who killed white indi­vid­u­als, the more stereo­typ­i­cal­ly Black a defendant’s facial fea­tures are per­ceived to be, the more like­ly that per­son is sen­tenced to death.

Columbia researchers con­duct­ed their study using images of sole­ly white male faces in order to con­trol for these race-based bias­es. According to Prof. Freeman, ​“by expos­ing a cog­ni­tive path­way toward erad­i­cat­ing facial stereo­types, future research must inves­ti­gate whether this train­ing could be broad­ly applied and how to ensure the bias reduc­tion per­sists over time.”