Journal of Law and Policy
*'/(
4
--/ ,.$'
*''*1.#$-)$.$*)'1*,&-. #6+-,**&'2)1*,&-,**&'1 /%'+
,.*!.# *,)(+'*2( ).1*((*)-).# 1)-2#*'*"2*((*)-
5$-,.$' $-,*/"#..*2*/!*,!, )*+ ) --2.# 1*/,)'-.,**&'2)*,&-.#- ) +. !*,$)'/-$*)$)*/,)'*!1)
*'$22)/.#*,$3 $.*,*!,**&'2)*,&-
*(( ) $..$*)
-#' 26' e Hidden World of Unconscious Bias and its Impact on the "Neutral" Workplace Investigator *'2
0$'' . #6+-,**&'2)1*,&-,**&'1 /%'+0*'$--
411
THE HIDDEN WORLD OF UNCONSCIOUS BIAS AND ITS
IMPACT ON THE “NEUTRAL” WORKPLACE
INVESTIGATOR
Ashley Lattal
*
Workplace investigations into complaints of harassment,
discrimination, and other allegations of workplace misconduct have
become a critical method for employers to establish that they have
complied with certain obligations to provide a discrimination-free
workplace. As a result, the fairness and effectiveness of the
workplace investigation process utilized by employers has
increasingly come under judicial scrutiny. The nature of workplace
investigations rests upon the assumption and expectation that
workplace investigators are capable of being impartial in making
findings of fact. For this reason, courts have identified the
impartiality of the investigator as a key tenet of a fair and effective
workplace investigation.
Yet, mounting scientific evidence indicates that our aut oma ti c
cognitive biases unconsciously steer us towards biased behavior.
Such unconscious biases present a significant hurdle to attaining
true impartiality and are likely to impact the workplace
investigator’s role.
Workplace investigators, who are not specifically regulated,
must educate th em se lv es as to the potentially detrimental impact
such biases can have at each stage of the process and make
*
Ms. Lattal is a founding partner of Shearer Lattal LLP, a Toronto firm
specializing in workplace investigations, workplace mediations, and training. Ms.
Lattal is a graduate of Cornell Law School (2000) and is called to the bars of New
York, Washington, D.C., and Ontario. She began her career practicing in
commercial litigation at a national law firm in Washington, D.C. and subsequently
moved to a litigation boutique where she developed an employment law practice.
She recently obtained an LL.M. in Alternative Dispute Resolution from York
University’s Osgoode Hall Law School in Toronto.
412 JOURNAL OF LAW AND POLICY
conscious and consistent efforts to reduce that impact. While
unconscious biases can be difficult to overcome, workplace
investigators can employ bias reduction strategies to attempt to
ensure that hidden biases do not taint their conclusions.
After exploring the workplace inve stigation process and the
critical role of neutrality in workplace investigations this article
examines the emerging research in the science of unconscious bias.
It then discusses court cases in which bias has or may have impacted
the processes and outcomes of workplace investigations. This article
concludes by proposing possible solutions to the problem of
unconscious biases in conducting workplace investigations.
INTRODUCTION
Workplace investigations occupy an increasingly important role
in identifying, understanding, and addressing workplace conflict.
Indeed, workplace in ve st ig at io ns into complaints of harassment,
discrimination, and other workplace misconduct are a critical
method for employers to establish that they have complied with
certain obligations, such as their “obligation to ensure a
discrimination-free work environment.”
1
As a result, the fairness
and effectiveness of the workplace investigation process utilized by
employers has increasingly come under judicial scrutiny.
2
Workplace investigators are tasked with making findings of fact that
may have an impact on the reputation and livelihood of individual
1
Northrop Grumman Corp. v. Worker’s Comp. Appeals Bd., 127 Cal. Rptr.
2d 285, 296 (Cal. Ct. App. 2002); see also U.S. EEOC, ENFORCEMENT GUIDANCE
ON VARIOUS EMPLOYER LIABILITY FOR UNLAWFUL HARASSMENT BY
SUPERVISORS, EEOC NOTICE 915-002 (June 18, 1999) [hereinafter EEOC
ENFORCEMENT GUIDANCE], http://www.eeoc.gov/policy/docs/harassment.html.
2
See STEWART S. MANELA, WORKPLACE INVESTIGATIONS 3,
http://www.americanbar.org/content/dam/aba/administrative/labor_law/meetings
/2009/ac2009/107.authcheckdam.pdf. “Employment practices experts frequently
testify regarding what constitutes reasonable investigation, and critique
employers’ investigations. Consequently, employers are increasingly being called
upon to explain and justify their investigatory techniques as reasonable and
appropriate in the face of these expert opinions.” Id. at 3 n.4. Cases considering
the reasonableness of workplace investigations are discussed further below in this
Article.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 413
workers.
3
The nature of workplace investigations rests on the
assumption and expectation that workplace investigators can make
impartial findings of fact.
4
Courts have therefore identified the
impartiality of the investigator as a key tenet of a fair and effective
workplace investigation.
5
In recent years, however, mounting psychological evidence has
suggested that achieving true impartiality requires more than simply
utilizing fair processes and ensuring there are no ties to the parties
involved.
6
The research indicates that automatic cognitive biases are
3
See id. at 3; SUBCOMM. OF THE EMP’T RIGHTS & RESPONSIBILITIES COMM.,
AM. BAR ASS’N, 2013 REPORT OF THE WORKPLACE I NVESTIGATIONS 3 (2013),
http://www.americanbar.org/content/dam/aba/events/labor_law/2013/03/employ
ment_rightsresponsibilitiescommitteemidwintermeeting/19_workplace.authchec
kdam.pdf (“The court also opined that the investigation lacked the ‘careful,
systematic assessments of credibility one would expect in an inquiry on which an
employee’s reputation and livelihood depended.’” (quoting Mastro v. Potomac
Elec. Power Co., 447 F.3d 843 (D.C. Cir. 2006))).
4
See generally EEOC ENFORCEMENT GUIDANCE, supra note 1 (“An
employer should set up a mechanism for a prompt, thorough, and impartial
investigation into alleged harassment.”); ASS’N OF WORKPLACE INVESTIGATORS,
GUIDING PRINCIPLES FOR INVESTIGATORS CONDUCTING IMPARTIAL WORKPLACE
INVESTIGATIONS, (Sept. 25, 2012) [hereinafter GUIDING PRINCIPLES],
http://www.aowi.org/assets/documents/guiding%20principles.pdf (“Guiding
Principle: The investigator should be impartial and objective, and have the
necessary skills and adequate time to conduct the investigation.”). For purposes
of this Article, the terms “impartial,” “neutral,” and “unbiased” are used
interchangeably. The meaning of these terms is discussed infra in Part II.B.
5
See, e.g., Ribeiro v. Can. Imperial Bank of Commerce, 1989 CanLII 4281
(Can. Ont. S.C.); Tong v. Home Depot of Can. Inc., 2004 CanLII 18228, para.
11–12 (Can. Ont. S.C.) (discussing the investigatory flaw of the floor manager’s
“untrained investigation methods,” where he “began his examination of [the
plaintiff]’s actions and movements with the firm belief that the plaintiff was a
wrongdoer”); Reeves v. Safeway Stores, 16 Cal. Rptr. 3d 717 (Cal. Ct. App. 6th
2004) (reversing summary judgment for a defendant employer where the evidence
“supports an inference that . . . [the] investigation of the alleged misconduct was
not truly independent”); Nazir v. United Airlines, 100 Cal. Rptr. 3d 296 (Cal. Ct.
App. 1st 2009) (reversing summary judgment for a defendant employer where the
investigation into discrimination claims was a “far cry” from impartial or
thorough, thus “raising a triable issue whether the investigation
was . . . appropriate”); Quela v. Payco-General Am. Credits, Inc., 84 F. Supp. 2d
956 (N.D. Ill. 2000).
6
See, e.g., Anthony G. Greenwald & Linda Hamilton Krieger, Implicit Bias:
Scientific Foundations, 94 CAL. L. REV. 945, 951 (2006) (discussing the new
414 JOURNAL OF LAW AND POLICY
deeply embedded within the human psyche, often run contrary to
our explicit beliefs, and unconsciously steer us towards biased
behavior and conclusions.
7
Such unconscious biases are likely to
impact the most important funct ions of workplace investigators,
including evaluation of witnesses and evidence, assessment of
behavior, recollection of facts, and judgment of wrongdoing.
Investigators may reach conclusions not on the basis of seemingly
strong objective evidence, but instead by unwittingly following
cognitive biases of which they are often unaware.
8
Workplace investigators should understand the potentially
detrimental impact such biases can have at each stage of the process
and make conscious efforts to reduce that impact. Understanding
these biases may be particularly important for workplace
investigators who investigate discrimination claims made by
disadvantaged groups—groups against which most of us have strong
unconscious biases that often run contrary to our explicit beliefs.
9
While unconscious biases can be difficult to overcome, workplace
investigators can employ bias reduction strategies to help to
eliminate the effects of unconscious biases in their conclusions.
Part I will explore the workplace investigation process, the
critical role of neutrality in the investigation process, and the largely
unregulated nature of the profession. Part II then examines relevant
developments in the science of unconscious bias. Part III will
discuss cases in which bias has or may have impacted workplace
investigations. Part IV will propose possible solutions to the
problem of unconscious biases in conducting workplace
investigations. Understanding unconscious biases and strategies
designed to reduce them can help to enhance the neutrality of
workplace investigations.
science of unconscious mental processes and how humans do not always have
conscious, intentional control over their mental processes that motivate their
actions); AMY OPPENHEIMER, THE PSYCHOLOGY OF BIAS: UNDERSTANDING AND
ELIMINATING BIAS IN INVESTIGATIONS (2012), http://amyopp.com/wp-
content/uploads/2013/07/Psychology_of_Bias_May_2012.pdf (discussing the
impact unconscious biases can have on investigators).
7
See, e.g., Greenwald & Krieger, supra note 6.
8
See generally OPPENHEIMER, supra note 6 (discussing the impact bias can
have on investigators).
9
See id. at 2 (discussing bias against disadvantaged groups).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 415
I. OVERVIEW OF THE WORKPLACE INVESTIGATION PROCESS AND
THE CRITICAL ROLE OF NEUTRALITY
A. The Workplace Investigation Process
A workplace in ve st ig at ion is a proc es s whereby an employer,
either internally or through an independent investigator, engages in
fact-finding to determine whether certain types of workplace
misconduct have or have not occurred.
10
Employers utilize
workplace investigations to respond to a variety of workplace
misconduct, incl uding allegations of harassment, discrimination,
threats or incidents of violence, unprofessional or inappropriate
workplace behavior, theft, and inappropriate use of company
information or assets.
11
In most cases, an internal complaint triggers
the investigation.
12
Allegations of discrimination and harassment are amongst the
most common types of complaints requiring investigation.
13
An
employer has a duty to provide a workplace free of discrimination
and harassment by taking “all reasonable steps to prevent
harassment and discrimination from occurring.”
14
In order for an
employer to meet its obligations in this regard, it must promptly,
thoroughly, and impartially investigate complaints of
10
See generally SOC’Y OF CORP. COMPLIANCE & ETHICS, GUIDE TO
CONDUCTING WORKPLACE INVESTIGATIONS (2008),
www.corporatecompliance.org/Portals/1/Users/169/29/60329/Workplace_Invest
igations_Guide.pdf (explaining that a workplace investigation involves fact-
finding to determine whether misconduct has occurred).
11
GARETH JONES, CONDUCTING ADMINISTRATIVE OVERSIGHT &
OMBUDSMAN INVESTIGATIONS 1, at 1 (2009); Manela, supra note 2, at 2.
12
FRAN A. SEPLER, FINDING THE FACTS: WHAT EVERY WORKPLACE
INVESTIGATOR NEEDS TO KNOW, at xi (2008).
13
See, e.g., Press Release, U.S. Equal Emp. Opportunity Comm’n, EEOC
Releases Fiscal Year 2015 Enforcement and Litigation Data (Feb. 11, 2016)
http://www1.eeoc.gov/eeoc/newsroom/release/2-11-16.cfm (providing “detailed
breakdowns of the 89,385 charges of workplace discrimination” the EEOC
received during the 2015 fiscal year, of which charges alleging harassment “made
up nearly 28,000 charges, or 31%”).
14
Northrop Grumman Corp. v. Worker’s Comp. Appeals Bd., 127 Cal. Rptr.
2d 285, 296 (Cal. Ct. App. 2d 2002) (quoting Cal. Gov’t Code § 12940 (West
2016)).
416 JOURNAL OF LAW AND POLICY
discrimination, harassment, and retaliation.
15
While an investigation
may not be necessary in all cases, employers who fail to conduct
reasonable and appropriate investigations run the risk of critically
damaging their legal position in the event that a party to the
investigation subsequently commences a lawsuit.
16
Although the workplace investigation process can vary
somewhat depending on company policy, an individual
investigator’s preferences, or the circumstances of the particular
case, it generally entails interviewing the complainant(s),
respondent(s), and relevant witnesses, and reviewing evidence to
make findings of fact about the workplace incident.
17
There is much
15
See EEOC ENFORCEMENT GUIDANCE, supra note 1 (“An anti-harassment
policy and complaint procedure should contain, at a minimum . . . [a] complaint
process that provides a prompt, thorough, and im partial investigation [among
other elements].”); Cal. Gov’t Code § 12940(j)(1), (k) (West 2016); Northrop
Grumman, 127 Cal. Rptr. 2d at 296 (“The employer’s duty to prevent harassment
and discrimination is affirmative and mandatory . . . Prompt investigation of a
discrimination claim is a necessary step by which an employer meets its obligation
to ensure a discrimination-free work environment.” (citations omitted)); see also
Malik v. Carrier Corp., 202 F.3d 97, 105 (2d Cir. 2000) (“[A]n employer’s
investigation of a sexual harassment complaint is not a gratuitous or optional
undertaking; under federal law, an employer’s failure to investigate may allow a
jury to impose liability on the employer.”); Melanie A. Kastl & Brian H. Kleiner,
Bias in Workplace Investigations and How to Minimise It, 25 M
GMT. RES. NEWS
227, 227–31 (2003).
16
See, e.g., Elgert v. Home Hardware Stores Ltd. (2011) 510 A.R. 1, para.
86–89 (Can. Alta. C.A.); see also CHRISTINE THOMLINSON ET AL., INVESTIGATING
ALLEGATIONS OF HARASSMENT AND WORKPLACE VIOLENCE: PRACTICAL
CONSIDERATIONS (June 2012), https://kmlaw.ca/wp-
content/uploads/2015/09/AH_InvestigatingAllegationsofHarassment1_JUN_6_1
2-1.pdf (“A seriously flawed or negligent investigation can be significant and
costly to employers.”).
17
See, e.g., SEPLER, supra note 12, at xiii fig.1. If there is no official
complainant, an investigator will interview those people that she determines likely
to have knowledge relevant to the investigation. She then prepares a report that
outlines the process followed, summarizes the evidence obtained in the interviews
and from any other sources, assesses the credibility of witnesses, and makes
findings as to whether the allegations are substantiated based on the evidence.
Depending on the mandate, the report will sometimes contain a legal conclusion
regarding whether the respondent has breached an applicable policy or statute. It
may also include recommendations as to how the organization ought to move
forward based on the investigator’s findings of fact. See, e.g., Margaret J. Grover
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 417
room for bias to seep into the investigation process, including during
initial discussions with the client or witnesses about the case, pre-
interview document review, interviews regarding the line of
questioning and the phrasing of questions, when deciding which
witnesses to interview and what additional evidence to review, and
when deciding which evidence to reference and which to ignore in
reports.
An effective investigation reaches accurate conclusions that
those affected by the investigation will accept as fair.
18
Courts have
found that a fair investigation is one that is unbiased, as well as
thorough and timely.
19
A fair investigation is achieved, in part,
through consistent use of a thoughtful, well-planned investigation
process.
20
Elements of such a process may include establishment of
a clear mandate, advising interviewees of the goal of conducting an
objective investigation, careful review of documents, and
communicating the result of the investigation to the parties.
21
Even when employers implement safeguards to attempt to
ensure a fair process, they may not be able to make any investigation
entirely unbiased, as investigators inevitably bring their
preconceived notions and unconscious biases with them.
22
Workplace investigations can leave a great deal of room for biases,
particularly unconscious ones, to steer the investigation towards
predetermined conclusions. While most investigators believe that
they make objective, impartial findings, the science concerning
unconscious biases indicates that this may not be the case.
23
& Lauren S. Antonio, Workplace Investigations – A Practical Approach, 30 FALL
BRIEF 8, 2000, at 9–10; Kastl & Kleiner, supra note 15.
18
See Kastl & Kleiner, supra note 15, at 227.
19
See MISSY DAVIDSON, THE IMPORTANCE OF INVESTIGATIONS IN THE
DISCIPLINARY PROCESS 2–3 (Jan. 2011),
http://www.npelra.org/files/public/CLRP/Davidson,Missy_CLRP_Paper_2011.p
df.
20
See, e.g., SEPLER, supra note 12, at 49–59.
21
See generally Kastl & Kleiner, supra note 15 (discussing the elements of
a workplace investigation).
22
Id. at 227–28.
23
See generally OPPENHEIMER, supra note 6 (discussing the kinds of
unconscious biases that can impact workplace investigators).
418 JOURNAL OF LAW AND POLICY
B. Critical Role of Impartiality in Workplace
Investigations
“[N]eutrality serves ‘as the antidote against bias.’”
24
A critical aspect of a fair workplace investigation process is the
investigator’s impartiality or lack of bias.
25
As with other dispute
resolution processes, the integrity of the workplace investigation
process rests upon the parties’ perception of the investigator as an
“unaligned participant.”
26
The employer’s theories, biases, or
preferences should not influence the workplace investigator in
determining the appropriate conduct of an investigation or its
outcome.
27
24
Carol Izumi, Implicit Bias and the Illusion of Mediator Neutrality, 34
WASH. U.J.L. & POL’Y 71, 76 (2010) (discussing mediation, another process in
which neutrality plays a crucial role).
25
See, for example, Silva v. Lucky Stores, 76 Cal. Rptr. 2d 382 (Cal. Ct.
App. 1998) and Fuller v. City of Oakland, 47 F.3d 1522 (9th Cir. 1995), two early
cases that established benchmarks by which many investigations are still judged,
including the need to remain objective. See also Reeves v. Safeway Stores, 16
Cal. Rptr. 3d 717, 736 (Cal. Ct. App. 2004) (holding that the investigation should
be “truly independent” and that the employer in this case “predetermined” the
terminated employee’s guilt without sufficient inquiry was evidence of retaliatory
discharge); Nazir v. United Airlines, 100 Cal. Rptr. 3d 296, 325 (Cal. Ct. App.
2009) (stating, in a race discrimination case, that the use of an investigator with
an “axe to grind” against the terminated employee could be evidence of pretext);
EEOC ENFORCEMENT GUIDANCE, supra note 1 (“An employer should set up a
mechanism for a prompt, thorough, and impartial investigation into alleged
harassment . . . . The employer should ensure that the individual who conducts the
investigation will objectively gather and consider the relevant facts. The alleged
harasser should not have supervisory authority over the individual who conducts
the investigation and should not have any direct or indirect control over the
investigation.”); Mark J. Flynn, Unlawful Retaliation and Workplace
Investigations—Lines, Bars and “X” Factors, COLO. LAW, Nov. 2013, at 85, 86
(“The integrity of an investigation is fundamental to its basic objective, getting to
the truth of a matter in dispute.”).
26
See GUIDING PRINCIPLES, supra note 4 (explaining that “the investigator
should be someone the participants view as impartial, though this may not be
possible in every case”). There are clear differences in the investigation and
mediation processes, including that an investigator’s aim is to make findings of
fact, possibly against the participant. See Flynn, supra note 26, at 86.
27
See SEPLER, supra note 12, at 127–28.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 419
An employee may face serious consequences as a result of an
investigation. The results can have a devastating impact on an
employee’s reputation, day-to-day work life, and his or her
livelihood. Arguably, the more detrimental the allegations, the more
seriously an investigator must take the responsibility of conducting
an impartial, unbiased investigation in order to come to fair and
objective findings.
In addition to preserving the integrity of the process as a stand-
alone goal, the legal and financial re pe rcussions of a biased
investigation can also be serious. Courts have not gone so far as to
determine that an employer’s duty to provide a thorough and
reasonable investigation requires the same level of due process as in
formal adjudicative processes.
28
Nonetheless, courts have
increasingly required certain elements of proced ural fairness in
workplace investigations.
29
For instance, they have criticized
investigations that do not adhere to basic due process principles,
such as providing notice of the allegations to the respondent,
granting an opportunity to be heard, and interviewing key
witnesses.
30
Likewise, courts and other adjudicative bodies have
28
See THOMILSON ET AL., supra note 16, at 7; see also Cotran v. Rollins
Hudig Hall Int’l Inc., 948 P.2d 412, 424 (Cal. 1998) (Mosk, J., concurring).
29
See, e.g., Tse v. Trow Consulting Eng’g Ltd., [1995] O.J. No. 2529 (Can.
Ont. C.J.) (QL); Nazir v. United Airlines, 100 Cal. Rptr. 3d 296 (Cal. Ct. App.
2009); Dewson v Boom Logistics Ltd. [2012] FWA 9027 (Austl.); Pinsker v.
Pacific Coast Soc’y of Orthodontists, 526 P.2d 253, 263 (Cal. 1974); Cotran, 948
P.2d at 424. For an overview of due process rights in investigations set out in
various labor arbitration decisions, see RICHARD L. KASPARI & KATHRYN M.
ENGDAHL, ANALYSIS OF INVESTIGATION, EVIDENCE, COMMUNICATION, AND
IMPLEMENTATION (Nov. 8, 2007),
http://apps.americanbar.org/labor/annualconference/2007/materials/data/papers/
v1/025.pdf.
30
See, e.g., Tse, [1995] O.J. No. 2529 (finding no just cause where an
employee was summarily dismissed for sexual harassment without any particulars
or explanation of the alleged harassment and stating that “[a]n obligation can be
asserted to exist on the part of the employer to give the accused employee a chance
to state his case”); Quirola v. Xerox Canada Inc. [1996] O.J. No. 21 (Can. Ont.
Gen. Div.) (QL) (“The case is instructive for a rather comprehensive discussion
of the importance of the need for a fair process with an opportunity for the
employee to respond. It is to be noted in this case that although the defendant had
elaborate policy guidelines in most areas, there was none with respect to sexual
harassment.”); Foerderer v. Nova Scotia Chemicals, [2007] A.J. No. 745 (Can.
420 JOURNAL OF LAW AND POLICY
frequently criticized workplace investigations for being conducted
in a “bia se d” manner.
31
Courts have found that fairness requires that
there be neither actual bias nor a reasonable apprehension of bias.
32
In articulating a test for reasonable apprehension of bias, one court
asked whether “a reasonably informed bystander looking at the
process [could] reasonably perceive bias on the part of one or more
of the persons involved in the assessment of the complainant.”
33
While it is clear that investigator neutrality is critical to a fair
process, it is not always clear what it means to be “unbiased.”
Although often used interchangeably to describe the roles of
Alta. Q.B.) (u pholding termination after investigation of sexual harassment
allegations and rejecting the employee’s assertion of entitlement to due process
and fair investigation). Th e Foerderer court cited with approval “Just Cause: The
Law of Summary Dismissal in Canada” as follows:
It may be that while a duty to provide a fair hearing may not
arise in the common law employment context the very nature
of the allegations of sexual harassment, based fo r the most part
on the accusations of one employee against another, compels
the employer to exercise a degree of care in prosecuting the
allegations, so as to balance the “rights” of the accused
employee with those of the victim. While there are divergent
views on this point, it does appear, that at the very least, by
providing to the accused employee an opportunity to respond to
allegations of harassment the employer enhances its case.
Foerderer, [2007] A.J. No . 745 (citing R
ANDALL SCOTT ECHLIN & MATTHEW
L.O. CERTOSIMO, JUST C AUSE: THE LAW OF SUMMARY DISMISSAL IN CANADA
(2010)).
31
See, e.g., Francis v. Can. Imperial Bank of Commerce, [1992] O.J. No.
153 (Can. Ont. Gen. Div.), rev’d, [1994] O.J. No. 2657 (Can. Ont. C.A.); Robert
Finlay, Investigating Employee Misconduct in the Workplace, 57 A
DVOC.
V
ANCOUVER 881, 883 1999; Elgert v. Home Hardware Stores Ltd., (2011) 510
A.R. 1 (Can. Alta. C.A.); Downham v. Cty. of Lennox & Addington, 2005 Ca nLII
45197 (Can. Ont. S.C.); Ribeiro v. Can. Imperial Bank of Co mmerce, 1989
CanLII 4281 (Can. Ont. S.C.); Tong v. Home Depot of Can. Inc., 2004 CanLII
18228 (Can. Ont. S.C.) ; Reeves v. Safeway Sto res, 16 Cal. Rptr. 3d 717 (Cal. Ct.
App. 6th 2004); Nazir v. United Airlines, 100 Cal. Rptr. 3d 296 (Cal. Ct. App. 1st
2009); Quela v. Payco-General Am. Cred its, Inc., 84 F. Supp. 2d 956 (N.D. Ill.
2000).
32
Haarsma v. the Deputy Minister of Nat’l Defence, 2013 PSST 5, para. 28
(Can.).
33
Id.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 421
workplace investigators,
34
mediators, and judges, the terms
“unbiased,” “neutral,” and “impartial” resist easy definition and,
while they are similar concepts, they have subtly different meanings.
Professional standards for mediators,
35
judges,
36
and arbitrators,
37
also require impartiality or neutrality in conducting their work and
attempt to define these terms. Workplace investigators have no
professional standards or model rules specific to their profession,
though cases indicate that the meaning of “unbiased” in the context
of workplace investigations includes the absence of having
34
Many of the definitions of “unbiased,” “neutral,” and “impartial” include
reference to one of the other terms. See Izumi, supra note 24, at 78 (discussing a
study which “found that that fourteen out of fifteen mediators defined neutrality
by using the word ‘impartiality’”).
35
One definition of neutrality that various U.S. states use in respect to a
mediator’s ethical obligation is, “freedom from favoritism or bias either by word
or action, and a commitment to serve all parties as opposed to a single party.” Id.
at 84; see also MODEL STANDARDS OF CONDUCT FOR MEDIATORS (AM. ARB.
ASS’N 2005),
https://www.adr.org/aaa/ShowProperty?nodeId=%2FUCM%2FADRSTG_0104
09&revision=latestreleased (requiring mediators to “decline a mediation if the
mediator cannot conduct it in an impartial manner,” defining impartiality as
“freedom from favoritism, bias, or prejudice,” and finding that, “[a] mediator
should not act with partiality or prejudice based on any participant’s personal
characteristics, background, values and beliefs, or performance at a mediation, or
any other reason”).
36
MODEL JUDICIAL CODE OF CONDUCT r. 2.3 cmt. (AM. BAR ASS’N 2011)
(“A judge who manifests bias or prejudice in a proceeding impairs the fairness of
the proceeding and brings the judiciary into disrepute.”). The Model Judicial Code
of Conduct states that manifestations of bias or prejudice may include overt acts,
such as “epithets; slurs; demeaning nicknames; negative stereotyping; attempted
humor based upon stereotypes; threatening, intimidating, or hostile acts;
suggestions of connections between race, ethnicity, or nationality and crime; and
irrelevant references to personal characteristics.” Id. The code also discusses
subtle indicators of bias, noting that “facial expressions and body language can
convey to parties and lawyers in the proceeding, jurors, the media, and others an
appearance of bias or prejudice.” Id.
37
The Code of Ethics for Arbitrators in Commercial Disputes, American
Arbitration Association is also a potentially useful reference point in the
development of standards for ethics for other neutrals, including investigators. See
CODE OF ETHICS FOR ARBITRATORS IN COMMERCIAL DISPUTES (AM. ARB. ASS’N
2004). Despite procedural differences between the processes, investigators and
arbitrators face many of the same issues of neutrality, propriety, confidence, and
fairness concerns well accounted for in the AAA Code.
422 JOURNAL OF LAW AND POLICY
commenced the investigation with a predetermined belief in the
outcome.
38
Ultimately, these terms seem to mean that workplace
investigators and other neutral professionals must be free from bias,
meaning freedom from actual or perceived prejudice or partiality.
39
These definitions, while useful in providing an expectation that
workplace investigators and other neutral professionals will conduct
their work in a manner that is free from the influence of biases and
prejudices, are incomplete. They do not address the existence of
unconscious biases, which may impact the work of neutral
professionals, including workplace investigators. Confronting the
issue of unconscious bias would increase such professionals’
awareness and understanding of the full range of possible biases that
may affect their neutrality.
C. The Unregulated Nature of Workplace Investigations
The practice of workplace investigations remains largely
unregulated.
40
Specific statutes and ethical rules govern individual
investigators who belong to other professions (such as lawyers and
private investigators), but no industry-wide standards, requisite
training, or governing rules of professional conduct apply uniformly
to workplace investigators. Case law does provide some guidance
regarding broad principles that workplace investigators must
follow.
41
However, the cases are fact specific and do not provide the
kind of guidance or consequences to professionals that regulation
may create.
38
See, e.g., Tong v. Home Depot of Can. Inc., 2004 CanLII 18228, para. 11
(Can. Ont. S.C.) (explaining that an investigation will be tainted where it is
determined that the in-house investigator commenced an investigation “with the
firm belief that the [accused employee] was a wrongdoer”).
39
See, e.g., Izumi, supra note 24, at 78–79; MODEL STANDARDS OF
CONDUCT FOR MEDIATORS, supra note 35.
40
Organizations such as the Association for Workplace Investigators, while
voluntary, are very useful for providing training, information and awareness of
issues specific to workplace investigators. See GUIDING PRINCIPLES, supra note
4.
41
See, e.g., Elgert v. Home Hardware Stores Limited, (2011) 510 A.R. 1,
para. 86–89 (Can. Alta. C.A.).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 423
Although hi ghly regulated within their profession, lawyers
acting as workplace investigators receive little guidance from their
Model Rule s of Professional Conduct regarding this unique role.
42
Such ethical rules are relevant to a traditionally adversarial system,
rather than the lawyer’s role as a neutral.
43
In fact, many Rules of
Professional Conduct only reference the role of a neutral in terms of
lawyers’ ethical duties when acting as arbitrators or mediators.
44
Such rules focus on the lawyer’s duty to ensure parties understand
the la wye r’s role in the process—that he or she is not an advocate
for either party.
45
Ethical issues that may face a third-party neutral
beyond this are not considered.
46
Private investigators also act as workplace investigators. Most
states require private investigators to be licensed. The New York
General Business Law governs the licensing of such professionals
and broadly defines the business of a “private investigator” to
include being retained to investigate matters relating to employee
conduct.
47
In other words, a licensed private investigator may
42
See, e.g., N.Y. COMP. CODES R. & REGS. tit. 22, § 1200 (2015).
43
See Alison Smiley, Professional Codes and Neutral Lawyering: An
Emerging Standard Governing Nonrepresentational Attorney Mediation, 7 GEO.
J. LEGAL ETHICS 213, 222 (1993) (noting several rules in conflict with the neutral
lawyering concept, including “lawyer-client loyalty by enjoining lawyers from
exercising independent judgment on their clients’ behalf,” and “mandate[ing] a
lawyer’s zealous representation of his client”).
44
See, e.g., MODEL RULES OF PROFESSIONAL CONDUCT r. 2.4 (AM. BAR
ASS’N 2013).
45
See id. The ABA adopted Model Rule 2.4 in 2002, in response to the lack
of guidance provided to lawyers who serve as neutrals. See Bruce E. Meyerson,
New Ethics Rules for Lawyers Who Serve as Mediators and Arbitrators, 1999
PROF. LAW SYMP. ISSUES 113, 114 (“The proposed Model Rule begins by placing
an ethical duty upon the lawyer who serves as a third-party neutral to conduct the
proceedings in an impartial, unbiased way. The lawyer should withdraw if he is
not able to conduct the proceedings impartially.”). Model Rule 2.4 requires a
lawyer to disclose any reason the lawyer might not be perceived as impartial. Id.
at 115.
46
See Meyerson, supra note 45, at 114.
47
N.Y. GEN. BUS. LAW § 71 (McKinney 2015) (“‘Private investigator’ shall
mean and include the business of private investigator and shall also mean and
include, separately or collectively, the making for hire, reward or for any
consideration whatsoever, of any investigation, or investigations for the purpose
of obtaining information with reference to any of the following matters,
424 JOURNAL OF LAW AND POLICY
conduct workplace investigations and will be regulated by that
industry. This type of legislation does not generally apply to lawyers
acting “in the regular practice of their profession.”
48
Some statutes
do mandate that private investigators obtain training.
49
While some
of that mandatory training includes references to education about
bias in the decision-making and report-writing processes, there is no
reference to the need for training on unconscious biases.
50
Thus, many workplace investigators may receive little to no
training or continuing education specific to the issues they regularly
face, including the challenge of controlling and reducing bias in the
investigation process. Even those who do seek out such education
may only be exposed to the more well known types of bias—actual
or perceived, rather than unconscious.
II. THE SCIENCE OF UNCONSCIOUS BIAS
The science of unconscious bias, otherwise referred to as
implicit bias or implicit cognition, studies an array of cognitive
processes that unconsciously impact human decision-making
processes and judgments.
51
In the past decade or so, researchers
have examined “the distinction between explicit and implicit mental
notwithstanding the fact that other functions and services may also be performed
for fee, hire or reward,” including matters relating to “the conduct, honesty,
efficiency, loyalty, or activities of employees, agents, contractors, and sub-
contractors”); see also Private Security and Investigative Services Act, S.O. 2005,
c. 34 (Can.) (defining “private investigator” broadly as “a person who performs
work, for remuneration, that consists primarily of conducting investigations in
order to provide information,” such as character or actions of a person, the
business or occupation of a person, or the whereabouts of persons or property).
48
See, e.g., N.Y. GEN. BUS. LAW § 83 (McKinney 2015). This legislation
would arguably apply to a licensed lawyer who is not engaged in the practice of
law.
49
In Ontario, private investigators must complete a Ministry approved
Training Curriculum for Private Investigators and a licensing test. See Private
Security & Investigative Services, ONT. MINISTRY OF COMMUNITY SAFETY &
CORRECTIONAL SERVICES,
http://www.mcscs.jus.gov.on.ca/english/PSIS/Training/AboutBasicTraining/PSI
S_training.html (last modified Mar. 24, 2016).
50
See id.
51
See Justin D. Levinson, Forgotten Racial Equality: Implicit Bias,
Decisionmaking, and Misremembering, 57 DUKE L.J. 345, 348 (2007).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 425
processes in relation to racial and other forms of bias.”
52
The
research has demonstrated the existence of unconscious biases,
which are mental processes that operate outside of humans’
consciousness, intentional awareness, or control, even in the face of
introspection.
53
This science has also raised serious doubts about
whether human behavior is under conscious control.
54
This science is important for professionals, like workplace
investigators, for whom impartiality comprises a critical component
of their work. The average person assumes that workplace
investigators can “cognitively process, evaluate and weigh the facts
presented” in a neutral manner, but research challenges the accuracy
of this assumption and shows that cognitive biases affect the way all
people process information.
55
For example, people often rely upon
critical thinking errors and false stereotypes “to justify their initial,
unreliable assessment.”
56
Courts have increasingly expressed
concern about actual or perceived biases in workplace
investigations.
57
In addition, courts and academics have recognized
the impact unconscious biases have on mediators, juries, and judges,
as well as on the effectiveness of laws.
58
While there are many kinds of unconscious biases, there are
arguably two broad types that may have the most significant impa ct
upon wor kp la ce investigators’ efforts at neutral fact-finding—
52
Deana A. Pollard, Unconscious Bias and Self-Critical Analysis: The Case
for a Qualified Evidentiary Equal Employment Opportunity Privilege, 74 WASH.
L. REV. 913, 917 (1999) (emphasis added).
53
See Debra L. Bassett, Deconstruct and Superstruct: Examining Bias
Across the Legal System, 46 U.C.D. L. REV. 1563, 1564 & n. 3 (2013) (“The
literature has used the synonyms ‘implicit,’ ‘automatic,’ and ‘unconscious’ in
describing such bias.”); see also Hon. John F. Irwin & Daniel L. Reel,
Unconscious Influences on Judicial Decision-Making: The Illusion of Objectivity,
42 MCGEORGE L. REV. 2, 3 (2011); Izumi, supra note 24, at 88; Greenwald &
Krieger, supra note 6, at 946.
54
Greenwald & Krieger, supra note 6, at 945.
55
Levinson, supra note 51, at 347–48.
56
Stephen Porter et al., Dangerous Decisions: The Impact of First
Impressions of Trustworthiness on the Evaluation of Legal Evidence and
Defendant Culpability, PSYCH., CRIME & L. 2010, at 1, 4.
57
Janet Bond Arterton, Unconscious Bias and the Impartial Jury, 40 CONN.
L. REV. 1023, 1028 (2008).
58
Id. at 1026–29.
426 JOURNAL OF LAW AND POLICY
unconscious biases towards social groups and those biases that lead
to “tunnel vision,” including confirmation bias, lie bias, and
trustworthiness bias.
A. Unconscious Bias towards Social Groups
Much of the most well-known research on unconscious bias has
focused on implicit attitudes towards members of socially
stigmatized groups.
59
The research often refers to such biases as
“implicit bias,” though the term is also used more broadly to
encompass other kinds of unconscious biases.
60
Until the late 1980s,
psychological research analyzed stereotyping and discriminating
through “observable behavior and self-reports;” in other words,
through explicit measures.
61
“[M]ost psychologists believed
that . . . attitudes, including stereotypes and prejudices, operated
consciously . . . [and] that individuals were aware of their own biases
and prejudices.”
62
Beginning in the “1990s, psychologists
documented that attitudes have both ‘explicit’ [(or conscious)] and
‘implicit’ [(or unconscious)] indices.”
63
Unconscious biases towards
or against certain groups are based on social stereotypes or attitudes
that have led to an association between a group and a trait.
64
They
are “automatically triggered by the slightest interaction with a target
group member.”
65
Research shows that implicit bias is widespread
59
Implicit Bias, STAN. ENCYCLOPEDIA OF PHIL. (Feb. 26, 2015),
http://plato.stanford.edu/entries/implicit-bias/.
60
See Understanding Implicit Bias, KIRWAN INST.,
http://kirwaninstitute.osu.edu/research/understanding-implicit-bias/ (last visited
Mar. 8, 2016).
61
Antony Page, Batson’s Blind-Spot: Unconscious Stereotyping and the
Peremptory Challenge, 85 B.U. L. REV. 155, 180 (2005).
62
Bassett, supra note 53, at 1567; see also Greenwald & Krieger, supra note
6, at 948 (noting that social psychologists define an attitude as an evaluative
disposition, for example, the “tendency to like or dislike, or to act favorably or
unfavorably toward someone or something”).
63
See Bassett, supra note 53, at 1567–68.
64
See Izumi, supra note 24, at 88; see also Greenwald & Krieger, supra note
6, at 949–51 (“Social stereotype is a mental association between a social group or
category and a trait . . . the content of the ascribed trait, rather than its evaluative
valence, is central . . . . For attitudes, the opposite holds.”).
65
Pollard, supra note 52, at 920.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 427
among the general public, and influences behavior in many contexts,
including employ ment, medicine, law enforcement, and forensic
science.
66
Implicit attitude measures
67
reveal a far greater bias in favor of
advantaged groups than do explicit measures.
68
Manifestations of
implicit bias often conflict with a person’s explicit attitudes about a
given group, a concept referred to as dissociation.
69
“Dissociation
[is] commonly observed in attitudes toward[s] stigmatized
groups.”
70
Studies indicate that dissociation begins in middle
childhood (age ten or so) “as partic ipan ts’ explicit bias began to
dissipate.”
71
1. The Implicit Association Test and Other
Research
Much of the evidence surrounding unconscious bias comes from
the now well-known Implicit Association Test (“IAT”) developed
in the 1990s by Mahzarin Banaji and Anthony Greenwald.
72
The
66
Anna Roberts, (Re)forming the Jury: Detection and Disinfection of
Implicit Juror Bias, 44 CONN. L. REV. 827, 834 (2012).
67
Greenwald and Banaji define “implicit attitudes” as “introspectively
unidentified (or inaccurately identified) traces of past experience that mediate
favorable or unfavorable feeling, thought or action toward social objects.”
Anthony G. Greenwald & Mahzarin R. Banaji, Implicit Social Cognition:
Attitudes, Self-Esteem, and Stereotypes, 102 PSYCHOL. REV. 4, 8 (1995). An
example of an im plicit attitude is where a voter votes for a candidate knowing
nothing about him other than his name. Greenwald & Krieger, supra note 6, at
948. One thing influencing the voter is that he shares one or more initial letters
with the candidate’s name. Id. “In such a case, the vote can be understood, at least
in part, as an implicit expression of the voter’s self-favourable attitude.” Id.
68
See Izumi, supra note 24, at 93.
69
Greenwald & Krieger, supra note 6, at 953; see Levinson, supra note 51,
at 360; Patricia G. Devine, Implicit Prejudice and Stereotyping: How Automatic
Are They? Introduction to the Special Section, 81 J. PERSONALITY & SOC.
PSYCHOL. 757, 757 (2001).
70
Greenwald & Krieger, supra note 6, at 949.
71
Izumi, supra note 24, at 94–95.
72
Irwin & Reel, supra note 53, at 3; see Christine Jolls & Cass R. Sunstein,
The Law of Implicit Bias, 94 CAL. L. REV. 969, 971 (2006); Levinson, supra note
51, at 354–55. Note that there are other types of testing of unconscious bias, such
as looking at physiological measures like “respiratory activity,
428 JOURNAL OF LAW AND POLICY
IAT has “provided powerful insight into and documentation of the
prevalence of our collective unconscious prejudice.”
73
Results of the
IAT have been repeatedly validated and consistently reflect that
most people harbor unconscious biases in areas such as, race,
gender, age, and disability.
74
The IAT is a computer-generated test that asks par ticipants to
categorize information as quickly as possible and then measures a
participant’s reaction.
75
The IAT “pairs an attitude object (such as a
racial group) with an evaluative dimension (i.e., good or bad) and
tests how response accuracy and speed indicate implicit and
automatic attitudes and stereotypes.”
76
The response time is
significantly faster when associations are more easily made between
the target group and the attribute than when associations are more
difficult to determine.
77
One commentator explained this as follows:
It takes slightly longer (in terms of milliseconds) for
an unconsciously biased individual to associate
positive descriptive words with a member of a
disfavored group than for the individual to associate
positive descriptive words with other persons. This
electromyographical activity, and eyeblink startle reflex.” Page, supra note 61, at
184.
73
Arterton, supra note 57, at 1031.
74
See Basset, supra note 53, at 1567–70; see also Greenwald & Krieger,
supra note 6, at 952. Some commentators have criticized the IAT, arguing that
the data do not support the conclusion that implicit bias leads to discriminatory
behaviour. See, e.g., Izumi, supra note 24, at 85.
A related and somewhat surprising finding was that members
of bias-affected groups sometimes harbored implicit biases
against their own group. For example, Banaji and Greenwald
found that women and men both display implicit gender
stereotypes that are sometimes negative toward women.
Similarly, Brian Nosek and his colleagues reported that older
people (those over sixty) and younger people (those in their
twenties) show a similar implicit preference for young over old.
Levinson, supra note 51, at 361.
75
Levinson, supra note 51, at 355–56.
76
Id.
77
Pollard, supra note 52, at 960.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 429
discrepancy in time is due to the differences in
cognitive barriers in making the connections.
78
This general method can be used to measure a wide variety of
associations that underlie attitudes and stereotypes.
79
One of the more well known IATs is the black/white IAT, in
which subjects respond more qui ck ly when Caucasian-Americans
are paired with pleasant words than when African-Americans are
paired with pleasant words (such as “good”).
80
The results indicate
an “implicit attitudinal preference for [European Americans]
relative to [African Americans].”
81
Other IAT iterations
demonstrate unconscious biases against many different groups,
including rampant implicit ageism.
82
The IAT and other research on implicit biases indicate that these
biases are widespread and often contradict explicit beliefs.
83
Results
of the IAT have shown that people’s implicit biases frequently
diverge from their self-reported attitudes —the phenomenon of
“dissociation” referred to above.
84
Many people sincerely express a
78
Id. at 918; see also Irwin & Reel, supra note 53, at 3. Note that some
authors have seriously criticized implicit bias research for failing to “satisfy key
scientific tests of validity” and have cautioned against making sweeping
statements that the science compels one or another proposed legal reform. See
Samuel Bagenstos, Implicit Bias, “Science” and Antidiscrimination Law, 1
H
ARV. L. & P
OL’Y REV. 477, 479 (2007) (quoting Gregory Mitchell & Philip E.
Tetlock, Antidiscrimination Law and the Perils of Mindreading, 67 O
HIO ST. L.J.
1023, 1023 (2006)). There is substantial dispute within psychology on some of
these critiques. See, e.g., Greenwald & Krieger, supra note 6, at 946.
79
Greenwald & Krieger, supra note 6, at 952.
80
See Roberts, supra note 66, at 848; see also Pollard, supra note 52, at 961.
81
Greenwald & Krieger, supra note 6, at 953.
82
Becca R. Levy & Mahzarin R. Banaji, Implicit Ageism, in AGEISM:
S
TEREOTYPING AND PREJUDICE AGAINST O LD ER PERSONS 49, 51–52 (Todd D.
Nelson ed., 2002); see also Izumi, supra note 24, at 94 (noting that implicit age
ism
has been shown to be far more negative than explicit attitudes and to crosscut all
age groups).
83
Pollard, supra note 52, at 917–18; Jolls & Sun stein, supra note 72, at 971.
“The impact of this type of bias on minorities can be as insidious as the older
traditional form of discrimination.” Pollard, supra note 52, at 927. For a
discussion as to why implicit biases are so prevalent, see Greenwald & Krieger,
supra note 6, at 959–60.
84
See Levy & Banaji, supra note 82, at 52.
430 JOURNAL OF LAW AND POLICY
commitment “to an anti-discrimination principle with respect to the
very trait against which they show a bias.”
85
Further, implicit biases predict behavior and the way people
make important decisions, often producing spontaneous behavior
that “diverges from a person’s avowed or endorsed beliefs or
principles.”
86
Implicit bias measures predict behaviour better than
explicit measures in situations that are “socially sensitive . . . where
impression-management processes might inhibit people from
expressing negative attitudes or unattractive stereotypes.”
87
Their
predictive validity is also higher with respect to “spontaneous
behaviors such as eye contact, seating distance, and
[similar] . . . actions that communicate social warmth or
discomfort.”
88
Many additional studies have been conducted that demonstrate
that unconscious biases can be used to predict behavior in a variety
of circumstances, including racial interactions,
89
physician
85
Jolls & Sunstein, supra note 72, at 975.
86
Greenwald & Krieger, supra note 6, at 951, 961.
87
Id. at 954–55.
88
Id. at 955; see also Nilanjana Dasgupta & Luis M. Rivera, From Automatic
Antigay Prejudice to Behavior: The Moderating Role of Conscious Beliefs About
Gender and Behavioral Control, 91 J. PERSONALITY & SOC. PSYCHOL. 268, 268
(2006).
89
See Greenwald & Krieger, supra note 6, at 961–62. One of the first
experiments to relate an IAT race attitude measure to discriminatory behavior
involved videotaping behavior of white undergraduates while being interviewed
separately by white and black experimenters. Id. at 961. The participants “also
completed a race attitude IAT measure.” Id. Those whose “IAT scores indicated
strong implicit preference for [w]hite relative to [b]lack . . . [showed] subtle and
spontaneous behaviors [that] suggested higher levels of comfort interacting
with . . . [w]hite experimenters,” including less hesitation, fewer speech errors,
and smiling more. Id.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 431
recommendations regarding health care,
90
hiring of minority
91
and
female candidates,
92
and judicial decision-making.
93
2. Explanations for Unconscious Biases About
Groups
Unconscious biases are “rooted in the fundamental mechanics of
the human thought process.”
94
Scientists have found that human
minds naturally stereotype as part of a “categorization” process,
useful in simplifying our complex environment for more efficient
cognitive functioning.
95
Categorization has been described as “the
most basic and essential of all cognitive processes”––without it,
humans would be “engulfed in a tidal wave of details” and their
“minds unable to cope.”
96
Humans form schemas in which they
group similar objects, events, ideas, and people, expecting these
associated items to coexist in different settings.
97
People or things
that do not fit easily in a specific category “may be forced into a
category regardless, thus reducing complexity, but introducing
inaccuracy.”
98
90
Izumi, supra note 24, at 95 (citing a study in which doctors with higher
measures of white preference recommended better treatment options to white
patients with the same medical profiles of black patients); see also Roberts, supra
note 66, at 850 (citing a study in which doctors and nurses were found to have
tested positive on the IAT for white preference).
91
See JUSTIN D. LEVINSON ET AL., IMPLICIT RACIAL BIAS ACROSS THE L AW
20 (Justin D. Levinson & Robert J. Smith eds., 2012) (discussing a study showing
that employers who harbored implicit biases towards Arabs were less likely to call
Arab candidates for job interviews).
92
See Greenwald & Krieger, supra note 6, at 949–51 for a discussion on
implicit stereotypes utilized in societal situations between males and females.
93
See Irwin & Reel, supra note 53, at 5; Izumi, supra note 24, at 98
(explaining how studies have shown that implicit biases can affect judges’
decision-making and may explain, at least in part, disparities in judicial decision-
making, such as in relation to convictions and sentencing).
94
Irwin & Reel, supra note 53, at 3.
95
Pollard, supra note 52, at 947; see Page, supra note 61, at 181 (“[The]
basic way in which people try to understand their world – categorization – can, of
its own accord, lead to stereotyping and discrimination.”).
96
Page, supra note 61, at 186.
97
See id. at 189–90.
98
Id. at 186.
432 JOURNAL OF LAW AND POLICY
While stereotypes have the benefit of permitting a “decision-
maker to make rapid inferences and reduce cognitive effort”
99
they
also lead to oversimplified conceptions and misapplied
knowledge.
100
“As a result, stereotyping ‘per se, propels the
individual down the road to bias.’”
101
These stereotypes are
“‘categories that have gone too far,’ using personal characteristics
such as race or gender as a proxy for personality traits such as
hostility, intelligence, or weakness.”
102
“[P]sychologists have found that stereotypes arise when a person
is as young as three years old and are usually learned from parents,
peers, and the media.”
103
As people grow older, their stereotypes
“harden” and, while they may develop nonbiased (or explicit) views
of the world, their stereotypes “become implicit and remain mostly
unchanged even as they develop nonprejudiced explicit views.”
104
The original beliefs “remain in the unconscious, waiting to be
activated”
105
and are easily and “automatically triggered by the
slightest interaction with a target group me mber.”
106
Once developed, stereotypes are not easil y changed. They are
“self-perpetuating” and maintained even in the face of
disconfirming data.
107
As they help us “fill in mi ssing information”
in our environment, “data supporting the stereotype is encoded
while data not supp ort i ng the stereotype is discarded.”
108
It is also
99
Id. at 236.
100
Id. at 18 8. For an overview of various theories and models provided by
psychologists to explain why humans form and utilize stereotypes, see Gary Blasi,
Advocacy Against the Stereotype: Lessons from Cognitive Social Psychology, 49
UCLA L. R
EV. 1241, 1254–66 (2001).
101
Page, supra note 61, at 188.
102
Pollard, supra note 52, at 919–20 (quoting Annie Murphy Paul, Where
Bias Begins: The Truth About Stereotypes, P
SYCHOL. T
ODAY, May–June 1998, at
52, 53).
103
Levinson, supra note 51, at 363.
104
Id.
105
Page, supra note 61, at 204.
106
Pollard, supra note 52, at 920; see Izumi, supra note 24, at 92.
107
Pollard, supra note 52, at 947. For a thorough review of the cognitive
processes by which people form and maintain stereotypes, see Page, supra note
61, at 199–202.
108
Pollard, supra note 52, at 947.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 433
difficult to “forget” stereotypes.
109
To do so requires “tremendous
mental resources” and may, in fact, “lead to a ‘rebound effect’ in
which the stereotype one tries to avoid is promoted.”
110
This is
because stereotypes are frequently and easily recalled, which
reinforces them in people’s minds.
111
Stereotypes also come from other psychological phenomena.
They relate to the “human instinct to identify with a group or
clan.”
112
Further, “the motivation to boost self-worth also affects the
activation and application of negative group stereotypes.”
113
Stereotypes are also often justified and propagated by way of self-
fulfilling prophecy. One commentator de scribed this process:
[W]e use schemas to anticipate what a stereotyped
group member will be like, how he or she will react,
and how we should behave. Our anticipatory
behavior toward the group member influences how
the group member responds; in essence, the
anticipatory behavior by the person operating under
the stereotype influences and causes the target group
member to engage in the “expected” behavior, thus
confirming the stereotype regardless of the
stereotype’s accuracy.
114
Unconscious biases towards groups of peopl e can have
particularly negative impacts on neutrality in workplace
investigations. Such biases may impact an investigator’s views of
the parties or witnesses to an investigation, despite explicitly held
views, and thus taint how they conduct their investigations, analyze
evidence, and assess credibility.
109
Id.
110
Id. at 948.
111
Id.
112
Id. at 917.
113
Blasi, supra note 100, at 1252. In cases where activating or applying a
negative stereotype would tend to diminish our self-concept we are less likely to
do so (e.g., praising a female teacher who gave us good marks). Id. at 1251.
114
Pollard, supra note 52, at 956.
434 JOURNAL OF LAW AND POLICY
B. Tunnel Vision: Confirmation Bias and Other Cognitive
Biases
“It is the peculiar and perpetual error of the human
understanding to be more moved and excited by affirmatives
than negatives, whereas it ought duly to be impartial; nay, in
establishing any true axiom, the negative instance is the most
powerful.”
115
In addition to unconscious biases about social groups, other
cognitive biases can impact how people make decisions.
Collectively, some of these cognit ive biases can work together to
contribute to what some refer to as “tunnel vision.”
116
One
commentator defined tunnel vision as “that ‘compendium of
common heuristics and logical fallacies,’ to which we are all
susceptible.” Tunnel vision leads actors to focus on a particular
conclusion and then to filter all evidence in a case through the lens
of that conclusion—significantly elevating supporting evidence and
overlooking or dismissing inconsistent evidence.
117
This is
generally not the result of “maliciousness or indifference” but rather
“the product of the human condition as well as institutional and
cultural pressures.”
118
People regularly rely upon a variety of
cognitive biases and “heuristics” (or mental shortcuts) to make
complex decisions, leading people to make decisions in “predictably
irrational ways.”
119
This section discusses some of the unconscious
biases that can lead to tunnel vision, including confirmation bias, lie
bias, and trustworthiness bias.
115
FRANCIS BACON, WORKS 348 (Basil Montagu ed., 3d vol., Philadelphia,
Carey & Hart 1841).
116
Keith A. Findley & Michael S. Scott, The Multiple Dimensions of Tunnel
Vision in Criminal Cases, 2006 WIS. L. REV. 291, 307–08 (2006).
117
Id. at 292.
118
Id.
119
Edie Greene & Brian H. Bornstein, Book Review: Nudging the Justice
System Toward Better Decisions, 103 J. CRIM. L. & CRIMINOLOGY 1155, 1161–
62 (2013).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 435
1. Confirmation Bias
“Confirmation bias is the tendency to [unconsciously] bolster a
hypothesis,” belief, or expectation by seeking and/or favoring
confirming information while mini mizing or ignoring disconfirming
information.
120
Once people form a hypothesis, “they search for
information to support it, interpret ambiguous information as
consistent with it, and minimize inconsistent evidence.”
121
Even in
situations where there is no prior relevant reason to confirm a
hypothesis, “people [nonetheless] tend to favo[r] confirmation as the
default testing strategy.”
122
The empirical research has produced
overwhelming data to support the existence of confirmation bias
across diverse judgment contexts, such as economics, criminal
investigations, medical decision-making, political beliefs, logical
problem solving, forensic analysis, and social interaction.
123
The
120
Barbara O’Brien, Prime Suspect: An Examination of Factors that
Aggravate and Counteract Confirmation Bias in Criminal Investigations, 15
PSYCHOL., PUB. POL’Y & L. 315, 315 (2009); see also Findley & Scott, supra note
116, at 309 (“[T]he tendency to seek or interpret evidence in ways that support
existing beliefs, expectations, or hypotheses.”); Martine B. Powell et al., Skills in
Interviewing Reduces Confirmation Bias, 9 J. INVESTIGATIVE PSYCHOL. &
OFFENDER PROFILING 126, 126 (2012); Eric Rassin, Blindness to Alternative
Scenarios in Evidence Evaluation, 7 J. INVESTIGATIVE PSYCHOL. & OFFENDER
PROFILING 153, 154 (2010).
121
O’Brien, supra note 120, at 328; see also Karl Ask et al., The ‘Elasticity’
of Criminal Evidence: A Moderator of Investigation Bias, 22 APPLIED COGNITIVE
PSYCHOL. 1245, 1246 (2008); Alafair Burke, Neutralizing Cognitive Bias: An
Invitation to Prosecutors, 2 N.Y.U. J.L. & LIBERTY 512, 516–17 (2007).
122
Eric Rassin et al., Let’s Find the Evidence: An Analogue Study of
Confirmation Bias in Criminal Investigations, J. INVESTIGATIVE PSYCHOL. &
OFFENDER PROFILING 231, 232 (2010).
123
Ask et al., supra note 121, at 1246; see Rassin et al., supra note 122, at
231; Galit Nahari & Gershon Ben-Shakhar, Primacy Effect in Credibility
Judgments: The Vulnerability of Verbal Cues to Biased Interpretations, 27
APPLIED CONGNITIVE PSYCHOL. 247, 247 (2013). The recent report produced by
Columbia University School of Journalism evaluating what went wrong with the
Rolling Stone article entitled “A Rape on Campus” cited confirmation bias as
having played a role in the journalistic failures of that article:
The problem of confirmation bias – the tendency of people to
be trapped by pre-existing assumptions and to select facts that
support their own views while overlooking contradictory ones
– is a well-established finding of social science. It seems to have
436 JOURNAL OF LAW AND POLICY
danger of confirmation bias for any professional who is required to
maintain neutrality is that they run the risk of becoming convinced
that a suspect or respondent is guilty or has behaved in a certain
manner, and may then no longer be open to alternative scenarios.
This may equally disadvantage a complainant where an investigator
is convinced that the respondent is not guilty. This is not through
“deliberate case building,” but, unconscious information
processing.
124
It involves “unwittingly selecting and interpreting
evidence to support a previously held belief.”
125
Simply forming a
hypothesis worsens bias.
126
Studies of confirmation bias have shown that most people
“sho[w] a preference for evidence that . . . confirms their hypothesis
over evidence that would disconfirm it, even [where] the latter [is]
more probative.”
127
Confirmation bias may also result when a
person obtains a description of another’s qualities before meeting
them.
128
Studies further show that in some circumstances, people do
not respond to information at variance with their beliefs by merely
ignoring it, but rather by critically examining it so as to undermine
it. “The end product of this intense scrutiny is that the contradictory
been a factor here. Erdely believed the university was
obstructing justice. She felt she had been blocked. Like many
other universities, UVA had a flawed record of managing
sexual assault cases. Jackie’s experience seemed to confirm this
larger pattern. Her story seemed well established on campus,
repeated and accepted.
Sheila Coronel, et al., Rolling Stone and UVA: A Columbia University Graduate
School of Journalism Report, ROLLING STONE (Apr. 5, 2015),
http://www.rollingstone.com/culture/features/a-rape-on-campus-what-went-
wrong-20150405?page=3.
124
O’Brien, supra note 120, at 316.
125
Id.; see also Page, supra note 61, at 215–16 (“Once a decision-maker has
a tentative hypothesis about a person, she must process information in order to
support or reject the hypothesis. Keeping in mind that our senses receive more
than eleven million pieces of information every second, and that the highest
number of pieces we can consciously process is roughly forty, much of our
information processing can and must occur unconsciously. Our brain filters what
we will consciously address, like hearing our name across the room at a cocktail
party.”).
126
See O’Brien, supra note 120, at 328.
127
Findley & Scott, supra note 116, at 310.
128
See id. at 311.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 437
information is either considered too flawed to be relevant, or is
redefined into a less damaging category.”
129
Another phenomenon related to confirmation bias, expectancy
effects, occurs when a prediction or prophecy is made that causes
people to alter their behaviors such that “a statement that would have
been false becomes true as a consequence of the prediction or
prophecy being made.”
130
Experiments have been conducted in
which investigative interviewers are given pre-interview reasons to
believe or to doubt the interviewee.
131
This “assignment of
expectation” impacted the investigators’ interviewing behavior,
including how they structured the interview and the questions they
posed.
132
Expectation also lays the groundwork for selective
attention to evidence. Humans are often confronted with far too
much information to process and will tend to draw selectively from
the available information, focusing on items that confirm an
expectation about a given event.
133
Many studies conducted on confirmation bias in the context of
criminal investigations have led researchers to conclude that an
“investigator bias” exists.
134
Investigator bias leads investigators to
“select and filter the evidence that will build a case for conviction,
while ignoring or suppressing evidence that points away from
guilt.”
135
The early naming of a suspect itself increases confirmation
129
Id. at 313.
130
Victoria Springer, Expectancy Effects in Forensic Evidence Handling:
Social Psychological Perspectives, 7 JIJIS 311, 311 (2007). Demonstrating the
expectancy effect, “a pathologist who is told she is being presented with a slide
of abnormal cells is more likely to conclude that she is seeing abnormal cells than
one who is told she is being presented with a slide of normal cells.” See D. Michael
Risinger et al., The Daubert/Kumho Implications of Observer Effects in Forensic
Science: Hidden Problems of Expectation and Suggestion, 90 CAL. L. REV. 1, 13
(2002).
131
Risinger, supra note 130, at 15.
132
Id.
133
Id. An example of this “expectancy effect” is found in one study that
showed examination scorers gave different grades to identical performances and
that those differences correlated with the exam graders’ expectations created by
the child’s IQ score. Id. at 23.
134
See Findley & Scott, supra note 116, at 295.
135
Id. at 292.
438 JOURNAL OF LAW AND POLICY
bias throughout criminal investigations.
136
Further, criminal
investigators tend to be more skeptical about evidence that counters
their original perception than evidence that is consistent with it.
137
Even temporary acceptance of the truth of a hypothesis may have
consequences for the entirety of the investigation, including how an
investigator “perceives the problem, interprets relevant data, and
searches for new information.”
138
This shift towards proving guilt in investigations may not always
impact the outcome. However, when the evidence is ambiguous,
subtle influences on how an investigator interprets that evidence
could influence the outcome.
139
As one commentator stated, “[a]
weak case can start to look strong when the investigator overlooks
potentially fruitful leads in another direction.”
140
Investigators may
136
See O’Brien, supra note 120, at 324, 328 (describing a study in which
participants stated early on who they believed committed the crime, compared to
participants not asked to name a suspect, remembered the facts as more consistent
with the named suspect’s guilt, “advocated more lines of investigation focused on
him, required more reports focused on him, and subtly shifted their opinions about
matters relevant to determining guilt in a way that support[ed] their suspicions”).
In one study, participants prepped with information that a videotaped alibi
provider was guilty recalled less alibi-relevant information, found the alibi less
believable, and viewed the alibi provider more negatively than did participants
without such an expectation. See Elizabeth Olson, “You Don’t Expect Me to
Believe That, Do You?” Expectations Influence Recall and Belief of Alibi
Information, 43 J.
APPLIED SOC. PSYCHOL. 1238, 1238 (2013). In another similar
study, those primed with guilty expectations “pushed hard for confessions,
selected more guilt-presumptive questions, and interpreted suspects’ behaviour as
more consistent with guilt.” See Saul M. Kassin et al., Behavioral Confirmation
in the Interrogation Room: On the Dangers of Presuming Guilt, 27 L. & HUM.
BEHAV. 183, 187 (2003). Indeed, they exerted the most pressure on innocent
suspects who offered plausible denials, in an apparent redoubling effort to elicit a
confession. See id.
137
Karl Ask & Par Anders Granhag, Motivational Bias in Criminal
Investigators’ Judgments of Witness Reliability, 31 J. APPLIED SOC. PSYCHOL.
561, 565 (2007); see O’Brien, supra note 120, at 318.
138
O’Brien, supra note 120, at 317.
139
See Risinger, supra note 130, at 16.
140
See O’Brien, supra note 120, 328–29 (“Investigators need not be
especially sure that they have the right person to sway their investigation toward
an early suspect.”).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 439
then become convinced of guilt based on the evidence, leaving them
less open to alternative scenarios.
141
Workplace investigators are highly susceptible to confirmation
bias and expectancy effects, as they, like criminal investigators,
question interviewees, assess credibility, and interpret evidence.
Since workplace investigators are given a “suspect” or respondent
at the very start of the investigation, workplace investigators may
form a programmed hypothesis from which they may not escape.
This can affect not only how they interpret existing evidence, but
also what evidence they look for next, all of which can profoundly
undermine the truth-find ing process.
142
2. Lie Bias
“Lie bias” or “deception bias” is a bias towards believing a
person to be deceptive rather than truthful.
143
This bias impacts
many investigators and fact-finders
144
as they rely upon body
language to “detect inconsistencies, contradictions, and other signs
of deception.”
145
Further, “trained detectives and interrogators tend
to be more confident in their judgments than untrained
individuals.”
146
The Reid Approach is an interrogation technique in which many
police investigators are trained.
147
It advocates a two-step process: a
nonconfrontational, fact-finding interview to determine if suspect is
lying or telling the truth; then if the suspect is deemed guilty, a shift
141
See Rassin et al, supra note 122, at 240; see also O’Brien, supra note 120,
at 315.
142
See O’Brien, supra note 120, at 316.
143
See generally Findley & Scott, supra note 116, at 337 (explaining that
detectives with deception bias tend to presume guilt); see also Christian A.
Meissner & Saul M. Kassin, “He’s Guilty!”: Investigator Bias in Judgments of
Truth and Deception, 26 L. & HUM. BEHAV. 469, 478 (2002) (describing how
there is an “investigator bias effect” where confident investigators are more likely
to respond that a suspect was being “deceitful” rather than “truthful”).
144
See Findley & Scott, supra note 116, at 337.
145
Louis Avitabile & Brian H. Kleiner, How to Assess Credibility When
Conducting Workplace Investigations, 26 MGMT. RES. NEWS, 219, 222 (2003).
146
Findley & Scott, supra note 116, at 337; see Meissner & Kassin, supra
note 143, at 478.
147
Olson, supra note 136, at 1238.
440 JOURNAL OF LAW AND POLICY
is made to interrogation, which is a guilt-presumptive interaction.
148
The Reid Technique has been strongly criticized for leading
investigators to take guilty expectations into the interview stage,
causing them to “interpret, encode, and recall only information that
fits their expectation—information that indicates a suspect is guilty
or lying.”
149
Further, the Reid Technique has been criticized for its claim that
those trained in its techniques of observing and analyzing verbal and
nonverbal behavior
150
can judge truth and deception with high levels
of accuracy.
151
Studies indicate that “training and prior experience
increased the likelihood of judging a story as a lie,”
152
though
research has shown that those trained to detect lies are only very
slightly more able to do so than lay people.
153
Indeed, one study
indicated that students trained in the Reid Technique did
significantly worse in distinguishing between innocent true denials
and guilty false denials, though they were more confident in their
judgments and articulated more reasons for their judgments.
154
The
same study was repl icated with detectives and investigators, the
majority of whom had special training. The overall level of accuracy
even amongst these professionals was only 56 percent.
155
148
Id. Once interrogation starts, a nine-step interrogation process is used.
Findley & Scott, supra note 116, 334. It is “designed to break suspects down,
convince them that they are doomed, and then make a confession appear to be a
rational or risk-reducing choice.” Id.
149
Olson, supra note 136, at 1238.
150
Saul M. Kassin, Human Judges of Truth, Deception, and Credibility:
Confident but Erroneous, 23 CARDOZO L. REV. 809, 812 (2002). Signs of
deception include “avoidance of eye contact[,] . . . slouching, maintaining a rigid
posture, crossing arms or legs, and issuing qualified denials.” Id.
151
Id. (“John A. Reid & Associates claim they can train detectives to judge
truth and deception at 80 to 85 percent levels of accuracy.”).
152
Olson, supra note 136, at 1238–39.
153
See Jaume Masip et al., Training to Detect What? The Biasing Effects of
Training on Veracity Judgments, 23 APPLIED COGNITIVE PSYCHOL. 1282, 1282
(2009).
154
See Kassin, supra note 150, at 813–14.
155
Id. at 811, 814; see also Meissner & Kassin, supra note 143, at 470.
Several studies show that “professional lie detectors” (or those who detect
deception for a living) are not better at accurately detecting deception than lay
observers. Masip et al., supra note 153, at 1282. One study asked college students
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 441
Workplace investigators are susceptible to lie bias and that
susceptibility may be heightened for those who are trained in and
utilize the Reid Technique. As part of their work, workplace
investigators must regularly assess credibility of the parties and
witnesses to reach their conclusions. While a certain level of
skepticism may serve an investigator well in this regard, there is real
risk that he or she will be led astray by expectations or assumptions
about a person’s deceptiveness based on little more than initial
impressions. Lie bias, in turn, may lead to or increase confirmation
bias, as the investigator seeks additional evidence to confirm a belief
that the respondent is lying.
3. “Trustworthiness Bias” or Dangerous Decisions
Theory
The Dangerous Decisions Theory (“DDT”) “suggests that
instantaneous impressions of trustworthiness based on facial
appearance may play a major role in both assessing the credibility
of and ensuing decisions about the target.”
156
DDT suggests that
people make “potent judgments of trustworthiness” rapidly upon
seeing a defendant’s face.
157
Yet research shows that such
assessments based on the face alone “while instantaneous, are highly
fallible.”
158
Much like the effect of an early hypothesis on confirmation bias,
an initial evaluation of a witness’ trustworthiness is “thought to
and trained professionals to determine whether videotaped stories were true and
or false. Kassin, supra note 150, at 811. The accuracy rates for college students
were 52.8%, compared to a range of 55.8% to 57.6% for trained professionals
such as police detectives, CIA, FBI, military polygraph examiners, and trial
judges. Id. US Secret Service Agents performed the best at 64%. Id. Focusing on
nonverbal cues also leads to a stronger lie bias. See Aldert Vrij, Nonverbal
Dominance Versus Verbal Accuracy in Lie Detection, 35 CRIM. JUST. & BEHAV.
1323, 1323 (2008). For a defense of the Reid Technique, see Jonathan Goodman,
Getting to the Truth: Analysis and Argument in Support of the Reid Technique of
Interview and Interrogation, 21 MAINE B. J. 20 (2006).
156
Porter et al., supra note 56, at 2.
157
Stephen Porter & Leanne ten Brinke, Dangerous Decisions: A
Theoretical Framework for Understanding How Judges Assess Credibility in the
Courtroom, 14 LEGAL & CRIMINOLOGICAL PSYCHOL. J. 119, 119 (2009).
158
Porter et al., supra note 56, at 3.
442 JOURNAL OF LAW AND POLICY
influence subsequent inferences concerning the defendant (or other
witnesses) by making decision-making about him/her increasingly
irrational, leading to biased decisions.”
159
DDT, like confirmation
bias, is concerning in its potential to “result in a non-critical, ‘tunnel
vision’ assimilation of potentially ambiguous and even
contradictory evidence concerning the defendant.”
160
One study conducted on this phenomenon showed that less
evidence was required to convict an “untrustworthy looking”
defendant of the same crime as a “trustworthy looking” person,
particularly for a more severe crime.
161
Further, the “‘untrustworthy
defendants’ were [deemed] guilty based on an average of fewer than
five pieces of ‘ambiguous’ evidence, while more incriminating
evidence was necessary to convict a trustworthy defendant of the
same . . . crime.”
162
While the studies on DDT relate primarily to the context of the
courtroom, DDT may equally impact workplace investigators.
External workplace investigators (as opposed to in-house
investigators) may be particularly susceptible to this, as they
generally have no prior contact with or knowledge of the respondent
(or other witnesses) and may be more likely to make biased
decisions about him or her based upon appearances. That initial
judgment of trustworthiness can be critical in an investigation, as it
may lead a workplace investigator astray as he or she seeks evidence
that confirms that initial impression.
159
Id. at 4.
160
Id. at 4.
161
Id. at 7–8.
162
Id. at 10. An example of the dangerous decisions theory at work occurred
in a Florida case. Tom Sawyer was accused of sexual assault and murder,
interrogated for 16 hours, and a confession was extracted, likely to have been
false. Meissner & Kassin, supra note 143, at 470. Charges were ultimately
dropped, as the trial judge suppressed the confession. Id. at 470–71.
Most intriguing to us was the manner in which Sawyer became
a prime suspect—his face flushed and he appeared embarrassed
during an initial interview, a behavioral reaction interpreted as
a sign of deception. However, what the investigators did not
know was that Sawyer was a recovering alcoholic with a social
anxiety disorder that caused him to sweat profusely and blush
in evaluative social situations.
Id. at 471.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 443
C. Potential Impact of Unconscious Biases on Workplace
Investigators
The potential impact of the unconscious biases discussed above
on the true neut ra li ty of workplace investigators is significant.
163
Workplace investigators, like all people, are susceptible to these
automatic cognitive biases in accessing and interpreting data.
164
Even if workplace investigators hold explicit egalitarian values, are
free from bias or prejudice against either party, and know all the
relevant facts, they may still make “systematically erroneous
decisions” or have “cognitive accidents,” not because of ill
intentions, but simply as a result of automatic human thought
processes.
165
Workplace investigators often deal with ambiguous situations
where evidence can be interpreted and credibility assessed in
different ways. This is important because studies indicate that
discrimination is more likely to occur in normatively ambiguous
situations, even for those holding explicit egalitarian values, as
opposed to those involving “clear indications of right and wrong
behavior.”
166
This is due to the automatic nature of stereotypes and
a lack of awareness about the need to monitor our biases.
167
163
See id.
164
See Chris Guthrie et al., Inside the Judicial Mind, 86 CORNELL L. REV.
777, 829 (2001). The problem of tunnel vision is well-recognized in the criminal
justice system. Several official inquiries have concluded that tunnel vision has
played a significant role in specific wrongful convictions. See Findley & Scott,
supra note 116, at 294.
165
Guthrie et al., supra note 164, at 829; see also Burke, supra note 121, at
528 (discussing the “narrative trend that increasingly depicts prosecutors as
victims of cognitive accidents as opposed to purposeful or reckless wrongdoers”).
166
See generally Guthrie et al., supra note 164, at 829 (explaining that even
judges who are free from bias and prejudice make erroneous decisions because of
the way they think).
167
See id. at 829–30. Interestingly, one experiment demonstrated that
“[w]hite jurors are more likely to demonstrate racial prejudice in cases without
salient racial issues.” Blasi, supra note 100, at 1247 (emphasis added). The
experiment involved providing jurors with two similar scenarios, one of which
contained an additional sentence “suggest[ing] that the defendant had been the
subject of racial remarks and unfair criticism from teammates of the other racial
group earlier in the season.” Id. The mock jurors in the “race-salient
version . . . were as likely to convict the White as the Black defendant on the same
444 JOURNAL OF LAW AND POLICY
Cognitive biases may also overlap and exacerbate faulty
decision-making and judgments about evidence and witnesses. For
example, an unconscious bias about a racial group may cause the
investigator to develop an initial hypothesis based on a stereotype
and seek to confirm it. Preliminary hypotheses may bias the
subsequent evidence collection process and analysis, which can
detrimentally impact the lives of those involved in workplace
investigations. Confirmation bias is “exacerbated in situations
where people must make sense of information sequentially, adding
new information to an existing framework,” precisely what occurs
in a workplace investigation.
168
Ultimately, workplace investigators must understand how
cognitive biases can automatically and unintentionally lead to
poorly supported findings and possible damage to the subject of the
investigation. Court decisions analyzing biased investigations help
to illustrate how bias may undermine the investigation process and
results.
III. CASES OF BIASED INVESTIGATIONS
In recent years, there has been increased judicial scrutiny of
employers’ workplace investigation processes. Courts expect
investigations to be conducted in a fair manner. “[J]udges seem to
be recognizing the power imbalance inherent in the employer-
employee relationships and attempting to tip the scales back to a
(perceived) equilibrium.”
169
Neutrality, or lack of bias, on the part
of the investigator is one of the core components that judges
consider in evaluating the fairne ss of an investigation.
170
While courts do not appear to have specifically referred to
unconscious bias to explain or criticize a workplace investigation
result, the cases addressed in this section provide examples of cases
in which confirmation bias seems to have played a significant role.
Many of the investigators in these cases sought evidence confirming
facts.” Id. In the scenario without the sentence, “mock jurors convicted the Black
defendant 90 percent of the time and White defendant only 70 percent of the
time.” Id.
168
See Olson, supra note 136, at 1239.
169
Finlay, supra note 31, at 882.
170
See Avitabile & Kleiner, supra note 145, at 220.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 445
their in itial beliefs at the expense of a thorough and fair
investigation.
It is not clear whether the bias at play in all of the cases cited
below was entirely “unconscious,” as at least some of the facts tend
to indicate some level of intentionality. Also, the cases do not
provide obvious examples of the arguably more difficult types of
cognitive biases to identify based on evidenc e about the
investigative process, such as biases towards groups or DDT. Even
so, these cases demonstrate the damage that bias in any form can do
to the integrity of a workplace investigation and ought to serve as a
cautionary tale to all investigators, who are all susceptible to
conducting similarly unfair investigations as a result of unconscious
biases.
A. Investigations in Which Confirmation Bias May Have
Played a Role
Several recent Canadian cases demonstrate confirmation or
investigator bias seeping into conclusions reached in workplace
investigations.
171
In one such case, Ribeiro v. CIBC, the Ontario
Superior Court of Justice found that the plaintiff’s manager, who
oversaw an investigation into allegations of fraud, had
predetermined conclusions at the beginning of the investigation and
“was bound and determined to justify” a finding against the
plaintiff.
172
The Court went on to criticize the company’s fraud
investigator, a for mer police officer, stating that he “started with a
desired result and, thereafter, proceeded so as to support it.”
173
The
Court further stated that the investigator was unable to remain
171
See Findley & Scott, supra note 116, at 294.
172
Ribeiro v. Can. Imperial Bank of Commerce, 1989 CanLII 4281 (Can.
Ont. S.C.). Mr. Ribeiro turned down a friend, Mr. LeBlanc, who wanted a $1,000
bank loan because he did not qualify. Id. at 39. Mr. LeBlanc subsequently returned
with his girlfriend who applied for the loan. Id. at 9. The loan was granted and the
girlfriend then lent the money to Mr. LeBlanc. Id. CIBC took the position that Mr.
Ribeiro had engaged in fraud by lending money indirectly to a known
unacceptable credit risk. Id. at 11. The court determined that there was no cause
for dismissal and determined that the manager’s “attitude and demeanour was
anything but that of an objective, open, and fair-minded executive that I expect
would represent a Canadian chartered bank.” Id. at 44.
173
Id.
446 JOURNAL OF LAW AND POLICY
objective “once pointed in the direction of a suggested or alleged
wrongdoing” by the respondent/plaintiff.
174
The investigator went
so far as to pursue information unrelated to the alleged fraud, such
as possible former criminal activity, in an effort to confirm the
plaintiff’s guilt.
175
Elgert v. Home Hardware Stores Limited
176
represents an
extreme case of tunnel vision in which the employer and
investigators carried out an investigation into allegations made
against the plaintiff in a biased manner.
177
Elgert, a seventeen-year
employee, was dismissed from his job as supervisor following an
investigation into allegations of sexual assault made by two female
employees.
178
After examining the investigation process, the
Alberta Court of Appeal concluded that the investigation was
carried out in a biased and bad faith manner.
179
In finding that the
investigation lacked the requisite neutrality and thoroughness, the
court no te d the following flaws: the investigators’ close relations
with management, the failure to interview relevant witnesses, the
failure to consider possible motive or fabrication of the allegations,
and refusal to provide particulars to the respondent in advance of his
interview.
180
While ulterior motives may have played a role in the
outcome of this investigation, it provides an example of the danger
that commencing an investigation with a hypothesis can lead to a
faulty and unfair conclusion.
In another case, Tong v. The Home Depot of Canada Inc., the
Ontario Superior Court of Justice conclude d that the workplace
investigator was “consumed” with building a case against the
plaintiff and was successful in doing so.
181
The court stated that an
174
Id.
175
Id.
176
Elgert v. Home Hardware Stores Ltd. (2011), 510 A.R. 1 (Can Alta.
C.A.).
177
Id.
178
Id. Mr. Elgert brought a lawsuit against Home Hardware for wrongful
dismissal and defamation. Id. The jury found that Mr. Elgert had not, in fact,
committed the alleged sexual assault. Id.
179
Id.
180
Id.
181
Tong v. The Home Depot of Canada, 2004 CanLII 18228, para. 16 (Can.
Ont. S.C.). Mr. Tong, a sales associate, had worked at Home Depot for
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 447
investigation will be tainted where the in-house investigator
commenced the investigation “with the firm belief that the [accused
employee] was a wrongdoer.”
182
Similarly, in Ogden v. Canadian
Imperial Bank of Commerce, the British Columbia Supreme Court
found that overall the defendant was not interested in using its
investigation as a fact-finding mission but rather was “engaged in
an exercise in case building against [the respondent/plaintiff].”
183
Confirmation bias may also play a role in cases where another
person in the organization who may be biased against one of the
parties influences the investigator. In Reeves v. Safeway Stores, the
California Court of Appeal found that “the evidence raised triable
issues as to the existence and effect of retaliatory motive on the part
of the supervisor, and as to whether the manager and the
approximately five years when a new floor manager, Mr. MacDonald, was hired.
Id. at para. 2–3. Mr. MacDonald concluded that Mr. Tong was taking longer
breaks than permitted, a view he shared with the store manager who encouraged
him to keep a watchful eye on Mr. Tong. Id. at para. 4–6. Mr. MacDonald kept
Mr. Tong under almost constant surveillance during work hours. Id. at para. 7.
Mr. Tong was given no opportunity to respond to Mr. MacDonald’s concerns or
the chance to offer any explanation. Id. Mr. MacDonald did not interview a single
employee about Mr. Tong and chose not even to speak to Mr. Tong’s supervisor.
Id. at para. 14.
182
Id. at para. 11–12.
183
Ogden v. Canadian Imperial Bank of Commerce, 2014 CanLII 285 (Can.
B.C. S.C.). Ms. Ogden, a financial advisor at CIBC who immigrated to Canada in
2000, was terminated for cause following an investigation into a transfer of funds
from China to her personal account on behalf of a client. Id. at para. 6, 15–16. At
the time of her dismissal, she had a portfolio of $233 million working with clients
in the Chinese community. Id. at para. 14. Ms. Ogden transferred funds to her
personal account in order to save a house purchase for the client. Id. at para. 16.
She did not benefit personally benefit from the transaction, but her conduct was
contrary to CIBC practice. Id. at para. 155. The personal wire transfer was
discovered six months later. Id. at para. 196. CIBC appointed an investigator,
following which Ms. Ogden was terminated for cause. Id. at para. 212–13. Ms.
Ogden challenged the dismissal. Id. at para. 214. The Court found that CIBC did
not responsibly investigate the allegations and could therefore not substantiate
that cause for dismissal existed. Id. at para. 390–96. Moreover, the timing of the
investigation was a deliberate attempt to deprive Ms. Ogden of a commission
payment. Id. at para. 389. The British Columbia Supreme Court was highly
critical of the investigation for many reasons, including the investigator’s purpose
to procure admissions and prevent Ms. Ogden from fully answering questions. Id.
at para. 413.
448 JOURNAL OF LAW AND POLICY
intermediate investigator acted as tools or ‘cat’s paws’ for the
supervisor.”
184
The court stated that the evidence “supports an
inference that [the] investigation of the alleged misconduct was not
truly independent, but was heavily skewed to favor the ostensibly
tentative conclusions of the reporting supervisor.”
185
This was based
on language indicating that the investigator may have been
predisposed towards confirming “disciplinary charges rather than
objectively ascertain[ing] their truth.”
186
In Nazir v. United Airlines, Inc., the California Court of Appeal
found that the investigator assigned to investigate a discrimination
complaint was a person “who at least inferentially had an axe to
grind” with the complainant.
187
Further, the “assistant” investigator
considered himself “an internal customer” of the lead
investigator.
188
Thus, the Nazir court held that “[s]uch an
investigation can itself be evidence of pretext . . . . [and] such
184
Reeves v. Safeway Stores, 121 Cal. App. 4th 95, 100 (Cal. Ct. App.
2004). Note that several cases, including a 2011 U.S. Supreme Court decision,
have held that where a supervisor actively discriminates or retaliates against an
employee, but convinces the decision-maker to terminate the employee for non-
discriminatory/non-retaliatory reasons, there may still be liability on the
employer’s part, notwithstanding that the actual decision maker acted without an
improper motive (known as the “Cat’s Paw rule”). See Staub v. Proctor Hospital,
562 U.S. 411, 419–22 (2011). As the Reeves Court stated, “[if] a supervisor makes
another his tool for carrying out a discriminatory action, the original actor’s
purpose will be imputed to the tool, or through the tool to their common
employer.” Reeves, 121 Cal. App. 4th at 113.
185
Id. at 119 (emphasis added).
186
Id. The Court cited the example of the investigator’s use of language in
describing his general role when employees are accused of stealing to the effect
of having a “solid case” against an employee rather than a more neutral
construction. Id. It went on to point to his language of referring to the complainant
as “victim” rather than complainant or accuser and the respondent as a “suspect.”
Id. at 120. The Court further noted the respondent/plaintiff’s rendition of events
is “presented as a story, while the [complainants’] version is presented as history”
in the investigation report. Id. The Court also noted that the investigator used
“vaguely disparaging language.” Id. The Court held that a fact-finder could find
that the investigator failed to investigate exculpatory information and that the
investigator may also have presented plaintiff’s case to management without
describing the “numerous potential ameliorating circumstances.” Id.
187
Nazir v. United Airlines, Inc., 178 Cal. App. 4th 243, 277 (Cal. Ct. app.
2009).
188
Id.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 449
investigation could ‘exploit a disciplinary process predisposed to
confirm all charges.’”
189
There are numerous additional cases in which courts have
concluded that workplace investigations were biased––cases in
which the investigation was tainted by personal bias and self-
interest;
190
cases where the respondent was provided no opportunity
to respond;
191
and cases in which the investigator empathized with
witnesses’ statements, demonstrating an apprehension of bias.
192
189
Id. at 277 (alteration in original). The Nazir court raised the point that this
particular employer (United Airlines) had extensive rules and policies about the
investigation of employee complaints. One of United’s policies stated that if
“there is any reason you would not be perceived as an unbiased investigator,
choose another investigator.” Id.; see also Gee v. Principi, 289 F.3d 342, 346 n.2
(5th Cir. 2002) (“If ultimate decisionmaker [is] influenced by others who had
retaliatory motives, then [the] investigation cannot in any real sense be considered
independent.”).
190
Quela v. Payco-General Am. Credits, Inc., No. 99 C 1904, 2000 U.S. Dist.
LEXIS 6932, at *2 (N.D. Ill. May 17, 2000) (finding an investigation of sexual
harassment allegations tainted by personal bias and self-interest when the
investigator held a high managerial position, a person who aided in the
investigation was the business partner of the accused, and pressure tactics were
employed when interviewing a witness).
191
See Downham v. County of Lennox and Addington, 2005 Can LII 45197
(Ont. S.C.). The Court determined that the investigation was “biased, shoddy, and
substantially undocumented.” Id. at para. 233(b). Plaintiff, Manager of Non-Profit
Housing Projects with the Defendant, was terminated following investigation for
breach of trust. Id. at para. 7. Bias appears to have arisen out of concerns around
the plaintiff (who was an ordained minister for the Salvation Army) providing
assistance to a convicted pedophile in finding public housing. Id. at para. 24–25,
30. Court found the employer “acted with bias and malice throughout this
unfortunate story” by, inter alia, failing to interview plaintiff and rushing to a
conclusion. Id. at para. 181.
192
Lohse v Arthur (No. 3), [2009] FCA 1118, ¶ 53(f) (Austl.). In Lohse, an
employee was demoted based on investigation argued denial of procedural
fairness in investigation. Id. at ¶ 8. The court found the investigation biased
because the employee was not fully advised of all allegations against him, was not
provided with an opportunity to orally respond to the allegations, and investigator
made comments empathizing with witnesses. Id. at ¶ 53(e). But see Kohler v.
Inter-tel Technologies, 244 F. 3d 1167, 1181 (9th Cir. 2001) (determining that,
based upon the well-executed investigation, as well as the employee’s failure to
utilize the employer’s well-drafted grievance procedures, the employer had an
affirmative defense to the allegations of sexual harassment that shielded it from
liability).
450 JOURNAL OF LAW AND POLICY
These cases indicate that courts may scrutinize investigations to
determine whether actual or perceived bias tainted an investigation.
While these cases do not specifically refer to confirmation bias, it
seems to have played a role, as the evidence suggested that the
investigators sought to confirm initial hypotheses while ignoring
alternatives. In most cases, however, cognitive biases of any sort
will be difficult, if not impossible, for courts to identify because
there will be no objective evidence of their existence, let alone how
they impacted an investigation. Yet, workplace investigators
committed to neutrality ought to be equally concerned about hidden
biases as they are about more manifest biases.
IV. BIAS REDUCTION STRATEGIES FOR WORKPLACE
INVESTIGATIONS
“We should envision neutrality as an unending search,
not a state of being.”
193
Some research suggests that unconscious biases are, to some
degree, malleable and can be overcome, or their impact diminished,
at least temporarily.
194
Numerous measures show promise in
overriding unconscious biases. A number of social psychologists
assert that “aware[ness] of discrepancies between conscious ideals
and unconsciously held negative stereotypes” is the first critical step
“toward [eliminating] discriminatory belief systems,” depending on
the “knowledge, guilt, and motivation to change one’s discovered
unconscious biases.”
195
However, research also indicates that it is no easy task to purge
unconscious biases from the decision-making process, as they are
193
Izumi, supra note 24, at 22.
194
Pollard, supra note 52, at 924; see Irwin & Real, supra note 53, at 8.
“Irene Blair summarized a number of similar studies and concluded that implicit
biases are malleable. For example, implicit gender stereotypes of feminine
weakness were reduced by imagining examples of counter-stereotypic (i.e.,
strong) women.” Greenwald & Kreiger, supra note 6, at 963–64; see Irene V.
Blair et al., Imagining Stereotypes Away: The Moderation of Implicit Stereotypes
Through Mental Imagery, 81 J. PERSONALITY & SOC. PSYCHOL. 828, 828 (2001).
195
Pollard, supra note 52, at 924; see Dasgupta & Rivera, supra note 88, at
277.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 451
“particularly resistant to conscious efforts.”
196
Overcoming them
requires significant effort. For implicit biases, people must
recognize the stereotype when it is activated and have the motivation
to control its activation and application.
197
People with low
prejudices are more likely to be motivated to reduce the discrepancy
between unconscious biases and internal belief systems.
198
As one
commentator stated, “[i]n order to conform one’s behavior with
one’s low-prejudiced belief system, one must inhibit spontaneous
stereotype-based re sponses and deliberately replace them with
belief-based responses.”
199
The research suggests that this is a
gradual process like breaking a habit, the first step of which is
awareness.
200
This section will explore possible de-biasing strategies based
upon social science research, as well as suggestions that workplace
investigators can build into their own processes to help to reduce
unconscious bias and the appearance of bias. It will also briefly
explore whether retaining internal or external investigators, or
regulation of workplace investigators might assist in reducing bias.
A. Bias Reduction Strategies as Shown in Psychology
1. Accessing a Counterstereotype
Several studies indicate that exposure to counterstereot ype cues
can reduce implicit bi as es in some circumstances.
201
One study
196
Roberts, supra note 66, at 851 (quoting Levinson, supra note 51, at 361).
One study of Dutch independent crime analysts used as a method to counter tunnel
vision of Dutch police “found that their knowledge of the primary investigators’
theory of the case influenced them.” O’Brien, supra note 120, at 330. They were
more likely to come to similar interpretations than were analysts blind to that
information. Id. “This finding suggests just how difficult it can be to purge these
biases from the decision-making process, even for people who are highly trained
and motivated to be accurate.” Id.
197
Blasi, supra note 100, at 1252–53.
198
Pollard, supra note 52, at 949.
199
Id. at 950.
200
Id. at 951; Izumi, supra note 24, at 141 (“Awareness of bias is critical for
mental decontamination success.”).
201
Blair et al., supra note 194, at 828; Blasi, supra note 100, at 1254; Izumi,
supra note 24, 148–49; Pollard, supra note 52, at 953.
452 JOURNAL OF LAW AND POLICY
demonstrated that exposure to admirable members of a stigm atized
group reduced automatic bias against that group.
202
The effect lasted
at le ast one day.
203
Other studies show that “participants who engage
in counterstereotypic mental imagery produced substantially weaker
implicit stereotypes compared with participants who engage[] in
neutral, stereotypic, or no mental imagery.”
204
The results suggest
that implicit stereotypes are malleable, and that controlled
processes, such as mental imagery, may influence the stereotyping
process at its early as well as later stages.
205
However, conjuring up a counterstereotype is not an elixir.
Correction strategies “require significant cognitive resources” and,
therefore, high levels of consistent motivation.
206
Further,
“overcorrection is a common problem” amongst “those who are
motivated to avoid discrimination.”
207
Such interventions may
temporarily activate a subtype of a larger category. “Once activated,
this subcategory would temporarily function as a mental
replacement for the larger (and presumably more negatively
valenced). . . category.”
208
Suppressing stereotypes has also been
found to be counterproductive because it can actually result in the
202
Izumi, supra note 24, at 148–49.
203
Id. Other studies have reached similar results. See Nilanjana Dasgupta,
Seeing is Believing: Exposure to Counterstereotypic Women Leaders and Its
Effect on the Malleability of Automatic Gender Stereotyping, 40 J. EXPERIMENTAL
SOC. PSYCHOL. 642, 645 (2004) (explaining that exposure to women in leadership
positions can reduce implicit biases in women); see also Dasgupta & Rivera,
supra note 88, at 268 (describing the results of studies showing “that, for both
sexes, automatic prejudice produced biased behavior in the absence of conscious
egalitarian beliefs and behavioral control” and that “[t]he presence of either
conscious process eliminated behavioral bias”); Greenwald & Krieger, supra note
6, at 963 (finding that when “White and Asian-American undergraduate
students . . . complete[d] . . . a preliminary task in which they identified a series
of photographs of well-known and admired African Americans, . . . mixed with
photographs of somewhat less well-known but thoroughly disreputable European
Americans (terrorists and serial murderers) . . . [a] subsequent Race IAT measure
revealed that this photograph-identification task reduced the level of automatic
preference for European American (relative to African American)”).
204
Blair et al., supra note 194, at 828.
205
See Pollard, supra note 52, at 924; see also Blair, supra note 194, at 828.
206
Page, supra note 61, at 239–40.
207
Id.
208
Greenwald & Krieger, supra note 6, at 964.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 453
“hyerpaccessibility” of the stereotypes and “increased stereotyping
at the next opportunity.”
209
While accessing counterstereotypes may not be a guaranteed
method of reducing bias, there is some indication that it can assist at
least on a temporary basis. Given this, it seems worthwhile for
workplace investigators to invoke this technique before or during
interviews and while assessing evidence in an effort to keep
unconscious stereotyping from seeping into in vestigative findings.
2. Education an d Tr ai ni ng for Investigators
Studies show that training may also assist in countering
unconscious biases for investigators.
210
To be effective, however,
training programs must have relevant content and be consistent,
continuous, and long-term, as most bias reduction strategies are only
temporary measures.
211
The most successful training will “focus[]
on the cognitive processes that can lead to bias.”
212
Multiculturalism training in particular has been successful in
reducing implicit biases against social groups.
213
In one study,
participants exposed to multiculturalism training displayed lesser
implicit bias on a black/white IAT than those who had been exposed
to training based on color blindness, thus ignoring differences
between groups.
214
In another study, participants in a fourteen-week
209
Page, supra note 61, at 241. In one study, “subjects were shown a picture
of a ‘skinhead’ and asked to write an essay on his probable qualities,” having been
instructed to avoid stereotypes. Blasi, supra note 100, at 1253. They successfully
did so. Id. However, when the participants “were later exposed to a skinhead
image, they were more likely than [the] control [group] to activate the stereotype.”
Id. at 1253 (emphasis added).
210
See, e.g., Springer, supra note 130, at 321 (demonstrating that “simply
educating people about the existence of expectancy effects virtually eliminate[s]
their existence in laboratory settings”).
211
See Greenwald & Krieger, note 6 supra, at 964; Page, supra note 62, at
242
212
Burke, supra note 121, at 522. One study found that a semester long
diversity course significantly reduced automatic stereotyping with regards to race.
See Page, supra note 61, at 242 n.446 (citing research showing that completing a
diversity education program reduced implicit bias, at least temporarily).
213
See Levinson, supra note 51, at 414–15.
214
Id. at 415.
454 JOURNAL OF LAW AND POLICY
diversity training course that involved learning about intergroup
conflict displayed less implicit and explicit bias than those in the
control group.
215
Research has even indicated that simply taking the
IAT and seeing the results can assist in addressing implicit bias.
216
This could be an appropriate starting point to educate workplace
investigators about implicit bias by enabling an understanding of
baseline cognitive biases so that they may compensate through their
higher cognitive processes.
217
Notwithstanding the possible benefits of education and training,
“the empirical evidence also suggests that cognitive bias is stubborn,
and that education is an unlikely panacea.”
218
Indeed, similar to
accessing a counterstereotype, studies indicate that certain bias
reduction trainings might, in fact, heighten some biases.
219
Simply
knowing one’s prejudice level is insuffic ie nt to respond in a less
prejudiced manner; internal motivation is necessary.
220
Training
215
Id.
216
See id. at 416.
217
See Roberts, supra note 66, at 858, n.244 (discussing use of the IAT for
judicial education); see also Justin D. Levinson & Danielle Young, Implicit
Gender Bias in the Legal Profession: An Empirical Study, 18 DUKE J. GENDER L.
& POL’Y 1, 38 n.203 (2010) (“Within the medical profession attention has been
given to devising training programs geared towards eliminating bias-driven health
care disparities. Enhance internal motivation to reduce bias, increase
understanding about the psychological basis of bias, enhance providers’
confidence in ability to successfully interact with socially dissimilar patients,
enhance emotion regulation skills and improve ability to build partnerships with
patients.”).
218
Burke, supra note 121, at 523.
219
See Irene V. Blair & Mahzarin R. Banaji, Automatic and Controlled
Processes in Stereotype Priming, 70 J. PERSONALITY & SOC. PSYCHOL. 1142,
1159 (1996).
220
Recent studies show that while stereotypes may be automatically
activated, as conscious actors we may be able to affect the application of those
stereotypes in our interactions, judgments, and decisions. Irene Blair and
Mahzarin Banaji conducted a series of four experiments to observe the automatic
activation of gender stereotypes and to assess conditions under which stereotype
priming may be moderated under particular conditions. See id. They concluded
that “stereotype priming can be eliminated when perceivers have an intention to
process counterstereotypic information and sufficient cognitive resources are
available.” Id.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 455
would, therefore, have to be coupled with other de-biasing strategies
incorporated into daily practice.
Given that workplace investigators are largely unregulated, any
requirement to attend de-biasing training and/or to take the IAT,
must be mandated, at least for internal investigators, by their
employers. Otherwise, workplace investigators should seek such
training out on their own. However, a deeper understanding of the
psychological processes at play appears to be necessary to obtain
any kind of lasting motivation and impact. While long-term training
may assist, workplace investigators are largely unregulated and are
not currently required to attend any sort of de-biasing training. For
the time being then, it is incumbent upon individual workplace
investigators to acquire such education themselves.
3. Appealing to Individuals’ Beliefs in Equality
and Fairness
Other studies show that “conscious processes such as the
motivation to be egalitarian. . . can circumvent the effect of
automatic prejudice on outward behavior.”
221
In other words,
“conscious egalitarian beliefs can override the effect of
automatically activated prejudice and prevent certain forms of
behavioral bias.”
222
Studies show that this can be accomplished by “priming subjects
with fairness or egalitarian goals, activat[ing] unconscious
cognitions [to] . . . counter the effects of the automatic activation of
stereotypes, leaving the entire battle between fairness and prejudice
to be played out at a subconscious level.”
223
For example, one study
showed that simply seeing or hearing the word “fairness” caused
subjects to behave as though they were more committed to being
fair.
224
This suggests that workplace investigators or others who
want to counter automatic stereotyping effects “should insert into
their presentations as many contrarian priming words, pictures, and
other stimuli as possible.”
225
221
Dasgupta & Rivera, supra note 88, at 277.
222
Id.
223
Blasi, supra note 100, at 1254.
224
Id. at 1277.
225
Id.
456 JOURNAL OF LAW AND POLICY
Workplace investigators, as well as organizations who require
investigations, might apply this research to their practices or
policies, as well as any investigation training an organization might
provide. For example, this might be done by incorporating the
practice of writing “fairness” or “neutrality” on investigators’ note
paper before interviews, incorporating those words into their
opening statements made to interviewees, or simply making it part
of their routine to consciously think about their egalitarian views
prior to working on an important component of the case.
4. Considering Altern ative Hypothesis or Why a
Favored Hypothesis Is Wrong
Research also suggests that considering alternative hypotheses
can counter cognitive biases.
226
“[T]aking the extra step of actively
considering evidence that points away from [a given] suspect shows
promise as a simple way to counteract bias.”
227
It “breaks the
inertia” that comes from focusing exclusively on data that is
consistent with an initial hypothesis. This may require investigators
to articulate reasons that counter their own conclusions. Research
also indicates that considering alternatives may “activate a mind-set
that reduces bias in later, unrelated judgments.”
228
Employers may
assist in reducing bias in investigations conducted on their behalf by
building these kinds of recommendations into their workplace
investigation policies and procedures.
However, people argue against their own position reluctantly,
and even when they do so, overconfidence in their positions is
generally reduced but not eliminated.
229
Further, some studies have
shown that considering alternatives “can worsen bias when people
226
See Burke, supra note 122, at 523–24; Findley & Scott, supra note 116,
at 388–89 (discussing how prosecutors can counter bias by stepping into their
opponent’s shoes); O’Brien, supra note 120, at 329.
227
O’Brien, supra note 120, at 329.
228
Galinsky and Moskowitz found that study participants made to consider
counterfactuals performed better later on the rule-discovery task, which requires
people to resist the tendency to seek only evidence that confirms their initial
hypotheses. “That is, people who considered how a situation might have turned
out differently were less prone to bias in a later, unrelated task.” Id. at 317.
229
See Findley & Scott, supra note 116, at 352 n.371.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 457
find it hard to generate plausible alternatives.”
230
It may be
particularly difficult for certain professionals to consider
alternatives. Studies indicate that professionals such as police,
lawyers, and judges, suffer from “blindness to alternatives.”
231
One
example of this can be seen in a study of police trainees serving as
mock investigators.
232
The trainees “received a piece of evidence
that either confir med or disconfirmed their prior suspicion against
the suspect” in a homicide case.
233
“Despite identical objective
characteristics of the evidence, participants rated the disconfirming
(vs. confirming) evidence as less reliable and generated more
arguments to questi on it s rel iabil it y.”
234
To neutralize confirmation bias, a workplace investigator should
recognize the initial hypothesis he or she creates and consciously
230
See O’Brien, supra note 120, at 330. In another study participants
provided with a case vignette where there was a serious possibility of another
perpetrator than the main suspect rated exonerating information no differently in
terms of incriminating power than those who read a version in which the suspect
had a clear motive and no other suspects were mentioned. See Ask & Granhag,
supra note 137, at 478 (“The pivotal decision investigators must make regarding
whether to further interrogate a suspect may be based on prejudgments of guilt,
confidently made, but frequently in error.”). Similarly, Kurt A. Carlson and J.
Edward Russo found that jury members tend to interpret information in the light
of their previously held convictions rather than completely objective. See Kurt A.
Carlson & J. Edward Russo, Biased Interpretation of Evidence by Mock Jurors, 7
J.
EXPERIMENTAL PSYCHOL. 91, 100–01 (2001). Ask, Rebelius, and Granhag
found that their participants had more faith in DNA, video, and identification
evidence if it produced incriminating information compared with when the
outcome of the same procedure was exonerating. See generally Ask et al., supra
note 121; see also Rassin, supra note 120, at 154.
231
See Rassin supra note 120 (finding that the estimation of the
incriminating power of investigation findings by police offers, district attorneys,
and judges was not affected by knowledge of an alternative suspect; a result that
did not emerge amongst university students).
232
Ask et al., supra note 121, at 1245.
233
Id.
234
Id. (“This asymmetrical scepticism was stronger for participants judging
witness evidence, compared to DNA, and photo evidence, supporting the
hypothesis that different types of evidence vary in ‘elasticity’—the extent to
which subjective interpretations can be justified. Interestingly, the observed
effects were not limited to the specific evidence in the case, but also affected the
ratings of the type of evidence in general, suggesting that reliability criteria for
witness information are highly malleable and sensitive to contextual influences.”).
458 JOURNAL OF LAW AND POLICY
step away from that hypothesis to consider alternatives, however
unlikely they may seem. Beyond that, workplace investigators
should also force themselves to articulate reasons that his or her
conclusions may be wrong and other conclusions may be correct.
B. Building “De-biasing Tools” into Investigative
Procedures
Workplace investigators may also help themselves reduce their
own biases through the implementation of a set of investigative
procedures and practices that incorporate “de-biasing tools.”
Organizations may also build these kinds of tools into their policies
and procedures to assist those who conduct investigations on their
behalf (whether internal or external), as well as any training that they
provide. These tools, if consistently implemented, may assist a
workplace investigator in reducing biases at the conscious and
unconscious levels, as well as the appearance of bias.
Some commentators have warned that the very informality of
alternative dispute resolution (“ADR”) processes increases the
likelihood of bias.
235
However, other commentators have argued that
an adversarial system is predisposed to bias because it “polarizes
parties and forces them to dogmatically pursue their own
interests.”
236
By contrast, inquisitorial processes like workplace
investigation avoid biasing pressures because they are foun ded upon
the concept of “neutrality and a search for the truth rather than
advocacy and a quest for victory.”
237
In either event, it is clear that
workplace investigators must do what they can to achieve neutrality
235
See, e.g., Gary LaFree & Christine Rack, The Effects of Participants’
Ethnicity and Gender on Monetary Outcomes in Mediated and Adjudicated Civil
Cases, 30 L. & SOC’Y REV. 767, 789 (1996) (finding that ethnic and gender
disparities are greater in mediation than in adjudication.); Izumi, supra note 24, at
103 (“Because formal adjudication explicitly manifests ‘societal norms of fairness
and even-handedness’ through symbols (flag, black robe), ritual, and rules, the
adversarial process counteracts bias among legal decision makers and disputants.”
(quoting Richard Delgado et al., Fairness and Formality: Minimizing the Risk of
Prejudice in Alternative Dispute Resolution, 1985 WIS. L. REV. 1359, 1387–88)).
236
Findley & Scott, supra note 116, at 323.
237
Id.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 459
and postpone judgment.
238
To be sure, “suppression of stereotyped
associations and engagement of non-prejudiced responses requires
intentional attention, and effort.”
239
Investigators need to act
deliberately and consistently in implementing bias reduction
strategies into each stage of their process. It is particularly important
for workplace investigators to have these processes well established,
as investigations often require a quick turnaround, leaving little time
to reflect and often instilling pressure that may derail attempts to
reduce bias.
1. Limiting Information Provided Pre-Interviews
and Client Contact throughout Investigation
External workplace investigators ought to consider carefully
how much contact to have with clients at the beginning and
throughout the investigation. It may be that less is more in terms of
avoiding actual bias, the appearance of bias, and unconscious bias.
A client will often have a particular belief in the “correct” outcome
for an investigation, which they may communicate to the workplace
investigator when he or she is retained. This can create an “initial
hypothesis” in the workplace investigator’s mind and lead him or
her down the dangerous road of confirmation bias.
Workplace investigators would be well served to implement a
provision into their retainer agreements that makes it clear from the
outset the circumstances in which there will be client contact and for
what purpose. Of course, thi s is not to say that workplace
investigators ought to avoid necessary client contact. However,
limiting the opportunities for discussion about the case while it is
ongoing may help to avoid exposure to potentially biasing
information from the client that may lead to tunnel vision.
In addition, workplace investigators ought to give thought to
how much information they must learn prior to interviews.
Investigators may be exposed to case information through the
referral process, initia l client contact, scheduling calls, and
document review. When investigators are exposed to information
about the underlying complaint early on, they risk forming early
238
See id. at 372.
239
Izumi, supra note 24, at 141.
460 JOURNAL OF LAW AND POLICY
hypotheses and biases that may grow stronger as a result of
confirmation bias.
240
There will certainly be times when a workplace investigator
must have initial information to conduct an investigation properly.
However, investigators should consider how much information is
necessary and/or beneficial to review pre-interview, as opposed to
reviewing it post-interview and following up thereafter if necessary.
Considerations will often be different for internal investigators,
who may have at least some knowledge of the parties involved and
the underlying dispute. It may be useful for organizations to
carefully draft their policies and procedures to clarify such things as
who is to conduct investigations, who is to have knowledge of the
underlying complaint, and leave some level of discretion for
referring a matter to an external investigator where the connection
to the parties or the dispute increases the likelihood of bias.
2. Interviewing Practices
Workplace investigators, as well as organizations that employ
investigators, should ensure that interviewing practices employed in
investigations are consistent to reduce the risk of bias preventing a
fair process or outcome. Investigators ought not rely upon memory
in any way to reach their findings or conclusions in a case,
particularly in light of the research on memory bias.
241
Instead,
workplace investigators should take careful notes in interviews. The
more detailed the notes, the better, particularly in a drawn-out
investigation with numerous witnesses.
In addition, investigators should consider having a standard
statement that they recite at the beginning of interviews both to
provide clarity to the participants about the investigators’ role as a
neutral, and to remind themselves of the importance of neutrality
and fairness in the investigati on process.
240
See Powell, supra note 120, at 132.
241
See Page, supra note 61, at 221 (showing studies that demonstrate how
people tend to recall information in a biased manner and in conformity with their
expectations); see also Findley & Scott, supra, note 116, at 312 (showing that a
known hypothesis can bias the manner in which people search their memories for
confirming information).
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 461
Further, investigators sh ould always be conscious of the critical
importance of listening during the interviews. Despite the popularity
of the Reid Technique, the research indicates that listening to
interviewees is more important than assessing visual cues in
detecting truth and deception.
242
Further, taking the focus off visual
cues may reduce lie bias. This means that an investigator should be
listening to auditory cues and, more importantly, to how evidence is
provided. “Facts that are vague, contradictory, or sketchy can be
invalidated while other information, given in an honest and
straightforward manner, without any attempt to hide or exaggerate,
prove credible.”
243
Recording interviews or using a verbatim
reporter may be helpful to investigators in this regard.
Further, many studies indicate that utilizing open questions, as
opposed to leading questions, can help minimize the detrimental
effects of confirmation bias.
244
The effect of confirmation bias is not
observed in interviewers who adhere to open questions because they
allow the “witnesses to talk, which in turn, masks any
preconceptions about the event that the interviewer may have had
during and prior to the interview.”
245
Even where “confirmation bias
occurs automatically when interviewers receive knowledge about a
case,” investigators can “overcome this bias if . . . previously trained
to . . . ask open questions.
246
In addition, workplace investigators must ensure that they are
utilizing neutral language. Investigators should avoid making
statements that may be interpreted as “sympathetic” to a witness,
including “I can understand that,” “I’m not doubting that,” and “I
agree.” This ki nd of language can backfire and have the effect of
242
Kassin, supra note 150, at 810; see Vrij, supra note 155, at 1333.
243
Avitabile & Kleiner, supra note 145, at 222.
244
See, e.g., Powell, supra note 120, at 131 (“Interviewers who were better
at adhering to open questions . . . showed less of a detrimental effect of
confirmation bias compared with interv iewers who were poorer at adhering to
open questions.”).
245
Id. at 127.
246
Id. at 131–32 (discussing the benefits of open questions, including “more
accurate witness statements, longer witness responses, encourage[ing] witnesses
to play an active role in the interview process, enhance[ing] witness’ perceptions
that they were listened to, [and maximizing] story-grammar detail and thus victim
credibility”).
462 JOURNAL OF LAW AND POLICY
suggesting the investigator does not have an open mind about certain
matters, thus leading to an apprehension of bias.
247
Eliminating
these kinds of phrases will require effort for investigators who use
such expressions in an unconscious attempt to encourage witnesses
to trust them.
Finally, workplace investigators should avoid suggesting
answers to interviewees. This is likely to be an issue for an
investigator who conducts a time sensitive investigation or
interviews a difficult witness who does not answer questions
concisely or clearly. In these cases, an investigator may simply want
to elicit evidence quickly, but this may taint the investigation as
interviewees may be influenced by the suggested answers and
thereby give inaccurate evidence.
3. Review of Evidence and Findings
Implementing a peer review process may also assist in ensuring
that findings are supported. Involving others in the decision-making
process provides a “fresh look” and may help to reduce cognitive
biases that can lead to tunnel vision and, perhaps, wrong
conclusions.
248
This may be done by reviewing evidence obtained
in an investigation with a trusted colleague, preferably without
providing the investigator’s vi ews , which could bias the colleague.
Of course, confidential information must be protected in any such
process.
Another useful process that may reduce bias relates to the
manner in which evidence is laid out, particularly in cases involving
large volumes of evidence where one may be particularly prone to
neglecting conflicting evidence. One study found that a graphical
evidence layout significantly reduced, though did not eliminate bias,
as compared to an informationally equivalent text format. A
graphical layout “makes the entire spectrum of evidence available
and easy to assimilate, so that evidence across the spectrum, not just
supporting evidence, stands a greater chance of being incorporated
247
Lohse v Arthur (No. 3), [2009] FCA 1118, ¶ 53(e) (Austl.) (“Even if the
evidence [of such language] . . . was not sufficiently strong to establish actual
bias . . . it was sufficient . . . to draw a conclusion of apprehended bias sufficient
to constitute a denial of natural justice.”).
248
Burke, supra note 121, at 525.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 463
into a decision.”
249
The authors of this study speculated that the
graphical format was helpful because “it plays to the recognition-
centred nature of decision making by making the full distribution of
evidence constantly available and easy to access.”
250
C. Does Retaining External Investigators Solve the
Problem of Bias?
Some argue that bias is inherent where staff conducts internal
investigations, as they have ongoing relationships with the
organization, are dependent upon the organization for their
livelihood, and often know the parties and inner workings of the
organization involved.
251
Some commentators believe that this
means that internal staff cannot be sufficiently separate from the
organization for true independence in decision-making and will
drive them to “predetermined outcomes.”
252
Thus, many companies rely on external investigators where
there would otherwise be a possibility or perception of bias if an
internal investigator were used, including in cases involving
allegations against a supervisor, where the internal employee has a
relationship with one of the parties, or where sensitive allegations
are raised that may harm the reputation of the company if true.
253
In
249
Maia B. Cook & Harvey S. Smallman, Human Factors of the
Confirmation Bias in Intelligence Analysis: Decision Support From Graphical
Evidence Landscapes, 50 J. HUM. FACTORS & ERGONOMICS SOC’Y 745, 747
(2008).
250
Id. at 752.
251
See, e.g., Harriet Stacey, Addressing the Shortcomings of Workplace
Investigations: A Response to Josh Bornstein’s Call for Reform, CORRUPTION
PREVENTION NETWORK 2 (June 2014), http://www.corruption
prevention.net/assets/Uploads/Addressing-the-shortcomings-of-workplace-
investigations.pdf (“Internally conducted investigations were often subject to bias
by the staff conducting the investigation.”).
252
Id.; see also GUIDING PRINCIPLES, supra note 4 (“Employers may choose
to use an in-house (internal) investigator. In such cases, the internal hierarchy of
the organization should be considered to avoid the perception of biased or
compromised objectivity.”).
253
See, e.g., Lisa Corrente, Workplace Investigations: Key Reasons for
Hiring an External Investigator, CANADIAN ASS’N OF WORKPLACE
INVESTIGATORS (2015), http://www.caowi.org/key-reasons-for-hiring-external-
464 JOURNAL OF LAW AND POLICY
many ways, an investigation that has been conducted by an external
investigator can be easier to defend than one conducted internally in
terms of evidencing neutrality and avoiding at least the appearance
of bias. It can be very difficult for an investigator w ho has an
existing and ongoing relationship with the employer to be truly
neutral and unbiased.
On the other hand, external investigators face issues “stemming
from lack of industry standards [for those conducting workplace
investigations] and bias resulting from the [i]nherent tension
between independence and financial imperative to meet the needs of
the client.”
254
Further, if an organization repeatedly uses one or two
investigators, these “externals” may face the same biases as internal
investigators—knowledge of the inner workings of the organization,
dependence upon the organization, and ongoing relationships with
those who may impact their livelihood.
Ultimately, where there is any possibility of bias, a company
would be best served by retaining an external investigator who has
no stake in the outcome. This however will not necessarily eliminate
the problem of bias, as both external and internal investigators will
have unconscious biases. When it comes to unconscious biases,
rather than actual or perceived biases, the mo re relevant determining
factor may be what processes and tools the investigator has in place
to reduce the seepage of unconscious biases into his or her
investigations.
D. Would Regulation Assist in Reducing Bias?
It is important to consider whether increased regulation of
workplace investigators could assist in reducing bias. As discussed,
there is no governing legislation specifically applicable to
workplace investigators. Although private investigators and lawyers
are covered by their professions’ professional standards and rules of
conduct, these do not address the unique issues that workplace
investigators face.
255
investigator (mentioning the need for impartiality and the impact on the
company’s reputation, among others, as factors in hiring external investigators).
254
Stacey, supra note 250, at 2.
255
See Smiley, supra note 43, at 219.
THE HIDDEN WORLD OF UNCONSCIOUS BIAS 465
There are many benefits to regulating an industry, including
engendering the public trust. Without regulation and a unified code
of con duct, workplace investigators and the public may be less
likely to trust the process of workplace invest igations. Without clear
guidance, some may avoid the role altogether or “be forced to
choose a guiding standard from a hodgepodge of other, often
conflicting standards.”
256
A unified code would inform investigators
of their duties “while simultaneously accommodating the interests
of the Bar, society, and the participating parties.”
257
It may also
impose a moral imperative to avoid bias or discrimination in
investigations.
258
Perhaps most importantly with respect to the issue of reducing
unconscious bias, regulation would allow the imposition of certain
minimum training requirements for workplace investigators.
Lawyers and private inve stigator s are required to have certain
training within their own professions,
259
yet that training is not
geared towards the unique issues facing workplace investigators.
While there are many certification programs and training programs
for workplace investigators, there is no requirement for workplace
investigators to participate. Therefore, investigators may be
investigating serious matters of workplace misconduct without any
relevant training whatsoever and certainly without training around
bias. Of course, the training must then specifically address
unconscious bias.
While regulation may serve many useful purposes, it would not
be a complete solution for addressing unconscious bias in workplace
investigations. Many other professionals, such as judges, are
required to be impartial pursuant to governing ethical rules. And
still, unconscious bias “can impair judges’ ability to align their
conduct with what is ethical.”
260
Ethical rules for such professions
arguably need to be updated to address unconscious biases.
Regulation of model rules of conduct for workplace investigators
would at least provide a possible means for addressing the issue of
256
Id. at 246.
257
Id.
258
See Izumi, supra note 24, at 103.
259
See Smiley, supra note 43, at 219.
260
Roberts, supra note 66, at 839.
466 JOURNAL OF LAW AND POLICY
unconscious bias on a profession-wide basis. Without it, workplace
investigators are left to their own devices, and perhaps the efforts of
their employers, to understand these cognitive processes and to
implement appropriate de-biasing tools into their practice.
CONCLUSION
Neutrality is essential to the role of the workplace investigator.
The science around unconscious bias clearly indicates that even
those workplace investigators with a strong professional
commitment to neutrality are at risk of succumbing to automatic
cognitive processes that can unconsciously lead to biased
conclusions. It is unlikely that bias can be completely eliminated
from the workplace investigation process, given the automatic
nature and deep entrenchment of unconscious biases in the human
psyche. Yet, if there is to be any hope of reducing such biases in
workplace investigations, workplace investigators must recognize
the multiple causes and manifestations of biases, and devise
deliberate, consistent, and informed practices aimed at guarding
against their detrimental effects.
