Beruflich Dokumente
Kultur Dokumente
1,"l02-ffl
The Impact of Ratee's Disability on Performance Judgments and Choice as Partner: The Role of Disability-Job Fit Stereotypes and Interdependence of Rewards
Adrienne Colella and Angelo S. DeNisi
Texas A&M University
Arup Varma
Loyola UniversityChicago
An experiment assessed the impact of disability-job fit stereotypes and reward interdependence on personnel judgments about persons with disabilities. Students (N = 87) evaluated 3 confederates. The experiment varied disability of the target confederate (dyslexia vs. nondisabled), task, and dependence of rater rewards on partner performance. Two disability-task combinations represented stereotypical poor fit and good fit. Dependent variables were performance evaluations, performance expectations, and ranking of target as a partner. There was negative bias against the confederate with dyslexia in poor-fit conditions. In the interdependent reward condition, there was a negative main effect for disability, regardless of fit. No effects for disability were found on performance ratings or expectations. Results indicate the need to consider disability-job fit stereotypes and consequences to raters when assessing the impact of disability on personnel judgments.
Several factors have drawn both organizations' and scholars' attention to assessing the treatment of persons with disabilities at work. These include the Americans with Disabilities Act (1990), growing political activism by persons with disabilities (Scotch, 1988), and the recent focus on managing a diverse employee population (e.g., Cox, 1993). Thus, recently we have seen more concern over personnel decisions about persons with disabilities. Yet there is little empirical research that actually examines the effect of disability on personnel decisions and judgments. Reviews of this limited research point to conflicting and inconsistent results in terms of bias (cf. Colella, 1996; D. L. Stone & Colella, 1996; E. F. Stone, Stone, & Dipboye, 1992), where bias refers to differences in evaluations and judgments due to disability rather than to actual performance. It is important to determine whether disability leads to bias in personnel decisions and judgments because, apart from presenting a case for illegal discrimination, biased
Adrienne Colella and Angelo S. DeNisi, Department of Management, Lowry Mays College and Graduate School of Business, Texas A&M University. Arup Varma, Institute of Human Resources and Industrial Relations, Loyola University Chicago. The research was funded by a grant from the New Jersey Developmental Disabilities Council. Correspondence concerning this article should be addressed to Adrienne Colella, Department of Management, Texas A&M University, College Station, Texas 77843. Electronic mail may be sent to acolella@cgsb.tamu.edu.
102
judgments may set in motion a negative performance cycle where employees respond to negative expectations in the form of serf-fulfilling prophecies (Eden, 1990), come to distrust the organization, and fail to become fully socialized into the organization (cf. Colella, 1996). Given that employees with disabilities often face poorer prospects for job advancement and high salaries compared with nondisabled employees (Braddock & Bachelder, 1994), examining how personnel judgments may be influenced when a person has a disability takes on added importance. This study addresses some of the confusion in past research by examining the effects of disability on personnel judgments (performance evaluations, performance expectations, and choice as a team partner) as a function of stereotypes about disability-job fit and the interdependence of rewards. There are several theoretical reasons why bias, both negative and positive, may occur in the performance evaluation of persons with disabilities (see Colella, DeNisi, & Varma, 1997, and D. L. Stone & Colella, 1996, for reviews), including stereotypes, low performance expectations, and the "norm to be kind." Reviews of empirical literature (Colella, 1996; D. L. Stone & Colella, 1996; E. F. Stone et al., 1992) examining the effect of disability on performance ratings have reported mixed findings. Research that involves surveying employers about disabled employees' performance does not suggest negative bias in assessments of performance per se (Braddock & Bachelder, 1994). Many of these surveys find that managers report that the majority of employees with disabilities perform their jobs well (e.g., Louis Harris and Associates,
RESEARCH
REPORTS
103
1987). This research makes it unclear as to what degree bias exists against (or for) persons with disabilities for several reasons. First, managers in these surveys rarely rate specific employee performance but rather are simply asked to rate how well employees with disabilities (in general) perform in their organizations. Also, these surveys are conducted for research purposes, that is, they are not actual performance evaluations. Second, this research does not control for objective performance, making the determination of bias (or accuracy) of the performance ratings impossible. Third, most of these studies do not contain a control group of nondisabled ratees, again making the assessment of bias difficult. Finally, this research is subject to social desirability bias. What is needed are experimental studies that control for these factors while assessing the degree of bias in performance evaluations of persons with disabilities. To our knowledge there have been only four experimental studies in which performance ratings of ratees with disabilities are compared with ratings of nondisabled ratees, controlling for objective performance levels. These studies have mixed results. Miller (1992) reported no bias toward persons with disabilities, whereas Hastorf, Northcraft, and Picciotto (1979) found negative bias toward persons with disabilities. Two other studies found bias only under certain circumstances. Czajka and DeNisi (1988) found positive bias toward persons with disabilities (i.e., they were rated unjustifiably high) when there were no performance standards. Varma, Colella, and DeNisi (1996) found negative bias, but only under conditions of a poor rater-ratee relationship. These few, inconsistent studies make it impossible to draw conclusions about the performance evaluation of persons with disabilities. There have been many more studies that assess the performance expectations toward persons with disabilities. This research has been clearer, albeit more pessimistic, about the impact of ratees' disabilities. Colella (1996) reviewed research examining performance expectations that persons without disabilities held toward ratees with disabilities and found that when expectations specifically concern future performance or promotional potential, there is little evidence of any positive bias in favor of persons with disabilities. In fact, 10 out of 13 studies that measured performance expectations found evidence of negative bias (e.g., Bordieri & Drehmer, 1986), whereas 3 studies found no effects (e.g., Rose & Brief, 1979). There are many theoretical reasons why employees with disabilities may be unfairly excluded from discretionary work-related activities (see Colella, 1996, for a review). These reasons include affect (i.e., people with disabilities may be disliked), strain (i.e., others may find it difficult to interact with people with disabilities), and low expectations. Empirical research on coworkers' inclusion of per-
sons with disabilities, that is, their willingness to choose to interact with disabled coworkers in discretionary situations, has been extremely scant. However, there is empirical research that indirectly speaks to this issue. Based on early stigma research (e.g., Goffman, 1963), numerous qualitative studies have described the exclusion of persons with disabilities. For example, Knox and Parmenter (1993) in their observation of the treatment of employees with mild intellectual disabilities noted that "subjects were part of the environment, yet only on the periphery. . . . They were 'in' the work setting, but not 'of' it" (p. 7). Surveys of social distance attitudes toward persons with disabilities generally report that respondents desire further distance from persons with disabilities, even in working relationships (Bowman, 1987), although attitudes vary by type of disability. Surveys of employers' attitudes about hiring people with disabilities also suggest that they are less likely to be included and accepted by coworkers (see review by Greenwood, Shiner, & Johnson, 1991; McFarlin, Song, & Sonntag, 1991). Finally, the few experimental studies that examined participants' willingness to interact with a person with a disability (e.g., Farina & Ring, 1965; Snyder, Kleck, Strenta, & Mentzet, 1979; D. L. Stone & Michaels, 1993, 1994) found negative bias. In sum, the results of past research on the effects of disability on performance evaluations and inclusion by coworkers are scant, unclear, or conflicting. We argue that some of this inconsistency can be partially explained by considering two specific factors: raters' stereotypes about disability-job fit and the potential personal consequences to raters deriving from their personnel decisions. Although it has been documented that nondisabled persons hold negative stereotypes and attitudes toward persons with disabilities (e.g., Fichten & Amstel, 1986), researchers have also concluded that stereotypes and attitudes vary as a function of the nature of the disability (cf. Jones et al., 1984), thus making it difficult to generalize from one disability to another. For example, compare stereotypes likely to be made about a person with AIDS with those about a person with chronic back pain. Furthermore, research on attitudes toward persons with disabilities has demonstrated that attitudes are a joint function of the disability type and the social context (Bowman, 1987; Grand, Bernier, & Strohmer, 1982; Yuker, 1983). On the basis of this work, several researchers have argued that rather than just examine the impact of stereotypes associated with different disabilities, research must also consider the nature of the job (Colella, 1996; Miller, 1992; D. L. Stone & Colella, 1996) to determine if there are stereotypes concerning the "fit" between a particular disability and a particular job. This notion is quite similar to Heilman's (1983) lack-of-fit model, which describes sex bias as a function of the discrepancy between the sex
104
RESEARCH
REPORTS
stereotypes applied to a person and the sex stereotypes associated with the job. Although there are stereotypes about "typical male and typical female" jobs (Heilman, 1983), in the case of disabilities, the critical stereotype may be more complex and may be a joint function of the nature of the disability and the nature of the job. For example, although persons with physical disabilities are generally rated more favorably than those with mental or emotional disabilities, they are rated more negatively for jobs where appearance is important (e.g., Sowa & Cutter, 1974). Thus, for such jobs, the stereotypic fit between physical disability and the job may be poor, and so personnel judgments about a person with a physical disability may be negatively biased. On the other hand, stereotypic fit may be good because the disability is not seen as being particularly important for the job, and thus, job-related judgments of the person would not be influenced by the disability. The issue of disability-job fit stereotypes has not received much empirical attention. Past work has either considered only general competency judgments about persons with specific disabilities (e.g., Bowman, 1987; Christman & Slaten, 1991; Greenwood et al., 1991) or compared competency judgments and attitudes of persons in various professions toward persons with disabilities (Schneider & Anderson, 1980). Examining the types of jobs that were involved in some of the studies cited above suggests that performance judgments and evaluations may vary according to the job and the disability. Certain types of disabilities may provide better stereotypic fit with some jobs than with others, and only by considering both the nature of the disability and the nature of the job can we truly come to understand stereotypes held by individuals about persons with disabilities at work. We believe that these fit stereotypes are critical to understanding jobrelated judgments about persons with disabilities. Thus, we posit:
Hypothesis la. Performance evaluations will be a function of disability-task fit stereotypes so that negative bias will be present only when there is a poor disability-job fit stereotype. Hypothesis Ib. Performance expectations will be a function of disability-task fit stereotypes so that negative bias will be present only when there is a poor disabilityjob fit stereotype.
pants may attribute good performance on the part of a disabled person to unstable factors (e.g., luck; Russell et al., 1985), and thus, while making accurate evaluations of past performance, they will not expect this level of performance to continue. Because we predicted that participants would expect persons in poor-fit situations to perform worse than others, we also posit:
Hypothesis Ic. Choice as partner will be a function of disability-task fit stereotypes so that negative bias will be present only when there is a poor disability-job fit stereotype.
Because participants in the present experiment were given nonambiguous objective performance information, and because stereotypes are most likely to be evoked when information is ambiguous (e.g., Dipboye, Fromkin, & Wibak, 1975; Locksley, Borgida, Brekke, & Hepburn, 1980), we expected the impact of stereotypic fit to be weaker on evaluations of past performance than on expectations about future performance. Research suggests that partici-
It should be noted that these hypotheses all predict a significant interaction effect between disability and task rather than a main effect for disability. In order to clarify Hypotheses la through Ic, two terms need to be defined. In the present study, we assess stereotypic fit in terms of how well people in general believe persons with disabilities can perform various jobs. Disability research has demonstrated that people hold a variety of different stereotypes about persons with disabilities (see D. L. Stone & Colella, 1996) and that one type of stereotype associated with categorization into various disability groups is how well an individual with a specific disability can perform various jobs (e.g., Bowman, 1987; Colella, DeNisi, Varma, & Lund, 1994). It should also be noted that sometimes these stereotypes may be based on "truth," that is, one's disability does prevent one from performing certain jobs (e.g., blind bus drivers). On the other hand, people may hold erroneous stereotypes. For example, Colella et al. (1994) found that participants held the stereotype that persons with cerebral palsy would perform very poorly as retail sales clerksa stereotype that is often shown to be false by those with cerebral palsy working as retail sales clerks. In the present study, actual performance is controlled, so that differences in evaluations cannot be accounted for by true performance. Also, we took care to devise a stereotypical poor disability-task fit scenario in which individuals with the given disability would actually be able to perform the poor-fit task. This brings us to the definition of the second term, negative bias. Bias here refers to when a person with a disability is evaluated more negatively than is a nondisabled person, holding objective performance level constant. Another issue that has been overlooked is the impact of personal consequences to raters for their personnel decisions. The vast majority of studies examining the impact of disability on personnel judgments have not provided any consequences to the raters making those judgments, that is, raters make judgments in artificial, personally irrelevant situations. Many have argued (e.g., E. F. Stone et al., 1992, in relation to disabilities) that social desirability effects override tendencies toward negative bias when the
RESEARCH REPORTS
105
judgment being made is of little importance to the rater. A theoretical justification for this observation is provided by social adaptation theory (Finer & Kahle, 1984), which posits that the adaptive significance of alternative responses in a given situation and context will determine which response an individual chooses to give. Thus, in situations of little consequence, it can be expected that socially desirable sympathy effects will surface. However, in situations of importance to the rater, negatively biased responses will most likely be evoked toward persons with disabilities. Several studies have found empirical evidence of this effect (e.g., Gibbons, Stephan, Stephenson, & Petty, 1980; Finer & Kahle, 1984). There are many ways in which consequences to the rater can be manipulated, such as having to justify one's judgments or being held responsible for inaccurate judgments. In this study, we manipulated consequences by having participants' chances for a performance-based reward either depend on one's partner's performance (reward-interdependence condition) or be solely determined by the participant's own performance (reward-independence condition). We chose this manipulation because it is one aspect of organizational reward structure predicted by D. L. Stone and Colella (1996) to interact with disability to influence how employees are treated. Also, this manipulation is familiar to student participants (i.e., their grades on team projects often depend on joint performance). Clearly, in the reward-interdependence condition, personnel judgments, which determine partners or provide reasons for choosing a partner, have potential consequences to the rater in that the partner's performance will influence the rater's chance of winning a prize. In the independent-reward condition, personnel judgments have no personal consequences to the rater. On the basis of social adaptation theory and D. L. Stone and Colella's (1996) model, we predicted that when participants' chances for rewards depend on their partner's performance, bias due to disability would be more strongly evidenced than in the independent-reward condition, particularly when disability-task fit is poor. This prediction is also in line with D. L. Stone and Michaels (1993, 1994), who found that participants were less willing to work with a person with a disability when the task was competitive (i.e., high-consequence condition) than when the task was noncompetitive. Thus, we posit:
Hypothesis 2a. There will be a three-way interaction among disability, task, and reward condition such that negative bias in choice as partner in the poor disability-task fit condition will be greater when rewards are interdependent.
Hypothesis 2b. There will be a three-way interaction among disability, task, and reward condition such that negative bias in performance evaluations in the poor disabilitytask fit condition will be greater when rewards are interdependent. Hypothesis 2c. There will be a three-way interaction among disability, task, and reward condition such that negative bias in performance expectations in the poor disabilitytask fit condition will be greater when rewards are interdependent.
Thus, we predicted that bias is more likely to occur when there are consequences and that bias will follow the pattern suggested by the disability-job fit stereotype hypothesis. That is, under conditions where raters are more likely to exhibit bias, they will provide less favorable evaluations and judgments when they perceive poor fit between the job and the disability. Method Participants
Eighty-seven graduate students served as participants; 61% were employed; 63% were White, 8% were African American, and 20% were Asian; 66% of participants were female. Students participated in various sessions, and one of the authors administered the session. Thirty-seven different students from the same pool participated in a pilot study. All participants signed an informed consent form. There were no demographic effects on dependent variables.
Pilot Study
The purpose of the pilot study was to find task and disability combinations representing a good-fit and a poor-fit scenario. We chose tasks rather than jobs because this would allow for a scenario where participants would believe that they would perform with the target individual. We chose disabilities that met several criteria: (a) They were realistic in a college setting, (b) they were visible or could be easily acknowledged, (c) they were not likely to be perceived as self-inflicted, (d) someone with the given disability could actually perform a task likely to be judged a poor tit. and (e) they were amenable to manipulation in a laboratory experiment. Pilot participants completed a matrix of seven tasks with a detailed descriptiontruck-routing exercises, words in sentences, survival task (on paper), hidden (embedded) figures, building blocks, logo design, and origamiby five disabilities deafness, legal blindness, paraplegia, amputated single arm, and dyslexia). They rated how well a person with each disability could perform each task. Ratings could range from 200% (twice as well as someone without the disability) to 0% (can't perform the task at all). A series of analyses (obtainable from Adrienne Colella) indicated that pilot participants clearly saw different combinations of disability and task as more likely to result in successful performance than did others. The combination chosen for the experimental study was dyslexia with words in sentences (poor
This effect may also hold true for performance ratings and expectations, because participants may be making these ratings to justify their choice of partner. Therefore, we predicted:
106
RESEARCH
REPORTS
fit) and embedded figures (good fit). The median rating was 20% for words in sentences (i.e., a person with dyslexia could perform this job 20% as well as someone without dyslexia) and 90% for embedded figures. It is important to note here that any individual with dyslexia may very well be able to perform the words-in-sentences task as well as anyone else. Dyslexia is formally defined as one of several distinct learning disabilities. It is a specific language-based disorder of constitutional origin characterized by difficulties in single word decoding, usually reflecting insufficient psychological processing. . . . Dyslexia is manifested by variable difficulty with different forms of language, often including, in addition to problems with reading, a conspicuous problem with acquiring proficiency in writing and spelling. (Orton Dyslexia Society, 1994) People with dyslexia vary greatly in the form and severity of impairments (Livesay & Livesay, 1995). A common myth about people with dyslexia is that they see numbers or letters backwards (which would make it nearly impossible to do the wordsin-sentences task)this is not true. Most impairments related to dyslexia take me form of not being able to connect sounds with written letters or words (although there are many other specific impairments). Also, research suggests that people with dyslexia, through training and education, can overcome reading and language difficulties and perform related tasks at the same level as someone without dyslexia and equivalent intelligence (Livesay & Livesay, 1995). In other words, it is entirely possible for a person with dyslexia to exhibit no or few impairments. To make sure that a person with dyslexia could actually perform the poor-fit task (words in sentences), we had a student who had been diagnosed with dyslexia perform the task. He performed slightly better than the average of the pilot sample used to develop task norms.
mation about each of the three men and told that they would later choose one of these men as a work partner on the upcoming task. We varied reward conditions by telling participants that their chances of winning a $100 prize depended to differing degrees on the performance of their chosen partner. In the independent reward condition, we told participants that the person with the highest score would win the $100 prize, regardless of the partner's scores. In the interdependent reward condition, we told participants that the scores of the team members would be aggregated and the team with the highest score would get a prize of $100 each. Thus, the participant's chance of winning the prize depended on the performance of his or her partner. The video showed each individual confederate participant in close up and in a long shot of the group. A voice-over introduced each of the men (Oriole, Pete, and Dave) and described the task. Initially, the three men were looking at an overhead with task instructions. After the overhead was removed, the man named Dave asked the person next to him what the last point in the instructions said. Oriole and Pete also asked the group to clarify a procedure. Thus, all three men in all conditions asked for clarification of a procedure. Holding this behavior constant rules it out as a possible explanation for any observed effects. The three men were then shown to begin working on the task. Participants saw one of two videotapes in which the target individual (named Dave) was portrayed either as dyslexic or without any mention of a disability. Dave was depicted as dyslexic by having him ask the person seated next to him to tell him what the last point of the instructions were, while explaining he was dyslexic. In the nondisabled condition, he asked just what the last point in the instructions had said. We scripted the video so that the observable performance of all three men was at an "average" (norms were provided to reinforce this). Thus, performance was held constant across individuals in the videotape and across conditions. Also, all three actors were White, approximately the same age (21-24), and rated as equally attractive in a pilot study. Performance information related to the points scored by the three actors on either of the two exercises in the last three trials. The average score obtained by all three actors was the same (100 points), with no outlying performance on any trials (scores ranged from 94 to 106). After viewing the videotape and the performance information, participants completed a questionnaire on which they made performance evaluations and ranked as a preferred partner the three men they had just viewed on the tape. While participants were filling out the questionnaires, one of the actors (never Dave) walked into the room and told the investigator that the arrangements were in place for the exercise. This was done to ensure that the participants believed that they would actually be participating in the exercise with one of the men featured on the videotape. Postexperiment interviews suggested that all participants believed that they would be working on an exercise with one of the people they saw on the tape. After handing in the first questionnaire, participants filled out a second brief questionnaire that asked about demographic information and contained manipulation checks. After the experiment was completed, we told participants about the true purpose of the experiment. We also told them that a $100 prize would be awarded in a random lottery of all study participants.
Procedure
The study was conducted in eight sessions with between 10 and 12 students participating in each session. We told students that they would participate in one of two exercises (words in sentences or embedded figures) along with a partner whom they would be allowed to choose. Participants received written descriptions and instructions for the task. The words-in-sentences task required participants to form new words and sentences using the letters from a word or phrase provided as part of the task. We scored the task so that each new word constructed was worth 5 points and each new sentence was worth 15 points. Pilot testing suggested that most students could score 100 points in 10 min. The embedded-figures task required participants to locate and circle-specific figures embedded in a larger picture. Each circled figure was worth 10 points. Pilot testing suggested that most students could score 100 points in 10 min. All participants were told that they would be performing their assigned task for 10 min. After receiving instructions, participants practiced the task for 5 min. This practice work was not scored. Following the practice trial, participants watched a videotape of three male participants who (they were told) had done the task at an earlier time. They were then given performance infor-
RESEARCH REPORTS Two participants were each awarded $100 prizes. Chances for winning were the same as if the prize had been awarded based on performance.
107
Dependent Variables
Performance rating. Participants made an overall performance rating for each person's performance on three Likertscale items scored from 1 (poor) to 7 (outstanding). We averaged the scores on these three items to form a performance rating scale (a = .91). Performance expectations. Participants estimated the number of points each man would make in a month and a year. These estimates were highly correlated (r = .87, p < .01), so we averaged them together. Rank as partner. Participants ranked each of the three men in terms of whom they would most prefer working with on the upcoming exercise (1 = most preferred) and how they thought the other people in the room would rank them. We asked participants to indicate how they believed others would rank the men to determine if there were social desirability effects. Desirability as partner. We included five items that asked participants to judge the men on skill at the task, ability to work with others, ability to be a team player, desirability as partner, and leadership potential on a 7-point Likert scale with 7 being the most positive rating. We averaged these items together to form a desirability-as-partner scale (a = .90). We also included manipulation checks to determine if participants had correctly noticed Dave's disability and the reward condition. The disability check was accomplished by including dyslexic in a list of adjectives that participants were asked to check for each individual. The reward structure check consisted of two questions asking participants to identify the degree to which their chance of winning the prize depended on their partner's performance on a 7-point Likert scale. Items were averaged together and scored so that higher scores meant that chances are more dependent on one's partner's performance.
Results All participants in the dyslexic condition and no participants in the nondisabled condition checked that Dave was dyslexic. The manipulation check for reward condition found the two groups to differ significantly, F(l, 86) = 6.36, p < .01, with those in the interdependent reward condition reporting more influence of their partner's per-
formance on their chances of winning (M = 4.58) than those in the individual condition (M = 3.85). Correlations and descriptive statistics for all dependent variables are depicted in Table 1. Note that the mean expected number of future points is 102.40, which is close to the average number (100) reported in the performance data given to participants. This suggests participants paid attention to the performance data and did not expect performance to change over time. We tested hypotheses using 2 (disability) X 2 (job) x 2 (reward) univariate analyses of variance on ratings of target Dave. We also conducted within-subject analyses for all dependent variables. Results of these analyses, which yielded the same conclusions as the univariate analyses, can be obtained from Adrienne Colella. Univariate results are presented for ease of interpretation. In hypotheses la through Ic, we stated that there would be a significant Disability X Task interaction effect on performance evaluations, expectations, and choice as partner, respectively. As can be seen in Table 2, Hypotheses la and Ib, regarding performance ratings and expectations, were not supported. None of the experimental manipulations had a significant effect on these judgments. On the other hand, Hypothesis Ic was supported with respect to rankings and ratings of partner desirability, variables presumably associated with greater consequences for the rater. There were significant Disability x Task interactions for rank as partner, F(l, 86) = 6.73, p < .01, ui2 = .07, estimations of others' rankings, F(l, 86) = 8.53, p < .01, a;2 = .09, and partner desirability ratings, F(l, 86) = 4.50, p < .05, u2 = 0.05. Cell means for these interactions and post hoc tests of the differences between means are presented in Table 3. As predicted, both participants' own and participants' estimation of others' rankings were lowest in the poor-fit condition (3 is the lowest ranking). Partner preference ratings in the poor-fit condition were significantly lower than nondisabled ratings on the same task and disability ratings in the goodfit condition. Close examination of the means in Table 3 reveals an effect that we did not predict. Estimations of others' ranking and desirability ratings, in the good-fit condition, were
Alphas in Diagonal)
SD
0.97 12.53 0.59 0.61 0.90
-.244*
(.90)
*p<.05.
**p<.01.
108
Table 2 Analysis of Variance Results
RESEARCH REPORTS
Dependent variables Performance rating Effect Disability Reward Disability X Reward Task Disability X Task Reward X Task Disability X Reward X Task Total model t p < .10. *p<.05. df F
0.52 0.41 0.29 1.96 0.57 0.37 0.12 0.65
Performance expectations
Rank as partner
Others' rank
Desirability as partner
w2
.000 .000 .000 .015 .000 .000 .000
F
0.36 0.00 1.08 1.02 0.14 0.22 0.15 0.46
u,2
.000 .000 .001 .000 .000 .000 .000
F
3.77f 7.03** 6.42** 0.01 6.73** 1.14 0.03 3.89**
u,2
F
0.01 5.39* 0.05 6.48** 8.53** 2.10 0.24 3.32**
c/
F
0.34 0.03 0.21 0.04 4.50* 0.91 0.78 1.06
u2
.000 .000 .000 .000 .054 .000 .000
1 86
**p<.01.
more positive when Dave had dyslexia than when he did not. This does not contradict Hypothesis Ic but additionally suggests positive bias in the good-fit condition. In hypotheses 2a through 2c, we predicted a significant Disability X Task x Reward reaction on all judgments. This interaction was not significant for any of the dependent variables. However, reward condition did have an effect on rank as partner, the judgment most closely tied to the consequences associated with reward condition. In relation to Hypothesis 2a, there was a significant Disability X Reward effect for rankings as partner, F(l, 86) = 6.42, p < .01, u2 = .06, in addition to the significant Disability X Task interaction. Analyses of means (see Table 3) show that rankings in the dyslexia condition were lower than in the other conditions only when rewards
were interdependent, that is, there was a main effect for disability under only interdependent reward conditions. There was negative bias against the target with dyslexia, regardless of fit, when the participants' chance of winning depended on the target's performance. Discussion Contrary to Hypotheses la and Ib, disability, in interaction with task, had no effect on performance ratings and predictions about future performance. Participants' accurately rated performance and expectations about future performance reflected current performance. It appeared that participants expected performance to remain constant over time. Had they thought that performance on these
Table 3
Cell Means for Significant Interactions
Disability X Task Interactions Dependent variables: rank as partner, other's rank, desirability as partner Rank!is partner Task Words in sentences Embedded figures Dyslexia
2.79, (poor fit) 2.45b (good fit)
Nondisabled
2.17C
Nondisabled 4.72b
4.25,
2.50b
Disability X Reward interation Dependent variable: rank as partner Reward Interdependent Independent Dyslexia
2.90. 2.20,
Nondisabled
2.35, 2.31b
Note. Means for each dependent variable with different subscript letters are significantly different at least at the .05 level, based on Student Newman-Keuls and TUkey post hoc tests of means.
RESEARCH REPORTS
109
tasks would vary over time, perhaps there would have been a disability-job fit effect on expectations or ratings. It may well be that if raters view someone with a disability (in a poor-fit job) perform as well as someone without a disability or a disability viewed as unrelated to job performance, they may believe that the person in the poorfit situation is performing at maximum potential whereas the others will improve over time. This is an interesting question that deserves attention in future research. On the other hand, this lack of findings may be due to situational factors such as the provision of clear performance information and standards (participants received normative data; cf. Czajka & DeNisi, 1988). Research that more closely examines the impact of disability on future performance expectations and performance attributions as a function of varying levels of ambiguity of performance information is warranted. Hypothesis Ic regarding choice of partner received clear support. Negative bias occurred against only the target with the disability in the poorfit condition. These were decisions where the outcome would influence with whom the participant would work and potentially influence his or her chances of winning. Thus, out of five dependent variables, the stereotypical fit hypothesis was supported on the three decisions that had implications for the participants (regardless of reward condition) and not supported on those judgments that had no implications. Interestingly, positive bias in desirability rankings and estimation of others' rankings was found for the target with dyslexia under good-fit conditions. We did not predict such an effect. It needs to be noted here that this effect was not found for actual rankings for partner choice. This pattern of findings loosely demonstrates that positive bias or general sympathy effects toward persons with disabilities may disappear when the action called for has some kind of direct possibility of impacting on the evaluator. Hypotheses 2a through 2c were not directly supported. Yet, there was some suggestion that the reward condition, in conjunction with disability, did influence rank as partner. There was a main effect for disability (irrespective of fit condition) only when chances for obtaining the reward were dependent on the partner's performance. Thus, it appears again that negative bias is most likely to be demonstrated under only conditions where the action has some consequences for the participant. This is an important finding in that most research on the evaluation of persons with disabilities does not provide participants with consequences for their behavior and judgments (cf. E. F. Stone et al., 1992), and yet, in field settings, it is quite common for ratings and recommendations to have personal consequences for raters. Our results suggest that it is critical that such consequences be included in any study of persons with disabilities at work. A note is needed about our definition of consequences
to raters. We denned consequences as interdependence in obtaining a valued reward. In retrospect, disability effects were found on only those judgments that would influence with whom the participant would work. It may be that potential interaction alone (apart from reward contingency) is enough of a consequence to elicit bias. This is in line with research demonstrating that nondisabled people often avoid interacting with people with disabilities (e.g., Snyder et al., 1979). There is little research on this issue in work settings, and more is needed. One other issue deserving of attention is that the current research outlines some conditions under which bias is (or is not) likely to occur; however, it does not address the processes underlying why the bias may occur. For example, we found bias in choice of partner (in poor-fit and interdependent reward conditions) but found no effect on performance expectations. This suggests that participants did not rank Dave lower when he was portrayed as dyslexic because they expected him to perform poorly in the future. Why did they provide lower ranks in the poor-fit and interdependent reward conditions? The data also show that when rewards were interdependent, there was a main effect for disability. Clearly more was operating here than fit stereotypes. It should be noted that the explanatory mechanisms (performance expectations and fit stereotypes) used in the current study were all based on performance. That is, we examined only the possibility that evaluations of people with disabilities were influenced by participants' beliefs about the impact of the disability on job performance. Many reasons have been put forth for biased treatment of persons with disabilities (see D. L. Stone & Colella, 1996, for a review) including attraction, affect, and fear, factors that have little to do with job performance. It may be that participants in the interdependent reward conditions felt that they would be working more intimately with the target individual, and thus, other reactions (besides uncertainty about the target's ability to perform the task), such as negative affect or anticipated strain, influenced participants' willingness to work with the target. Thus, one pressing need for future research is to examine processes underlying potential bias other than those related to performance expectations. There are a few limitations to the present research. First, the participants were students. However, the majority had work experience and it has been demonstrated that employers tend to hold the same stereotypes as the general public (Christman & Slaten, 1991). Also, the task and situation were personally relevant to students who often work on group projects where their reward (i.e., grade) depends on the performance of others. Furthermore, the possibility of winning a $100 prize most likely ensured that students were engaged in the experiment. Second, attempts were made to make the performance evaluation situation as realistic as possible; however, this was a labo-
110
RESEARCH REPORTS
ratory context so that objective performance could be held constant. Research that attempts to replicate these findings in organizational settings using actual raters in actual evaluation settings is needed. However, so little experimental work has been done in this area, and there is so little understanding of the dynamics involved, that future experimental work is also warranted. Finally, it needs to be noted that the issue of accommodation was not addressed. Although participants were asked to assume reasonable accommodation when making fit judgments in the pilot study that determined disabilityjob fit stereotypes, it may be that disability-fit stereotypes do not hold or are weakened when raters actually see a person use an accommodation. This situation may reflect real life encounters in that people form disability-job fit stereotypes without considering how the person with a disability would perform given various types of accommodation. Most research on accommodation has been limited to examining the cost of accommodation. Research that examined the impact of accommodation on performance expectations and stereotypes would be quite enlightening. In sum, we found little evidence of negative bias against persons with disabilities in terms of performance judgments and expectations. We posited a variety of reasons why this may be the case, but what is clear from these results is that consequences to raters certainly matter. For personnel judgments that could have implications for the raters, there was negative bias, because of disability alone and the combination of disability and the job. From an applied standpoint, these results suggest that persons with disabilities may receive differential treatment at work. It is an empirical question as to whether these results would generalize from this laboratory setting to a situation where a person with dyslexia was working at a job requiring a great deal of precise reading and writing, although anecdotal accounts suggest they would (Blalock, 1981; Vaz, 1996). From a conceptual standpoint, the results of this study suggest factors that may consolidate the scattered and mixed results from past research regarding the treatment and evaluation of persons with disabilities at work. Here we suggest that the nature of the disability, the nature of the job, raters' stereotypes regarding disability-job fit, and the consequences of the rater's judgments all work together to influence judgments. Future research should address these issues, because it is then that we will be able to begin making statements about under what conditions and in what form a person's disability will impact on the way he or she is evaluated, judged, and treated at work.
References Blalock, J. W. (1981). Persistent problems and concerns of young adults with learning disabilities. In W. Cruickshank &
A. Silver (Eds.), Bridges to tomorrow: The best of ACLD (Vol. 2, pp. 35-55). Syracuse, NY: Syracuse University Press. Bordieri, J. E., & Drehmer, D. E. (1986). Hiring decisions for disabled workers: Looking at the cause. Journal of Applied Social Psychology, 16, 197-208. Bowman, J. T. (1987). Attitudes towards disabled persons: Social distance and work competence. Journal of Rehabilitation, 53, 41-44. Braddock, D., & Bachelder, L. (1994). The glass ceiling and persons with disabilities. Washington, DC: U.S. Department of Labor, Glass Ceiling Commission. Christman, L. A., & Slaten, B. L. (1991). Attitudes towards people with disabilities and judgments of employment potential. Perceptual and Motor Skills, 72, 467-475. Colella, A. (1996). The organizational socialization of employees with disabilities: Theory and research. In G. R. Ferris (Ed.), Research in personnel and human resources management (Vol. 14, pp. 351-417). Greenwich, CT. JAI Press. Colella, A., DeNisi, A. S., & Varma, A. (1997). Appraising the performance of employees with disabilities: A review and model. Human Resource Management Review, 7, 2753. Colella, A., DeNisi, A. S., Varma, A., & Lund, M. (1994, August). Evaluations and personnel decisions regarding persons with disabilities: The role of stereotypical fit. Paper presented at the meeting of the National Academy of Management, Dallas, TX. Cox, T. (1993). Cultural diversity in organizations: Theory, research & practice. San Francisco: Barrett-Koehler Publishers. Czajka, J. M., & DeNisi, A. S. (1988). Effects of emotional disability and clear performance standards on performance ratings. Academy of Management Journal, 31, 394404. Dipboye, R. L., Fromkin, H. L., & Wibak, K. (1975). Relative importance of applicant sex, attractiveness, and scholastic standing in the evaluation of job applicant resumes. Journal of Applied Psychology, 60, 39-43. Eden, D. (1990). Pygmalion in management: Productivity as a self-fulfilling prophecy. Lexington, MA: Lexington Books. Farina, A., & Ring, K. (1965). The influence of perceived mental illness on interpersonal relations. Journal of Applied Social Psychology, 70, 47-51. Fichten, C. S., & Amstel, R. (1986). Trait attributions about college students with a physical disability: Circumplex analysis and methodological issues. Journal of Applied Social Psychology, 16, 410-427. Gibbons, F. X., Stephan, W. G., Stephenson, B., & Petty, G. R. (1980). Reactions to stigmatized others: Response amplification vs. sympathy. Journal of Experimental Social Psychology, 16, 591-605. Goffman, E. (1963). Stigma: Notes on the management of spoiled identity. Englewood Cliffs, NI: Prentice Hall. Grand, S. A., Bernier, J. E., & Strohmer, D. C. (1982). Attitudes toward disabled persons as a function of social context and specific disability. Rehabilitation Psychology, 27, 165-174. Greenwood, R., Shiner, K. F., & Johnson, V. A. (1991). Employer concerns regarding workers with disabilities and the business-rehabilitation partnership: The PWI practitioners' perspective. Journal of Rehabilitation, 57, 21-25.
RESEARCH REPORTS Hastorf, A.M., Northcraft, G. B., & Picciotto, S.R. (1979). Helping the handicapped: How realistic is the performance feedback received by the physically handicapped? Personality and Social Psychology Bulletin, 5, 373-376. Heilman, M. (1983). Sex bias in work settings: The lack of fit model. In L. L. Cummings & B. Staw (Eds.), Research in organizational behavior (Vol. 5, pp. 269-298). Greenwich, CT: JAI Press. Jones, E. E., Farina, A., Hastorf, A. H., Markus, H., Miller, D. T., Scott, R. A., & de Sales-French, R. (1984). Social stigma: The psychology of marked relationships. San Francisco: W. H. Freeman. Knox, M., & Parmenter, T. (1993). Social networks and support mechanisms for people with mild intellectual disabilities in competitive employment. International Journal of Rehabilitation Research, 16, 1-12. Livesay, Y., & Livesay, R. C. (1995, March). National Conference for Learning DisabilitiesA report. Retrieved from World Wide Web: http://www2.dyslexia-add.com/dyslexiaadd/research.html#definition. Locksley, A., Borgida, E., Brekke, N., & Hepburn, C. (1980). Sex stereotypes and social judgment. Journal of Personality and Social Psychology, 39, 821-831. Louis Harris and Associates, Inc. (1987). The ICD survey II: Employing disabled Americans. New %rk: Author. McFarlin, D. B., Song, J., & Sonntag, M. (1991). Integrating the disabled into the work force: A survey of Fortune 500 company attitudes and practices. Employee Responsibilities and Rights Journal, 4, 107-122. Miller, M. (1992). The effects of disability-related information on employment decisions regarding persons with disabilities. Unpublished doctoral dissertation, University of Oregon. Orton Dyslexia Society. (1994, April). Research definition. Retrieved from World Wide Web: http://www2.dyslexiaadd.com/ dyslexia-add/research.html#definition. Finer, K. E., & Kahle, L. R. (1984). Adapting to the stigmatizing label of mental illness: Foregone but not forgotten. Journal of Personality and Social Psychology, 47, 805-811. Rose, G., & Brief, A. (1979). Effects of handicap and job characteristics on selection evaluations. Personnel Psychology, 32, 385-392. Russell, D., Lenel, J. C., Spicer, C., Miller; J., Albrecht, J., & Rose, J. (1985). Evaluating the physically disabled: An attri-
111
butional analysis. Personality and Social Psychology Bulletin, 11, 23-31. Schneider, C. R., & Anderson, W. (1980). Attitudes toward the stigmatized: Some insights for recent research. Rehabilitation Counseling Bulletin, 23, 299-313. Scotch, R. K. (1988). Disability as a basis for social movement: Advocacy and politics of definition. Journal of Social Issues, 44, 159-172. Snyder, M. L., Kleck, R. E., Strenta, A., & Mentzer, S. J. (1979). Avoidance of the handicapped: An attributional ambiguity analysis. Journal of Personality and Social Psychology, 37, 2297-2306. Sowa, P. A., & Cutter, H. S. (1974). Attitudes of hospital staff toward alcoholics and drug addicts. Quarterly Journal of Studies on Alcohol, 35, 210-214. Stone, D. L., & Colella, A. (1996). A framework for studying the effects of disability on work experiences. Academy of Management Review, 21, 352-401. Stone, D. L., & Michaels, C. (1993, August). Factors affecting the acceptance of disabled individuals in organizations. Paper presented at the meeting of the National Academy of Management, Atlanta, GA. Stone, D. L., & Michaels, C. (1994, August). Effects of nature of the disability and competetiveness of the reward system on selection of disabled team members. Paper presented at the meeting of the National Academy of Management, Dallas, TX. Stone, E. F., Stone, D. L., & Dipboye, R. L. (1992). Stigmas in organizations: Race, handicaps, and physical unattractiveness. In K. Kelly (Ed.), Issues, theory, and research in industrial and organizational psychology (pp. 385457). New \brk: Elsevier Science. Varma, A., Colella, A., & DeNisi, A. (1996, April). Performance evaluation of individuals with disabilities: The role of ingratiation. Paper presented at the Conference of the Society for Industrial and Organizational Psychology, San Diego, CA. Vaz, A. M. (1996). The doctor he begged to be. Pittsburgh, PA: Dorrance. Yuker, H. E. (1983). The lack of stable order of preference for disabilities: A response to Richardson and Ronald. Rehabilitation Psychology, 28, 93-103.
Received October 7, 1996 Revision received August 6, 1997 Accepted August 17, 1997