Affirmative Action: Psychological Data and Policy Debates

1
Affirmative Action Psychological Data and
Policy Debates
Aarti Iyer University of Exeter in
collaboration with Faye J. Crosby, Susan Clayton,
Roberta Downing, Sirinda Sincharoen
7 May 2007 University of Utrecht
2
social policy to combat inequality

• history provides examples of inequality between
  social groups
• in nearly every society
• ? gender, race/ethnicity, caste, socio-economic
  class
• remarkable progress has been made in the general
  
• awareness that such group-based inequality is
  illegitimate
• ? discrimination has been outlawed (and
  punished)
• individual and societal levels
• ? expression of overt prejudice is socially
  unacceptable
• social policy has taken a key role in
  eliminating group-based
• discrimination and promoting equality
• such affirmative action (AA) programs are
  controversial and
• subject to passionate debate
• ? social psychologists can offer theory and data
  to address some of these questions

3
overview of talk

• what AA is
• why AA is needed
• AA strategies legislation
• evaluation of AA
• strategies to improve effectiveness of AA

4
what AA is

• aim of equal opportunity (EO) policies
• 1. eliminate differential (often biased)
  treatment on the basis of group membership
• assumption fairness is achieved once
  intentional, explicit discrimination is
  prohibited
• aims of AA policies
• 1. eliminate differential (often biased)
  treatment on the basis of group membership
• 2. proactively improve the status of target
  groups
• ? assumption good intentions are not enough to
  ensure equality of representation and status

5
why AA is needed

• 1. policies and practices are biased
• individuals are still (unconsciously) prejudiced
• stereotypes still widely-held about the
  competence of women (DeArmon et al., 2006
  Wallace, 2001) and ethnic minorities
• (Frazer Wiersma, 2001)
• prejudice is often expressed in covert or subtle
  ways that are hard to identify and police
• -- benevolent sexism (Fiske et al., 1991
  Rudman Glick, 1999)
• -- aversive/modern racism (Dovidio Gaertner,
  2000 Sears, 1994)
• prejudice can lead to discriminatory behavior
• -- performance evaluations of women (Heilman et
  al., 1995)
• and ethnic minorities (Dovidio
  Gaertner, 2000)
• -- selection of women (Lott et al., 2002) and
  ethnic minorities (Staples, 1998) for
  positions
• ? especially in ambiguous or non-traditional cases

6
why AA is needed

• 1. policies and practices are biased
• A. individuals are still (unconsciously)
  prejudiced
• B. structural/institutional barriers still exist
• even with sincere attempts at impartiality at
  the individual
• level, policies and practices may still
  favour certain groups
• advertising opportunities on an ad hoc or
  limited basis
• ? excludes those who are not part of existing
  network
• providing a narrow category of opportunities to
  everyone
• ? limits those who can take advantage of it
• disparate impact of (ostensibly objective)
  standards
• ? vestiges of historical bias
• ? arbitrary choice of criteria

7
why AA is needed

• 2. discrimination is not always challenged
• victims may not come forward
• EO expects victims of discrimination to identify
  and challenge their unfair treatment
• ? does not necessarily happen, for at least 2
  reasons
• 1. individuals often do not recognize their own
  disadvantages
• minimization of personal disadvantage
• (Crosby, 1982 Taylor et al., 1990 Foster
  Matheson, 1995)
• attributional ambiguity (Major et al., 1994)
• 2. individuals often do not file formal
  complaints
• lack of knowledge or resources (Bergmann, 1996)
• discomfort with adversity (Fine Barreras, 2001)
• fear of derogation by colleagues (Kaiser
  Miller, 2001)
• fear of retaliation from authorities (Reskin,
  1998)

8
why AA is needed

• 2. discrimination is not always challenged
• A. victims may not come forward
• B. observers may not notice discrimination
• assumed that those who are not directly
  affected by discrimination
• may be best positioned to identify and
  challenge it
• even observers can have a hard time noticing
  patterns of disparate
• treatment between groups (e.g., Rutte et al.,
  1989 Cordova, 1992)
• a series of experiments presented participants
  with information
• about qualifications and salaries of male and
  female employees
• -- on a case-by-case basis (individual or
  department)
• -- in summary or aggregate form (across an
  organization)
• only when the information was presented in
  summary or aggregate
• form were gender differences in salaries
  noticed and challenged

9
why AA is needed

• 3. left unchecked, social inequality is
  perpetuated
• A. stereotype threat
• awareness of negative stereotype about an
  in-group can
• affect individuals performance in that domain
  
• (Steele, 1997 Steele Aronson, 1995)
• experimental evidence in studies of
• -- women and math performance (e.g., Quinn et
  al., 2001)
• -- African Americans and academic performance
• (Steele Aronson, 1995 Steel, 1997)
• underlying mechanisms protective strategies
  for this effect
• are still being worked out
• ? vicious cycle members of stereotyped groups
  may perform
• worse, which perpetuates stereotype of
  incompetence

10
why AA is needed

• 3. left unchecked, social inequality is
  perpetuated
• A. stereotype threat
• B. psychological disengagement
• existence of inequality may encourage members of
  
• disadvantaged groups to disengage from the
  domain
• research has demonstrated two such strategies
• (Crocker Quinn, 1998 Major et al., 1998
  Schmader et al., 2001)
• devaluing (and distancing from) the domain
• discounting feedback received in the domain
• ? vicious cycle disengagement perpetuates idea
  that disadvantaged group doesnt really belong

11
why AA is needed

• 4. broad benefits of increased diversity
• better performance in organizations
• individual AA hires are typically as strong as
  non-AA hires
• -- similar qualifications (Holzer Neumark,
  1996, 2000)
• -- similar performance on standard indicators
• (Holzer Neumark, 1996 Lott, 2000
  Leonard, 1984)
• diverse workforces offer economic advantages to
  orgs.
• -- varied points of view can increase ability
  to respond to challenges and problems
  (Reskin, 1998)
• -- broadening market for products and services
  to previously
• under-served populations (Crosby
  Herzberger, 1996)
• a caveat
• careful management needed to avoid (personal or
  group) conflict
• (Williams OReilly, 1998)

12
why AA is needed

• 4. broad benefits of increased diversity
• better performance in organizations
• improved learning conditions at university
• exposure to new perspectives, ideas, and beliefs
  helps students engage in deeper and more complex
  learning
• (Gurin, Dey, Hurtado, Gurin, 2002)
• better preparation for future interactions in an
  increasingly diverse society (Gurin et al., 2002)
  
• increased motivation after graduation to seek out
  diversity
• (Orfield, Bachmeier, James, Eitle, 1997)
• AA admits are more likely to contribute to
  disadvantaged communities than are non-AA admits
  (Bowen Bok, 1998)

13
AA programs (1)

• monitoring program to assess whether EO exists
• calculate proportion of positions
  held/promotions earned by members of target
  groups (positions)
• calculate proportion of qualified target group
  members in candidate pool (candidates)
• estimate proportion of qualified target group
  members in broader work-force (work-force)
• Type of Discrepancy Nature of Potential Problem
• candidates lt work-force recruitment procedures
• positions lt candidates selection procedures
• positions lt work-force selection or recruitment
  procedures
• work-force lt population longer-term
  recruitment procedures

Crosby, Iyer, Clayton, Downing (2003) Iyer (in
press)
14
AA programs (2)

• 2a. eliminate discriminatory policies and
  practices
• 2b. proactively increase representation of
  target group
• recruitment procedures
• formalize advertising procedures
• ? increases access and information for all
  groups
• b. expand recruitment beyond traditional sources
  strategies
• ? increases proportion of target group members
  in pool
• offer training mentoring programs
• ? increases knowledge and opportunities in
  target group
• selection procedures
• formalize selection criteria ensure no
  disparate impact
• ? eliminates biases against all groups
• make process transparent and subject to review
• ? eliminates biases against all groups
• b. actively adjust selection criteria (e.g.,
  diversity as plus factor)
• ? increases proportion of target group members
  selected

Crosby, Iyer, Clayton, Downing (2003) Iyer (in
press)
15
positieve aktie in the Netherlands

• since the 1980s, programs have aimed to
  improve the
• employment opportunities of women and ethnic
  minorities
• (e.g., Surinamers, Antillians/Arubans, Turks,
  and Moroccans)
• goal is typically to bring the of target
  group members in
• organisation up to the of target group in
  region
• programs are narrowly targeted to job/occupation
  and group
• -- increasing number of women railroad
  conductors
• -- increasing number of ethnic minority police
  officers
• positieve aktie efforts span a range of
  activities
• -- recruitment
• -- training
• -- selection procedures and targets

De Vries (1992) Zwamborn (2003)
16
positive outcomes of AA

• AA has achieved positive outcomes in
  organizations
• increased representation and status of target
  group members in the USA (Leonard, 1984 Holzer
  Neumark, 2000), Canada (Leck Saunders, 1992
  1996), Australia (French, 2001)
• no decrease in overall productivity or efficiency
  in the USA
• (Lott, 2000 Holzer Neumark, 1996)
• in some cases, AA organizations report better
  performance and expanding profits (Reskin, 1998)
• but the story is not completely positive

Iyer (in press)
17
problems with AA

• representation of target group members still not
  equivalent to representation in population
• opposition to AA can prevent effective
  implementation
• required changes in policy and practice may be
  difficult to
• sustain without formal oversight and enforcement
• organisations often face competing priorities for
  their scarce resources
• AA only works to the extent that it is properly
  and consistently implemented

Iyer (in press)
18
problems with AA

• representation of target group members still not
  equivalent to representation in population
• 2. members of target groups feel stigmatized by
  (presumed) association with AA
• experimental research suggests that individuals
  selected on the basis of group membership feel
  stigmatized compared to those selected on the
  basis of merit alone
• -- doubts about self-competence (Heilman et
  al., 1987)
• -- lower task performance (Brown et al., 2001)
• -- higher stress and lower well-being
  (Nacoste, 1985)
• effects moderated by positive feedback about
  performance
• (Heilman et al., 1990 Heilman Alcott, 2001)
• effect not replicated outside lab in surveys of
  full-time employees
• (Taylor, 1994 Gallup, 1995)

Iyer (in press)
19
problems with AA

• representation of target group members still not
  equivalent to representation in population
• 2. members of target groups feel stigmatized by
  (presumed) association with AA
• 3. negative perceptions of (presumed)
  beneficiaries
• experimental research suggests that individuals
  selected on the basis of group membership are
  perceived as less competent than those selected
  on the basis of merit alone
• (Dietz-Uhler Murrell, 1998 Heilman Welle,
  2006 Summers, 1991)
• effects moderated by positive information about
  performance
• (e.g., Heilman et al., 1990 Heilman Alcott,
  2001)
• in absence of information, default response is to
  presume low competence

Iyer (in press)
20
improving the effectiveness of AA

• Problems with AA are not insurmountable
• incomplete or inconsistent implementation
• educate employees about discrimination
  importance of AA
• devote more resources to enforcement and
  oversight
• 2. self-views and performance of target group
  members
• provide positive feedback about performance
• establish support networks and programs
• 3. negative views of target group members
• provide positive information about all employees
  performance
• make institutional support for AA explicit and
  ubiquitous

Iyer (in press)