Content uploaded by Boris Baltes
Author content
All content in this area was uploaded by Boris Baltes on Oct 09, 2017
Content may be subject to copyright.
A meta-analysis of empirical studies of weight-based bias
in the workplace
q
Cort W. Rudolph
*
, Charles L. Wells, Marcus D. Weller, Boris B. Baltes
Department of Psychology, Wayne State University, 5057 Woodward Avenue, 7th Floor, Detroit, MI 48230, USA
article info
Article history:
Received 25 July 2008
Available online 8 October 2008
Keywords:
Weight-based bias
Bodyweight
Evaluative workplace outcomes
abstract
For nearly 30 years researchers have investigated how bodyweight affects evaluative work-
place outcomes, such as hiring decisions and performance appraisals. Despite this, no
meta-analytic review has been undertaken to quantify the negative impact that body-
weight has on such outcomes. The results of this meta-analytic study suggest that in rela-
tion to non-overweight individuals in the workplace, overweight individuals may be
disadvantaged across evaluative workplace outcomes (d=.52). Further, differences in
magnitude of the effects of weight-based bias were found for hiring (d=.70) and perfor-
mance (d=.23) outcomes.
Ó2008 Elsevier Inc. All rights reserved.
1. Introduction
In terms of organizational research, bias is often operationalized as a significant main effect difference between the eval-
uations of two target individuals who, all else being equal, vary only by some stigmatized quality or characteristic extrane-
ous to their qualifications or job performance. This operationalization is commonly used in experimental studies in which
some stigmatized quality or characteristic of individuals is varied, while performance or job qualifications are held constant.
In such a paradigm, any differences between the evaluations or judgments of target individuals are said to exist as a function
of the effects of a systematic bias related to the particular stigma under investigation. This main effect operationalization of
bias should not be confused with more recent studies in which attitudes or stereotypes related to a particular stigma are
used to infer bias (e.g. Baltes, Bauer, & Frensch, 2007; Bauer & Baltes, 2002). Research concerning bias operationalized as
main effect differences has investigated these effects by varying different qualities of target individuals, such as race (e.g.,
Greenhaus, Parasuraman, & Wormley, 1990; Landau, 1995; Landy & Farr, 1980; Schmidt & Lappin, 1980), and gender
(e.g., Arvey, 1979; Davison and Burke; 2000; Deaux & Taynor, 1973).
Along these lines, a number of primary studies have demonstrated main effect evidence for bias against overweight indi-
viduals in the workplace (e.g., Bellizzi, Klassen, & Bellonax, 1989; Boridieri, Drehmer & Taylor, 1997; Klesges, et al., 1990;
Zhdanova et al., 2007). Indeed, recent qualitative reviews, (e.g., Puhl & Brownell, 2001; Roehling, 1999; Roehling, 2002), have
suggested that a negative relationship exists between peoples’ bodyweight and a wide range of evaluative workplace out-
comes, citing that overweight individuals are systematically denigrated in comparison to their non-overweight coworkers.
For example, Roehling’s (1999) review of weight-based discrimination in employment settings concluded that evidence for
discrimination against overweight individuals can be found at virtually every stage of the employment process, including
selection (e.g., Klesges et al., 1990), placement (e.g., Bellizzi et al., 1989), compensation (e.g., Register & Williams, 1990), pro-
0001-8791/$ - see front matter Ó2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jvb.2008.09.008
q
Portions of this paper were presented at the 23rd annual meeting of the Society for Industrial and Organizational Psychology, San Franscisco, CA. The
authors would like to gratefully acknowledge the suggestions of Lucy Zhdanova and Malissa Clark on earlier drafts of this manuscript.
* Corresponding author. Fax: +1 313 720 7082.
E-mail address: Cort.Rudolph@wayne.edu (C.W. Rudolph).
Journal of Vocational Behavior 74 (2009) 1–10
Contents lists available at ScienceDirect
Journal of Vocational Behavior
journal homepage: www.elsevier.com/locate/jvb
motion (e.g., Boridieri, Drehmer, & Taylor, 1997), discipline (e.g., Bellizzi & Norvell, 1991), and termination (e.g., Kennedy &
Homant, 1984).
Further research concerning negative attitudes toward overweight people in the workplace has suggested that such indi-
viduals are seen by their coworkers and supervisors as lacking self-discipline and self-control (e.g., Bellizzi & Norvell, 1991;
Klassen, Jasper, & Harris, 1993; Klesges et al., 1990; Larkin & Pines, 1979; Rothblum, Miller, & Garbutt, 1988), being lazy and
not trying as hard as others at work (e.g. Bellizzi & Norvell, 1991; Klassen et al., 1993; Larkin & Pines, 1979; Larwood, 1995),
possessing poor work habits, and having less conscientiousness, competency, skill, and ability than individuals of ‘‘average”
weight (e.g., Klesges et al., 1990; Larkin & Pines, 1979; Larwood, 1995). Moreover, overweight individuals in the workplace
are also viewed as more likely to be absent from work (e.g., Klesges et al., 1990), and less likely to get along with, and be
accepted by their coworkers and subordinates (e.g., Bordieri et al., 1997; Klesges et al., 1990). These issues associated with
the impact of bodyweight in the workplace are particularly relevant given recent epidemiological research that suggests that
the rate of obesity among adults in the United States has been increasing over the last 20 years (CDC., 2007). Further, Puhl
and Brownell (2001) called for more research to be conducted investigating discrimination against overweight individuals
across a variety of evaluative workplace outcomes.
With regard to the impact of weight-based bias at a societal level, Allon (1982) suggests that a strong ‘‘anti-fat” sen-
timent exists in the United States. Indeed, researchers in several domains have investigated the nature of prejudice
against overweight individuals. For example, several studies (e.g., Crandal, 1994) have focused on identifying the nega-
tive characteristics that people associate with overweight individuals. Research of this type suggests that overweight
people are seen as unattractive (Harris, Harris, & Bochner, 1982), aesthetically displeasing (Wooley & Wooley, 1979),
morally and emotionally impaired (Keys, 1955), alienated from their sexuality (Millman, 1980), of a lower socio–eco-
nomic status (Sobal & Stunkard, 1989), and generally unhappy with themselves as people (Maddox, Back, & Liederman,
1968).
1.1. The present study
While several qualitative reviews have examined the effects of bias against overweight individuals in the workplace (e.g.,
Puhl & Brownell, 2001; Roehling, 1999; Roehling, 2002), and nearly 30 years of empirical research has been conducted inves-
tigating this phenomenon (e.g., Bellizzi et al., 1989; Larkin & Pines, 1979; Pingitore, Dugoni, Tindale, & Spring, 1994; Zhda-
nova et al., 2007), no quantitative meta-analytic review has been undertaken to assess the magnitude of the effect of weight-
based bias in empirical investigations. Thus, the present study aims to accomplish two goals.
The first goal of the current study is to conduct a meta-analysis of the extant literature concerning the effects of weight-
based bias across various evaluative workplace outcomes. The results of such a meta-analysis will allow for the integration of
past research with more recent research (e.g., Sartore & Cunningham, 2007; Shapiro, King, & Quinones, 2007) that has inves-
tigated these effects. The second goal of the current study is to test moderators of the weight-based bias—evaluative work-
place outcome relationship. Specifically, this investigation will attempt to determine if job type moderates this relationship
as suggested by Pingitore et al., 1994, and if the effects of weight-based bias vary as a function of outcome type (e.g., hiring
decision, performance evaluation, and promotion decisions). With these goals in mind, we now turn to a discussion of our
hypotheses.
1.1.1. Overall effect
As suggested above, Roehling (1999) concludes that evidence for discrimination against overweight individuals can
be found at virtually every stage of the employment process. Thus, we would expect an overall negative effect of
weight-based bias across evaluative workplace outcomes. Consistent with this conclusion, the following hypothesis
was tested:
Hypothesis 1. There is an overall negative effect of weight-based bias across evaluative workplace outcomes.
1.1.2. Moderation by job type
Several studies have examined whether the effects of weight-based bias are different for different job types, but with
varying results (e.g., Pingitore et al., 1994; Rothblum et al., 1988). For example, Rothblum et al. (1988), present evidence that
the effects of weight-based bias are stronger for jobs that require extensive public contact, such as sales positions. Further,
Pingitore et al. (1994) suggest that this effect may be due to a desire among organizations to maintain the physical appear-
ance of their employees who hold jobs with high public contact, such as sales positions, and that this effect should not be
found for positions with less public contact, such as managerial positions. However, Pingitore et al. did not detect a signif-
icant interaction between bodyweight and job type to support this hypothesis. Despite this, we believe that using more sta-
ble meta-analytic estimates may capture a better picture of this relationship. Thus, in line with other studies that have
looked at this relationship (e.g., Rothblum et al., 1988) it should be expected that jobs that require less extensive public con-
tact, such as managerial positions, would experience the effects of weight-based bias less so than sales positions. It should be
noted here that for consistency, the distinction between ‘‘high public contact” and ‘‘low public contact” positions as ‘‘man-
agerial,” and ‘‘sales” respectively, are based of the on the operationalization of this phenomenon by Pingitore et al. (1994).
Thus, we hypothesized:
2C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10
Hypothesis 2. Job type moderates the effects of weight-based bias on evaluative workplace outcomes, such that the effects
of weight-based bias are stronger for sales positions than for managerial positions.
1.1.3. Moderation by evaluative workplace outcome
Conceptually, if an overall negative relationship is found for weight-based bias across evaluative workplace outcomes (as
suggested by hypothesis 1), it makes sense to test whether or not the type of evaluative workplace outcome under investi-
gation may affect the impact of bias on a given type of evaluation. While we do not make any a priori hypotheses about the
pattern of this moderation, we can infer some clues of this relationship from past research. Specifically, research in organi-
zational and social psychology has long demonstrated that the amount of job relevant information that is available to deci-
sion makers when making an evaluative workplace decision can have a profound impact on the accuracy of that decision
(e.g., Randal & Owens, 1988; Tosi & Einbender, 1985). Research in this area suggests that the effects of bias are less profound
as more job-relevant or performance related information is made available to decision makers. Further, research has consis-
tently shown that stereotypes influence judgments most when limited information about target individuals is available to
raters (e.g., Locksley, Borgida, Brekke, & Hepburn, 1980; Locksley, Hepburn, & Ortiz, 1982). In addition, research has indicated
that individuals place less emphasis on stereotypes when specific information that is relevant to the judgments they are
making is made available (Fiske & Taylor, 1991).
Without going too far to infer the amount of performance relevant information available to raters, one can still infer that
you would expect weight-based bias to impact evaluative workplace outcomes that are associated with lower amounts of
specific performance relevant information (i.e., hiring decisions) more so than evaluative outcomes that may be associated
with more access to performance relevant information (i.e., performance evaluations or promotion decisions). Thus, for the
current investigation, because no explicit hypotheses are made about this relationship, the following research question was
investigated.
Research question 1. Does the type of evaluative workplace outcome investigated moderate the strength of the weight-
based bias-evaluative workplace outcome relationship? Specifically, are there differences in the effect of weight-based bias
for hiring outcomes, performance outcomes, or promotion outcomes?
2. Method
2.1. Sample of studies
Two procedures were used to gather data for the current meta-analysis. First, a series of searches were conducted in
the following digital databases for the periods noted: PsychINFO (Psychological Abstracts; 1967–2007), Proquest (Inter-
disciplinary Dissertations and Theses; all dates) and ERIC (Educational Resources Information Center; 1966–2007). These
searches were conducted using the keywords obese,obesity,overweight, and fat, combined with such keywords as selec-
tion,evaluation,promotion,workplace,managerial,applicant, and performance evaluation. Second, a ‘‘snowballing” tech-
nique was used, which identified relevant articles in the reference lists of primary studies and review articles
concerning bodyweight in the workplace. A total of 59 studies were identified for potential inclusion in this meta-anal-
ysis by using both these search methods.
2.2. Inclusion criteria
Two decision rules were used in order for studies to be included in this meta-analysis: (a) information about the weight of
the target had to be a manipulated variable where there was at least one overweight target and a comparison group of non-
overweight targets; and (b) at least one or more of the dependent variables used had to be a rating of the target on an eval-
uative workplace outcome (e.g., hiring decision, promotion decision, predicted success, suitability, or performance evalua-
tion). Applying these criteria, several primary studies that had been initially considered were excluded from further
analysis (e.g., Frieze, Olson, & Good, 1990). Further, studies that focused on other workplace outcomes such as wage differ-
ences (e.g., Maranto & Stenoien, 2000) were eliminated, because they did not fit with the present conception of evaluative
workplace outcomes. As a result of using the aforementioned selection criteria, the initial pool of 59 primary studies was
reduced to 25.
2.3. Coding procedure
For each primary study, the type of evaluative workplace outcome (i.e., hiring decision, promotion decision, predicted
success, suitability, or performance evaluation) and job type (managerial, sales, or other) was coded. It should be noted that
initially all job types were coded, but due to a lack of studies examining similar job types, this variable had to be collapsed to
include an ‘‘other” category. Outcomes that involved a decision regarding selecting potential employees for hiring were
classified as hiring decisions. Similarly, outcomes that involved a decision regarding whether or not to promote a target
individual were classified as promotion decisions. Outcomes that involved a decision regarding a target’s potential for suc-
cess in a particular position were classified as predicted success, whereas outcomes in which people were asked to make an
C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10 3
evaluation of a target individual’s appropriateness or aptitude for a position were classified as suitability outcomes. Finally,
outcomes in which people were asked to evaluate the job performance of a target individual were classified as performance
evaluations.
In order to calculate the reliability of coder accuracy, all authors coded each study separately. An intra-class correlation
coefficient was calculated for both the outcome type variable and job type variable that were coded, to serve as an index of
inter-rater agreement. Coders showed high agreement, evidenced by a high inter-rater reliability estimate for both the
outcome type (r
icc
= .94), and job type (r
icc
= .89) variables. In the event of disagreement, a resolution was reached via dis-
cussion. It should be noted here that while moderation by outcome type and job type were tested by the present study,
other potential moderators were identified during the initial coding process (i.e., target sex, target BMI). Analysis of these
potential moderating variables was not possible because few studies provided the information needed to test such
moderators.
2.4. Computation of effect size estimates
The meta-analytic strategy employed by the current study was that of Hedges and Olkin (1985). Although other
meta-analytic methods exist (e.g., Hunter & Schmidt, 1990) Hedges and Olkin’s was specifically chosen because among
the experimental studies included in the current analysis, there were few potential issues associated with predictor
unreliability. It should be noted, however, that issues associated with criterion unreliability (i.e., the measurement of
the particular evaluative workplace outcome) could present an issue. It should be noted that only three studies included
in the present meta-analysis reported reliability information for the evaluative workplace outcomes measured (Klesges
et al., 1990; Sartore & Cunningham, 2007; Shapiro et al., 2007). Further, many studies included in the current study re-
lied on one-item measures to assess evaluative workplace outcomes, such as hiring decisions (e.g., Pingitore et al., 1994).
Thus, for the purposes of this study, applying corrections for criterion unreliability was not feasible. Further, there was
no reason to believe that range restriction was attenuating any of the relationships that were coded for in this meta-
analysis. Thus, applying corrections for statistical artifacts, as outlined by Hunter and Schmidt (1990) was not possible
for a majority of cases, and indeed probably not necessary for the current meta-analysis.
As a first step, all statistics reported in the studies were converted into a common statistic using Johnson’s (1993)
DSTAT computer program. The next step was to compute g, a standardized effect size estimate (Hedges & Olkin,
1985). For this study, gis the standardized mean difference between the overweight targets and average-weight targets
on the evaluative workplace outcome of interest. These differences were then divided by the relevant denominator used
to calculate the effect size estimate. For any given primary study, this denominator was either (a) the standard deviation
of the differences when bodyweight was a within-subjects variable, or (b) the pooled standard deviation when body-
weight was a between-subjects variable. The sign of the difference between means was negative when non-overweight
targets were favored over overweight targets and positive when overweight targets were favored over non-overweight
targets on the specified evaluative workplace outcome.
Hedges and Olkin (1985) suggest that the gindex tends to overestimate the magnitude of the population effect size,
especially when samples are small. To avoid this, the effect size estimates that were extracted from the primary stud-
ies were d-statistics. These d-statistics were then combined to estimate both an unweighted mean effect size, and
sample size weighted mean effect sizes. In addition, Q, a homogeneity statistic (Hedges & Olkin, 1985), was calculated
in order to determine whether each set of d-statistics shared a common population effect size, which indicates
whether or not the effect size estimates are homogeneous across the studies. The Q-statistic has a distribution that
approximates that of a chi-square, with (k1) degrees of freedom, where kis the number of effect size estimates
(Hedges & Olkin, 1985).
In order to interpret the results, the magnitude of the effect sizes (d) was based on the suggestions of Cohen (1988). Spe-
cifically, effect sizes (d) of .20 or less were considered small (corresponding to a weighted uncorrected rof .10), .50 were
considered to be a medium effect size (corresponding to a weighted uncorrected rof .25), and .80 or higher were considered
a large effect size (corresponding to a weighted uncorrected rof .40).
An effort was made to extract as much information as possible from each primary study. However, not all of the studies
initially identified for inclusion reported the statistics needed to calculate effect size estimates. Therefore, from the 25 empir-
ical studies, 33 unique samples were identified, and 42 effect size estimates were computed.
3. Results
3.1. Evaluative workplace outcomes
The 42 effect size estimates that were derived from the 25 studies used in this meta-analysis are listed in Table 1. The
overall weighted mean effect size was d=.52 (see Table 2). Since the 95% confidence interval (.56 to .48) that surrounds
this value does not include the value of 0.00, the overall effect size is significant for weight-based bias across all relevant
evaluative workplace outcomes. Based on this, Hypothesis 1 was supported, indicating a significant overall negative effect
of weight-based bias across evaluative workplace outcomes. The Q-statistic based on the 42 effect size estimates was
4C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10
significant, Q
b
(41) = 575.93, p< .01, indicating that these estimates are not homogenous, which suggests the presence of
moderators.
Table 1
Study characteristics and effect sizes for weight-based bias and evaluative workplace outcomes.
Study Year d
i
for Evaluative workplace outcome 95% C.I. for d
i
HD PR PS ST PE Lower Upper r
Larkin & Pines 1979 0.38 0.74 0.02 0.19
*
Polinko & Popovich 2001 0.16 0.10 0.42 0.08
Ding & Stillman 2005 0.30 0.04 0.56 0.15
*
Bellizzi & Hasty 2001 0.18 0.38 0.02 0.09
Bellizzi & Hasty 1998 0.73 0.88 0.58 0.34
****
Bellizzi & Norvell 1991 0.21 0.34 0.08 0.10
**
Bellizzi, Klassen, & Belonax 1989 0.59 0.90 0.27 0.28
***
Bellizzi & Hasty 2000 0.88 1.09 0.67 0.40
****
Pingitore et al. 1994 1.44 1.69 1.20 0.59
****
Rothblum, Miller, & Garbutt 1988 0.29 0.68 0.10 0.14
Lennon 1992 0.28 0.52 0.03 0.14
*
Bordieri, Drehmer, & Taylor 1997 0.51 1.12 0.11 0.25
Bordieri, Drehmer, & Taylor 1997 0.47 1.09 0.14 0.23
Bordieri, Drehmer, & Taylor 1997 0.42 1.03 0.19 0.21
Jasper & Klassen 1990a 0.59 0.94 0.25 0.28
***
Brink 1988 0.86 1.31 0.41 0.40
***
Brink 1988 0.79 1.15 0.42 0.37
****
Benson et al. 1980 1.35 2.32 0.38 0.57
**
Benson et al. 1980 1.25 2.21 0.29 0.54
**
Klesges et al. 1990 0.18 0.41 0.04 0.09
Klesges et al. 1990 0.26 0.49 0.03 0.13
*
Sartore & Cunningham 2007 1.46 1.84 1.08 0.59
****
Sartore & Cunningham 2007 1.48 1.86 1.10 0.60
****
Sartore & Cunningham 2007 1.13 1.52 0.74 0.49
****
Sartore & Cunningham 2007 1.24 1.77 0.97 0.57
****
Sartore & Cunningham 2007 1.24 1.68 0.80 0.53
****
Sartore & Cunningham 2007 1.05 1.48 0.62 0.47
****
Shapiro et al. 2007 0.70 1.30 0.10 0.34
*
Shapiro et al. 2007 1.43 2.08 0.77 0.59
****
Kutcher & Bragger 2004 0.40 0.74 0.05 0.20
*
Kutcher & Bragger 2004 0.63 1.13 0.14 0.31
*
Kutcher & Bragger 2004 1.11 1.47 0.75 0.49
****
Mirch-Kretschmann 2004 1.81 2.19 1.44 0.67
****
Alfonso 1997 1.53 1.72 1.33 0.61
****
Alfonso 1997 0.01 0.18 0.16 0.01
Alfonso 1997 0.01 0.16 0.18 0.01
Bevins 2003 0.43 0.80 0.08 0.21
*
Bevins 2003 0.08 0.45 0.29 0.04
Banta 2004 0.85 1.11 0.60 0.39
****
Cates 1999 0.42 0.60 0.24 0.21
****
Cates 1999 0.48 0.68 0.29 0.24
****
Hebl 1997 0.38 0.06 0.70 0.19
*
Year, year of publication; HD, hiring decision; PR, promotion; PS, predicted success; ST, suitability; PE, performance evaluation; d
i
, the mean weighted effect
size estimate; C.I., confidence interval; r, correlation between evaluative workplace outcome and weight-based bias.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
Table 2
Overall effect size estimate (d) across evaluative workplace outcomes.
95% C.I. for d
i
kd
i
Lower Upper rQ
b
Overall 42 0.52 0.56 0.48 0.25 575.93
****
k, the number of effect size estimates in each evaluative workplace outcome type; C.I., confidence interval; d
i
, the mean weighted effect size estimate; r,
correlation between evaluative workplace outcome and weight-based bias.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10 5
3.2. Moderation by job type
Hypothesis 2 suggested that job type would moderate the effects of weight-based bias on evaluative workplace outcomes.
Because only a small number of studies reported job type for performance outcomes, the effect of job type could only be
tested for hiring outcomes. For hiring outcomes, job type did not significantly moderate the effects of weight-based bias,
Q
b
(1) = 2.43, n.s., (see Table 3), suggesting that there was no significant differences between managerial (d=.62), and sales
(d=.72) job types, in terms of hiring outcomes.
3.3. Moderation by evaluative workplace outcome
Research question #1 asked whether or not the effect of weight-based bias would vary for hiring outcomes, performance
outcomes, or promotion outcomes. In order to explore this question, we tested for moderation by outcome type. Table 4
shows the effect of weight-based bias by evaluative workplace outcome. The effect of weight-based bias on hiring outcomes
(d=.76) was not significantly different than the effect of weight-based bias on suitability outcomes (d=.65). Likewise,
the effect for predicted success outcomes (d=.26) was not significantly different than for performance evaluation out-
Table 3
Job type between class effects (Q
b
).
Attributes Outcome Level 95% C.I. for d
i
kd
i
Lower Upper rQ
w
Q
b
Job type
Hiring 2.43
Managerial 7 0.62 0.72 0.53 0.30 147.85
****
Sales 10 0.72 0.81 0.64 0.34 54.02
****
k, the number of effect size estimates in each evaluative workplace outcome type; d
i
, the mean weighted effect size estimate; C.I., confidence interval; r,
correlation between evaluative workplace outcome and weight-based bias; Q
w
, within class effects: significance indicates effects rejection of hypothesis of
homogeneity; Q
b
, between class effects: significance indicates effects differ as a function of outcome type.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
Table 4
Effect size estimates (d) by evaluative workplace outcomes.
Outcome 95% C.I. for d
i
kd
i
Lower Upper rQ
w
Hiring decision 17 0.76 0.84 0.68 0.35 248.47
****
Promotion 4 0.07 0.19 0.04 0.04 15.21
**
Predicted success 3 0.21 0.39 0.02 0.10 3.10
Suitability 14 0.65 0.72 0.58 0.31 146.41
****
Performance evaluation 4 0.24 0.35 0.13 0.12 12.82
**
Model Q
b
(5) = 144.03
****
k, the number of effect size estimates in each evaluative workplace outcome type; C.I., confidence interval; d
i
, the mean weighted effect size estimate; r,
correlation between evaluative workplace outcome and weight-based bias.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
Table 5
Overall effect size estimate (d) for collapsed evaluative workplace outcome.
Outcome 95% C.I. for d
i
kd
i
Lower Upper rQ
w
Hiring 31 0.70 0.75 0.64 0.33 399.25
****
Performance 7 0.23 0.32 0.14 0.12 61.66
*
Overall 38 0.58 0.63 0.54 0.28 489.87
****
Model Q
b
(1) = 73.61
****
k, the number of effect size estimates in each evaluative workplace outcome type; d
i
, the mean weighted effect size estimate; C.I., confidence interval; r,
correlation between evaluative workplace outcome and weight-based bias.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
6C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10
comes (d=.24). Further, the effect size estimate calculated for promotion outcomes was not significant (d=.07). Because
hiring and suitability outcomes were not significantly or qualitatively different from each other, and neither were predicted
success and performance evaluation outcomes, and because both hiring and suitability outcomes were significantly and
qualitatively different from predicted success and performance evaluation outcomes (see Table 4), these four groups were
collapsed into two new outcomes which were labeled hiring outcomes, and performance outcomes (see Table 5).
Interestingly, there were significant differences between hiring outcomes (d=.70) and performance outcomes
(d=.23), with the effect of weight based bias on hiring outcomes being significantly stronger than the effect on perfor-
mance outcomes (see Table 6).
4. Discussion
The primary purpose of this study was to conduct a meta-analytic review of experimental studies that have examined the
relationship between weight-based bias and evaluative workplace outcomes. Based on Cohen’s (1988) effect size classifica-
tion, the magnitudes of the effect sizes calculated here vary by outcome type. For instance, the overall effect for bodyweight
across workplace outcome was medium (see Table 2). However, more variability was found with hiring and performance
outcomes whose effect sizes were large and small, respectively (see Table 4). Based on these estimates, it is clear that there
is an overall medium effect of weight-based bias across the evaluative workplace outcomes studied to date. This finding sup-
ports Roehling’s (1999) review, which suggests that bodyweight has negative implications across a variety of evaluative
workplace outcomes, including hiring, performance, and promotion decisions.
The findings presented here add to the literature on weight-based bias in the workplace by successfully addressing the
previously stated goals: (1) to conduct a meta-analysis of the extant literature concerning the effects of weight-based bias
across various evaluative workplace outcomes, (2) to test moderators of the weight-based bias—evaluative workplace out-
come relationship.
4.1. Evaluative workplace outcomes
Hypothesis one was supported (See Table 2), indicating a significant overall negative effect of weight-based bias across
evaluative workplace outcomes. These results, consistent with previous qualitative reviews (e.g., Puhl & Brownell, 2001;
Roehling, 1999; Roehling, 2002), indicate that overweight individuals are systematically denigrated in relation to their
non-overweight coworkers across evaluative workplace outcomes.
4.2. Moderation by job type
Hypothesis two, that job type would moderate the effects of weight-based bias on evaluative workplace outcomes, was
not supported. There were no significant differences in the level of weight-based bias between managerial and sales posi-
tions for hiring outcomes. These results are contrary to those of prior studies, (e.g., Rothblum et al., 1988) which have sug-
gested that differences exist in the amount of weight-based bias experienced by high public contact versus low public
contact positions. However, these results are similar to those of Pingitore et al. (1994) who also did not find differences
in hiring outcomes between managerial and sales positions for overweight applicants. It should be noted, however, that
there appears to be a trend that supports this hypothesis, specifically that the effects of weight-based bias on hiring were
stronger for sales positions (d=.72) than for managerial positions (d=.62), even though these differences were not sta-
tistically significant. Perhaps with further primary investigation into the relationship between job type and level of weight-
based bias, this question could be answered more definitively. Indeed, this highlights a discrepancy in the literature that Puhl
and Brownell (2001) have also identified, specifically that more research needs to be conducted regarding the differences in
how overweight individuals are treated in the workplace. Such research may be conducted by varying job types and levels of
public contact for overweight and average weight targets, and identifying differences between these two target groups.
Table 6
Moderation by evaluative workplace outcome.
Outcome 95% C.I. for d
i
kd
i
Lower Upper rQ
w
Hiring 31 0.70 0.75 0.64 0.33 399.25
****
Performance 7 0.23 0.32 0.14 0.12 17.00
*
Promotion 4 0.07 0.19 0.04 0.04 15.21
**
Overall 42 0.47 0.504 0.4283 0.230 624.04
****
Model Q
b
(5) = 173.66
****
k, the number of effect size estimates in each evaluative workplace outcome type; C.I., Confidence Interval; d
i
, the mean weighted effect size estimate; r,
correlation between evaluative workplace outcome and weight-based bias.
*
p< .05.
**
p< .01.
***
p< .001.
****
p< .0001.
C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10 7
4.3. Moderation by evaluative workplace outcome
Investigating research question #1 led to some interesting conclusions, namely that the effect of weight-based bias was
found to be strongest for hiring outcomes, less so for performance outcomes, and least for promotion outcomes. These results
may suggest a diminishing impact of weight-based bias across such outcomes. However, while these results are compelling,
more research is needed to establish the parameters under which this diminishing impact may occurs, specifically these re-
sults are based on meta-analytic estimate.
4.4. Limitations
One limitation of the current investigation is that the results were based on only 25 studies. Every attempt was made to
identify all published and unpublished experimental studies meeting the pre-established criteria for inclusion, and to con-
tact authors and experts in this area of study for additional data. While seemingly limited, the 25 studies used in this meta-
analysis represent a very representative picture of the extant literature concerning the effects of weight-based bias on eval-
uative workplace outcomes.
Another potential limitation of this study is that the studies included in the meta-analysis were all laboratory based,
experimental designs. Indeed, the rationale behind conducting a meta-analysis of such experimental studies was based par-
tially on the fact that only a handful of field studies have been conducted examining the relationships under investigation
here. However, it is interesting to note that the many of field studies that have been conducted (e.g., Falkner et al., 1999;
Harris, Waschull, & Walters, 1990; Rothblum, Brand, Miller, & Oetjen, 1990) seem to align with the conclusions of the current
study; overweight individuals are systematically disadvantages in the workplace, across a variety of evaluative outcomes.
Indeed, it would be interesting to see whether results from a meta-analysis using field or non-experimental studies would
demonstrate similar results. Indeed, future research should attempt to quantify the extant non-experimental literature that
has investigated the effects of weight-based bias on workplace outcomes. However, it should be noted that this is not a lim-
itation if one is generalizing the results of this meta-analysis to future laboratory based, experimental investigations.
It is fair to assume that the participants used for the studies on which these results are based were not making critical
decisions, and were thus not accountable for the outcomes of their decisions. Research has indicated that decision-making
can be impacted if people are held accountable for their decisions (e.g., Simonson & Nye, 1992). Because participants in the
studies used for this meta-analysis were primarily from laboratory studies, one could argue that their actions are governed
by a different set of demand characteristics than participants in field studies, who may be held more accountable for their
actions and decisions.
Along these same lines, an additional potential limitation of this study is that a majority of participants in the studies
included in this meta-analysis did not have contact with those whom they evaluated. Indeed the contrived nature of labo-
ratory research concerning evaluative workplace outcomes makes it difficult to determine whether or not these effects
would have similar magnitude and directionality in actual workplace settings, or if they are artificially affected by the nature
of the laboratory environment. Perhaps in actual workplace settings, where individuals interact, the effects of weight-based
bias would be differently affected by such contact. Although it would be interesting to see whether the results of this meta-
analysis would be different if more studies used organizational samples, the scope of the literature at this point in time pre-
vents such an analysis. Indeed, future research efforts should attempt to address the issue of weight-based bias in applied
workplace settings to determine if the strength of the relationships reported here would transfer to an organizational setting.
Despite the limitations noted above, the present effort represents the best attempt to date to quantitatively synthesize
the extant literature concerning the effects of weight-based bias on evaluative workplace outcomes. This paper stands to
demonstrate, at its best, an aggregated estimate of the average effect of weight-based bias across evaluative workplace out-
comes, suggesting that weight-based bias is likely to occur across a variety of evaluative decision making scenarios, partic-
ularly for hiring and performance decisions. By no means should this investigation be considered the final word concerning
this relationship. Indeed, the results presented here should be viewed as a call for more careful and directed research inves-
tigating the nature of the relationship between weight-based bias and evaluative workplace outcomes.
References
*Alfonso, I. (1997). The effects of weight on employment decisions in a variety of job settings. Dissertation Abstracts International, 58(06), 3364. UMI No.
9738104.
Allon, N. (1982). The stigma of overweight in everyday life. In B. Wolman (Ed.), Psychological aspects of obesity: a handbook (pp. 130–174). New York: Van
Nostren Reinhold.
Arvey, R. D. (1979). Unfair discrimination in the employment interview: Legal and psychological aspects. Psychological Bulletin, 86, 736–765.
Baltes, B. B., Bauer, C. C., & Frensch, P. A. (2007). Does a structured free recall intervention reduce bias in performance ratings and by what mechanism?
Journal of Applied Psychology, 92, 151–164.
*Banta, J. A. (2004). The impact of weight, race, gender, and context on target selection and attributions. Dissertation Abstracts International, 64(07), 3577.
UMI No. 3098645.
Bauer, C. C., & Baltes, B. B. (2002). Reducing the effects of gender stereotypes on performance evaluations. Sex Roles, 47, 465–476.
*Bellizi, J. A., & Hasty, R. W. (2000). Does successful work experience mitigate weight and gender based employment discrimination in face-to-face
industrial selling? Journal of Business and Industrial Marketing, 15, 384–398.
*Bellizzi, J. A., & Hasty, R. W. (1998). Territory assignment decisions and supervising unethical selling behavior: The effects of obesity and gender as
moderated by job related factors. Journal of Personal Selling and Sales Management, 2, 35–49.
8C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10
*Bellizzi, J. A., & Hasty, R. W. (2001). The effects of a stated organizational policy on inconsistent disciplinary action based on salesperson gender and weight.
The Journal of Personal Selling and Sales Management, 3, 189–198.
*Bellizzi, J. A., Klassen, M. L., & Belonax, J. J. (1989). Stereotypical beliefs about overweight and smoking and decision-making in assignments to sales
territories. Perceptual and Motor Skills, 69, 419–429.
*Bellizzi, J. A., & Norvell, D. W. (1991). Personal characteristics and salesperson’s justifications as moderators of supervisory discipline in cases involving
unethical salesforce behavior. Journal of the Academy of Marketing Science, 19, 11–16.
*Benson, P. L., Severs, D., Tatgenhorst, N., & Loddengaard, N. (1980). The social costs of obesity: A non-reactive field study. Social Behavior and Personality, 8,
91–96.
*Bevins, B. C. (2003). Employability of individuals with varying disabilities and costs of needed workplace accommodations. Masters Abstracts International,
41(05), 1524–1555. UMI No. 1413413.
Bordieri, J. E., Drehmer, D. E., & Taylor, D. W. (1997). Work life for employees with disabilities: Recommendations for promotion. Rehabilitation Counseling
Bulletin, 40, 181–191.
*Brink, T. L. (1988). Obesity and job discrimination: Mediation via personality stereotypes? Perceptual and Motor Skills, 66, 494.
*Cates, S. V. (1999). Do human resource professional practice the self-fulfilling prophecy: An examination of the influence of applicant gender and applicant
weight in the context of an employment selection decision. Dissertation Abstract International, 60(05), 1656. UMI No. 1413413.
CDC. (2007). US Obesity Trends 1985–2005. In Centers for Disease Control and Prevention. Retrieved July 21, 2007, Available from: <http://www.cdc.gov/
nccdphp/dnpa/obesity/trend/maps/>.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press.
Crandal, C. S. (1994). Prejudice against fat people: Ideology and self interest. Journal of Personality and Social Psychology, 66, 882–894.
Davison, H. K., & Burke, M. J. (2000). Sex discrimination in simulated employment contexts: A Meta-analytic investigation. Journal of Vocational Behavior, 56,
225–248.
Deaux, K., & Taynor, J. (1973). Evaluation of male and female ability: Bias works two ways. Psychological Reports, 32, 261–262.
*Ding, V. J., & Stillman, J. A. (2005). An empirical investigation of discrimination against overweight female job applicants in New Zealand. New Zealand
Journal of Psychology, 34, 139–148.
Falkner, N. H., French, S. A., Jefferey, R. W., Neumark-Sztainer, D., Sherwood, N. E., & Morton, N. (1999). Mistreatment due to weight: Prevelance and sources
of perceived mistreatment of women and men. Obesity Research, 7, 572–576.
Fiske, S. T., & Taylor, S. E. (1991). A continuum of impression formation, from category based to individuating processes: Influences of information and
motivation on attention and interpretation. In M. P. Zanna (Ed.). Advances in experimental social psychology (Vol. 23, pp. 1–74). New York: Academic
Press.
Frieze, I. H., Olson, J. E., & Good, D. C. (1990). Perceived and actual discrimination in the salaries of male and female managers. Journal of Applied Social
Psychology, 20, 46–67.
Greenhaus, J. H., Parasuraman, S., & Wormley, W. M. (1990). Effects of race on organizational experiences, job performance evaluations, and career
outcomes. Academy of Management Journal, 33, 64–88.
Harris, M. B., Harris, R. J., & Bochner, S. (1982). Fat, four-eyed and female: Stereotypes of obesity, glasses, and gender. Journal of Applied Social Psychology, 6,
503–516.
Harris, M. B., Waschull, S., & Walters, L. C. (1990). Feeling fat: Knowledge and attitudes of overweight men and women. Psychological Reports, 67, 1191–1202.
*Hebl, M. R. (1997). Nonstigmatized individuals’ reactions to the acknowledgment and valuation of a stigma by overweight individuals and physically
disabled individuals. Dissertation Abstracts International, 58(06), 3367. UMI No. 9735860.
Hedges, L. V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando: Academic Press.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias in research findings. Newbury Park, CA: Sage.
*Jasper, C. A., & Klassen, M. L. (1990). Perceptions of salespersons appearance and evaluation of job performance. Perceptual &Motor Skills, 71, 563–566.
Johnson, B. T. (1993). DSTAT 1.10: Software for the meta-analytic review of research literatures. Hillsdale, NJ: Erlbaum.
Kennedy, D. B., & Homant, R. J. (1984). Personnel managers and the stigmatized employee. Journal of Employee Counseling, 21, 89–94.
Keys, A. (1955). Obesity and heart disease. Journal of Chronic Diseases, 1, 456–460.
Klassen, M. L., Jasper, C. R., & Harris, R. J. (1993). The role of physical appearance in managerial decisions. Journal of Business and Psychology, 8, 181–198.
Klesges, R. C., Klem, M. L., Hanson, C. L., Eck, L. H., Ernst, J., O’Laughlin, D., et al (1990). The effects of applicant’s health status and qualifications on simulated
hiring decisions. International Journal of Obesity, 14, 527–535.
*Kutcher, E. J., & Bragger, J. D. (2004). Selection interviews of overweight job applicants: Can structure reduce the bias? Journal of Applied Social Psychology,
34, 1993–2022.
Landau, J. (1995). The relationship of race and gender to managers’ ratings of promotion potential. Journal of Organizational Behavior, 16, 391–400.
Landy, F. J., & Farr, J. L. (1980). Performance Ratings. Psychological Bulletin, 16, 72–107.
*Larkin, J. C., & Pines, H. A. (1979). No fat persons need apply: Experimental studies of the overweight stereotype and hiring preferences. Sociology of Work
and Occupations, 6, 312–327.
Larwood, L. (1995). Rational bias and interorganizational power in the employment of management consultants. Group and Organizational Studies, 10, 3–17.
*Lennon, S. J. (1992). Categorization as a function of body type. Clothing and Textile Research Journal, 10, 18–23.
Locksley, A., Borgida, E., Brekke, N., & Hepburn, C. (1980). Sex stereotypes and social judgment. Journal of Personality and Social Psychology, 39,
821–831.
Locksley, A., Hepburn, C., & Ortiz, V. (1982). Social stereotypes and judgments of individuals: An instance of the base-rate fallacy. Journal of Experimental
Social Psychology, 18, 23–42.
Maddox, G. L., Back, K. W., & Liederman, V. R. (1968). Overweight as a social deviance and disability. Journal of Health and Social Behavior, 9, 287–298.
Maranto, C. L., & Stenoien, A. F. (2000). Weight discrimination: A multidisciplinary analysis. Employee Responsibilities and Rights Journal, 12, 9–24.
Millman, M. (1980). Such a pretty face: Being fat in America. New York: Norton.
*Mirch-Kretschmann, S. E. (2004). Framing messages aimed at reducing discrimination and increasing diversity: The role of certainty, affect, and stigma.
Dissertation Abstracts International, 65(03), 1598. UMI No. 3125264.
Pingitore, R., Dugoni, B. L., Tindale, R. S., & Spring, B. (1994). Bias against overweight job applicants in a simulated employment interview. Journal of Applied
Psychology, 79, 909–917.
*Polinko, N. K., & Popovich, P. M. (2001). Evil thoughts but angelic actions: Responses to overweight job applicants. Journal of Applied Social Psychology, 31,
905–924.
Puhl, R., & Brownell, K. D. (2001). Bias, discrimination, and obesity. Obesity Research, 12, 788–805.
Randal, G. A., & Owens, S. D. (1988). The effect of job level and amount of information on the evaluation of male and female job applicants. Journal of
Employee Counseling, 25, 160–171.
Register, C. A., & Williams, D. R. (1990). Wage effects of obesity among young workers. Social Science Quarterly, 71, 130–141.
Roehling, M. V. (1999). Weight based discrimination in employment: Psychological and legal aspects. Personnel Psychology, 52, 969–1016.
Roehling, M. V. (2002). Weight discrimination in the American workplace. ethical issues and analysis. Journal of Business Ethics, 40, 177–189.
Rothblum, E. D., Brand, P. A., Miller, C. T., & Oetjen, H. A. (1990). The relationship between obesity, employment discrimination, and employee related
victimization. Journal of Vocational Behavior, 37, 251–266.
*Rothblum, E. D., Miller, C. T., & Garbutt, B. (1988). Stereotypes of obese female job applicants. International Journal of Eating Disorders, 7, 277–283.
*Sartore, M. L., & Cunningham, G. B. (2007). Weight discrimination, hiring recommendations, person-job fit, and attributions: Fitness-Industry Implications.
Journal of Sport Management, 21, 172–193.
C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10 9
Schmidt, N., & Lappin, M. (1980). Race and sex as determinants of the mean and variance of performance ratings. Journal of Applied Psychology, 65, 428–435.
*Shapiro, J. R., King, E. B., & Quinones, M. A. (2007). Expectations of obese trainees: How stigmatized trainee characteristics influence training effectiveness.
Journal of Applied Psychology, 92, 239–249.
Simonson, I., & Nye, P. (1992). The effect of accountability on susceptibility to decision errors. Organizational Behavior and Human Decision Processes, 51,
416–446.
Sobal, J., & Stunkard, A. J. (1989). Socioeconomic status and obesity: A review of the literature. Psychological Bulletin, 105, 260–275.
Tosi, H. L., & Einbender, S. W. (1985). The effect of the type and amount of information in sex discrimination research: a meta-analysis. Academy of
Management Journal, 28, 712–723.
Wooley, S. C., & Wooley, O. W. (1979). Obesity and women—I: A closer look at the facts. Women’s Studies International Quarterly, 2, 69–79.
Zhdanova, L., Baltes, B. B., Chakrabarti M., Finkelstein, L., Roehling, M. V., & Sheppard, A.K. (2007). Examining and Reducing Rater Stereotype Affects on
Performance Ratings. Poster presented at The Annual Conference of Industrial and Organizational Psychology, New York, New York.
10 C.W. Rudolph et al. / Journal of Vocational Behavior 74 (2009) 1–10
