The effects of rater sex and ratee sex on managerial performance evaluation.

Abstract:

Thirty-three men and twenty-four women evaluated and rated the performance of two managers (a man and a woman) after the managers had made decisions based on information provided by the management accounting system. The experimental materials comprised two cases--one showing successful performance and the other, unsuccessful performance. It is hypothesised that both men and women raters will demonstrate a bias against the successful woman manager, that is, evaluate her more harshly compared with the successful male manager. No difference in performance evaluation is predicted between the unsuccessful female manager and the unsuccessful male manager. The results support the hypotheses. A second dependent variable measured the perceived benefit that was received from expenditure incurred in carrying out the managers' decisions. It is hypothesised that both men and women raters will demonstrate a bias against the successful woman manager, that is, perceive lower benefit from a woman-initiated decision relative to a man-initiated decision. No evidence of biases against the successful woman manager is observed. The paper concludes with a discussion of the implications of these findings.

Keywords:

SUCCESSFUL AND UNSUCCESSFUL PERFORMANCE; BIAS AGAINST WOMEN; PERCEIVED BENEFIT; MENTAL ACCOUNTS; SEX STEREOTYPES.

1. Introduction

Senior management tends to be male-dominated. Earlier human resource and personnel literature frequently report that men may evaluate women more harshly relative to their evaluation of other men (e.g. Decker 1987; Dubno 1985). Consequently, women ratees (managers) may be disadvantaged because they may be evaluated less favourably compared to their male colleagues for employment, promotion, and pay rises. A recent study by Johnson, Kaplan and Reckers (1998) examines the performance evaluation of men and women ratees when they perform at an equal but unsuccessful level. The results show no significant difference in their performance evaluation. However, the literature reports that performance evaluation is affected by more factors than just sex of ratee alone. It may also be affected by level of performance, for example, successful and unsuccessful (e.g. Lipe 1993), and sex of the rater(s) (e.g. Decker 1987; Dubno 1985). These two variables have received little research attention.

Another factor that could affect performance evaluation is mental accounting. This refers to the use of a psychological account to evaluate actions that may lead to different consequences (Kahneman & Tversky 1984). Lipe (1993) demonstrates that more benefit is perceived to be received from expenditures incurred by a successful manager relative to expenditure incurred by an unsuccessful manager. In addition to affecting performance evaluation, raters' biases and subconscious preconceptions may affect their mental accounts (Kahneman & Tversky 1979) since the values that govern people's actions are often framed by their firmly established social biases and prejudices (Neilsen & Bartunek 1996). Assuming raters' biases against women, Kahneman and Tversky's (1979) suggestion implies that this bias may extend to their mental accounts as well.

The purposes of this study are three-fold. First, it examines ratees' performance at two levels--successful and unsuccessful--to explore whether biases against women are dependent upon their level of performance. Second, it provides experimental evidence on whether the participation of women as raters promotes equity in performance evaluation for women ratees. (1) Third, it investigates whether biases against women extend to the ratees' mental accounts. Overall, the results of this study provide insight into the nature of biases against women managers, and the form and extent such biases may take.

The next section develops the theories used in this study. Section 3 discusses the experiment and section 4 reports the results. The paper concludes in section 5 with a discussion on the implications and limitations of the study and makes suggestions for future research.

2. Theory Development

2.1 Bias Against Women in Patriarchal Societies

Sex bias occurs when one sex is held to a higher performance standard relative to the other sex. A bias against women occurs when this bias disadvantages women. In patriarchal societies, which include most Western countries, females are a group often discriminated against in various facets of life, including the workplace. In response to this, various countries including the USA, Australia and the United Kingdom, have enacted legislation making illegal discrimination against women in hiring practices. This legislation has been relatively effective in reducing overt discrimination against women. For example, Almer, Hopper and Kaplan (1998) find no bias against women in hiring decisions. However, less overt forms of workplace discrimination against women remain possible. For example, discrimination may have become subtler and may have shifted to performance evaluation instead (Swim 1993; Shore 1992). This is because performance evaluation is essentially a subjective process and proving discrimination is often difficult. In these situations, raters' biases, prejudices and stereotypes are more likely to be introduced into the evaluation process compared to hiring decisions (Swim 1993; Shore 1992). In addition, raters may not be aware that their internalised biases have skewed their evaluations as these biases and stereotypes are inherently subconscious (Johnson, Kaplan & Reckers 1998).

In patriarchal societies, the word `gender' may conjure up (to some people) stereotypes of men as being competent, capable and valuable to society whereas women are often seen as communing, nurturing and submitting (Hull & Hicks 1993; Ridgeway 1993). Categorisation theory is used to explain the pro-male bias of society. This theory states that people have a tendency to compartmentalise their environment so that they can understand it better. It involves classifying objects and events in their surroundings into categories. Each category is then identified by a summary of its most salient features--a stereotype (Rosch 1978). New stimuli are analysed and compared to the stereotype (Rosch 1978). These stereotypes together with people's prejudices and emotional values dictate their perceptions of how men and women should behave and act (Hull & Umansky 1997). For example, stereotypes have been found to be sufficiently strong enough to cause irrational behaviour in the area of management evaluations (Morrison & Von Glinow 1990). When extended to the workplace, categorisation theory suggests that raters compare ratee managers to their successful manager stereotype. As raters are traditionally exposed to more men as managers rather than women as managers, a feature of their stereotypic manager is that he is a man. Women ratees do not match this stereotype, and consequently, may be evaluated less favourably relative to men. Evidence of Decker (1987) and Dubno (1985) suggests that stereotyping works against women, and in favour of men.

Sex stereotyping extends to occupations as well and is believed to be a major contributor to the present division of labour between men and women (Cejka & Eagly 1999; St Pierre, Herendeen, Moore & Nagie 1994; Glick 1991). This stereotyping of the workplace often works by promoting and maintaining men as leaders and women as subordinates (Hull & Umansky 1997). Through this, sexual division of labour is justified. This belief maintains that success in occupations dominated by one sex requires the personal characteristics typical of that sex (Cejka & Eagly 1999; Hull & Umansky 1997). For example, secretarial work and childcare are perceived to require feminine qualities while engineering and management require masculine qualities. This stereotyping of occupations is reflected in performance evaluation as well so that in male-typed occupations such as management, men are rated more highly than women. This may explain why senior management of most organizations continue to be dominated by men.

Studies by Decker (1987) and Dubno (1985) report a bias against women among men, that is, men may hold women to higher performance standards than they hold other men. In addition, men may have a tendency to devalue women leaders (see meta-analysis by Eagly, Makhijani and Klonsky 1992). Both Eagly, Makhijani and Klonsky (1992) and Judge and Ferris (1993) cite men's higher status in society as an explanation. They may evaluate women more harshly (relative to men) in order to maintain their societal status (Judge & Ferris 1993; Eagly, Makhijani & Klonsky 1992). Another example of men's bias against women concerns the way they interact with women (Ridgeway & Johnson 1990). Ridgeway & Johnson (1990) report that higher-status people--such as men--may use negative expressions when interacting with lower-status people--such as women. Such expressions serve two purposes. First, they may convey to women that their judgments are not valued, and second, they may serve as a mechanism for maintaining the social hierarchy (Ridgeway & Johnson 1990). Given men's possible negativity towards and biases against women, the latter's performance evaluations may suffer because the evaluation process is inherently subjective and men traditionally dominate as raters.

However, biases against women are not confined to men alone as the related literature further suggests that, in the presence of men, women may evaluate other women more harshly than they would when no men are present (Todor 1980). The theory for this bias is based on men's/women's status in society (Paludi & Strayer 1985). In patriarchal societies, the role of men is more highly valued by both men and women relative to the role of women. As a result, men's positive behaviour may be valued more highly compared to the identical behaviour of women (Paludi & Strayer 1985). (2)

2.2 Level of Performance

Biases against women (whether from men or from other women) may not hold true in all situations. For example, Deaux and Emswiller (1974), and Garland and Price (1977) show that, in a male domain such as management, successful and unsuccessful women are not subjected to the same biases. There is a tendency among raters to undervalue the success of women by attributing such success to good luck or easy job while the success of men is attributed to skill (Deaux & Emswiller 1974; Garland & Price 1977). However, there is no tendency to attribute negative causes to explain women's non-performance. For example, in an auditing context, Johnson, Kaplan and Reckers (1998) show that non-performing men and women ratees were not rated significantly differently. There are two potential explanations for this. First, since women have lower societal status, they may be held to higher performance standards to prove their competency (Foddy & Smithson 1989; Foschi 1989). At the same time, they receive greater forbearance when they do not succeed (Foddy & Smithson 1989; Foschi 1989). The second explanation is based on attribution theory which suggests that raters search the environment for causes or reasons to explain observed events. As raters' stereotypical manager is not a woman (Brown & Geis 1984; Schein 1973), reasons have to be found to explain the success of a woman manager. Very often, her success is attributed to some factors other than skill. However, an unsuccessful woman in a male domain (such as management) is consistent with raters' stereotype, and therefore the need for attribution is unnecessary. This explanation is supported by the results of Johnson, Kaplan and Reckers (1998).

2.3 Use of Management Accounting Information

`The concept of responsibility accounting is widely held to be an essential feature of any respectable management accounting system' (Choudhury 1986, pg. 189). Managers make use of the information provided by the management accounting system (MAS) to manage their area(s) of responsibility and, in turn, their performance will be evaluated by senior management. Very often, this evaluation would be based on the financial results. For example, Lipe (1993) shows that managers receive higher evaluations when their decisions (which are based on the management accounting information) result in some perceived value to the company -- a successful performance. They receive lower evaluations when their decisions do not result in any perceived value to the company -- an unsuccessful performance. Managers' perception of the fairness by which they are evaluated affects their use of the MAS information. A manager who is evaluated unfairly could attribute this to other causes (e.g. his/her use of MAS information) and this could affect future behaviour (such as reduced reliance on MAS information) in a way that damages the organization.

Based on the above discussion, managers may perform successfully or unsuccessfully, and the biases experienced by women may be affected by the level of their performance. As a result, two performance levels will be used in this study--successful and unsuccessful--with the prediction that raters will demonstrate a bias against successful woman ratees by rating their performance lower relative to men ratees. In contrast, no bias will be demonstrated against the unsuccessful woman ratee is expected. Therefore, the following hypotheses are tested:

H1: When making performance evaluations:

a. Both men and women raters will demonstrate a bias against women ratees in the successful performance condition; and

b. Both men and women raters will not demonstrate a bias against women ratees in the unsuccessful performance condition.

2.4 Mental Accounting of Investigation Expenditure

Mental accounting refers to the use of a psychological account to evaluate actions that may lead to different consequences (Kahneman & Tversky 1984). The use of this account is a method of framing decisions. A two-stage process of editing and evaluating information is employed (Kahneman & Tversky 1979). First, a mental account is employed in the editing process to assess the costs and benefits of the available choices. This account simplifies cognitive processes by decomposing complex decision problems into smaller, more manageable segments. Relevant information is accumulated and evaluated and posted to this account. Information considered to be irrelevant is posted to a separate account. Second, the choices are then evaluated using prospect theory, and the literature (e.g. Lipe 1993) observes a framing effect on these choices.

Mental accounts have been widely used to explain economic behaviour. For example, they cause people to behave in ways that are contrary to the basic economic assumption of rationality (Tversky & Kahneman 1981). In addition to causing irrational economic behaviour, Kahneman and Tversky (1979) also speculate that mental accounts may be affected by decision-makers' biases. Schweitzer (1999) explains how this effect could occur. According to him, mental accounts are constructed using some internalised reference points. The process of constructing these accounts is highly unpredictable, easily influenced and context dependent, and it is at this juncture that biases can enter the decision process (Schweitzer 1999). If decision makers' framing of the decision context and their mental accounts are affected by their biases (Kahneman & Tversky 1979), then biases against women may also extend to the mental accounting of the perceived benefit resulting from expenditure incurred in carrying out the ratee's decision. Consequently, the following hypotheses state:

H2: In the mental accounting of the perceived benefit resulting from expenditure incurred in carrying out the ratee's decision:

a. Both men and women raters will demonstrate a bias against women ratees in the successful performance condition; and

b. Both men and women raters will not demonstrate a bias against women ratees in the unsuccessful performance condition.

3. Experiment

3.1 Participants

The participants comprised 57 (33 men and 24 women) students enrolled in Master of Professional Accounting courses at two Australian universities. They came from a wide range of industries including banking, hospitality, engineering, computer technology, health and education. About half of them worked in accounting and accounting-related areas. Senior managers made up 23% of the participants and included financial controllers, an accounting firm partner, bank managers and business proprietors. Middle managers made up 56% of the participants and included accountants, tax consultants, software engineers and financial planners. The rest comprised participants employed in entry-level positions and included trainee accountants, accounts clerks, assistant accountants, and administrative assistants. The descriptive data are shown in table 1. At the time of testing, the participants had completed a course in managerial accounting.

3.2 Tasks and Procedures

A two (rater sex) by two (ratee sex) full factorial experiment was carried out. As discussed above, two cases were used. One case was developed for this study. It reported an unsuccessful ratee. This case was pretested on five colleagues and where necessary, adjustments were made. The amended case was pilot-tested on two research assistants who were graduate students. The final case is shown in appendix A. The other case was the case used by Lipe (1993) and it described a successful ratee (see appendix B). In the unsuccessful performance condition, the ratee incurred expenses in investigating an inventory shortage. The probability of the need to increase inventory levels was given as 50%. Based on the results of the investigation, inventory levels were increased. However, due to unforeseen circumstances, the expected increase in demand did not eventuate. In the successful performance condition, an unfavourable variance was reported. The ratee incurred expenses in investigating it. The probability of discovering an `out of control' system was also given as 50%. The investigation uncovered an `out of control' system that was later adjusted.

Ratee sex was manipulated by changing the name of the ratee in both cases. Sex-specific names and pronouns were used. All participants performed both tasks. Half of the participants received the male version of successful case and the female version of the unsuccessful case. For the other half of the participants, the sex of the ratees was reversed. Half of the participants saw the two cases in the male/female order while the other half saw the cases in the female/male order.

The experiments were carried out at the beginning of regular class time and in a controlled setting. The experimental materials were presented to the participants in the form of a booklet. Introductory information provided the purpose of the experiment (which was to examine performance evaluation) as well as an assurance of confidentiality. After reading the first case, participants evaluated the performance of the manager and rated the benefits received from the expenditure incurred in carrying out the manager's decision. Next, they performed a filler task that required them to calculate some financial report balances. This task was intended to clear their minds of the previous tasks and was not analysed. Subsequently, they read the second case and performed tasks similar to the first case. According to Lipe (1993), the evaluation system under which ratees work may have an effect on their susceptibility to being evaluated based on situations outside of their control. In both cases used in this study, the result of the managers' decision was outside of their control. Consequently, participants in this study were asked if they were held responsible for situations outside of their control, and whether it was appropriate for their employer to hold them responsible for such situations. As the cases were set in a retail and a manufacturing context, participants also recorded their familiarity with these areas. (3) Finally, the participants provided both demographic and diagnostic data. These included their work experience (in years), their motivation and the amount of effort expended on the experimental tasks. They each received A$10 for their efforts.

4. Results

Initially, the dependent variables were tested for normality. Both dependent measures in the unsuccessful condition were found to be normal (z < [+ or -] 1 for both). However, in the successful condition, both dependent variables were not normal (z > [+ or -] 1 for both), and were skewed right. Consequently, square root transformations were applied.

H1a states that both men and women raters' evaluation of the ratees will demonstrate a bias against the successful woman, and H1b predicts no bias against the woman ratee who performs unsuccessfully. The dependent variable was measured on a scale anchored by `very poor' (0) and `very good' (100) (refer appendices A & B). As raters' perception of the appropriateness of being held responsible for events outside of their control and the frequency of this occurring may affect their evaluation of others (Lipe 1993), these two variables were included as covariates. In addition, work experience, motivation and effort expended were included because of their potential effects on the results. (4) The means and standard deviations (sd) of the performance measure are shown in panel A (successful performance condition) and panel B, table 2 (unsuccessful performance condition). In the successful performance condition, the overall mean (sd) for the performance evaluation was 71.47 (20.64) (panel A, table 2). The ANCOVA results showed that the covariates did not have significant effects on the dependent variable. (5) Similarly, rater sex did not have a significant effect on performance evaluation (F = 0.24, p > 0.05). This indicated that there was no significant difference in the men and women raters' evaluation of the ratees. Consistent with H1a, there was a significant main effect for ratee sex (F = 2.93, p = 0.046). The man ratee received a mean (sd) score of 76.32 (16.98) while the woman ratee received a mean score of 66.79 (22.97) (table 2). This resulted in a difference of 9.53 points. There was no significant interaction effect between rater sex and ratee sex on performance evaluation (F = 0.00, p > 0.05).

In the unsuccessful performance condition, the overall mean (sd) for performance evaluation was 60.91 (21.34) (panel B, table 2). The covariates similarly did not have significant effects on the dependent measure (p > 0.05). Consistent with H1b, there were no rater sex and ratee sex main effects (F = 0.89, p > 0.05; F = 0.18, p > 0.05 respectively). The mean (sd) for the man ratee was 58.93 (22.01) and 62.96 (20.82) for the woman ratee (panel B, table 2). There was also no significant interaction effect between the two independent variables on performance evaluation (F = 0.16, p > 0.05). The results showed that both men and women raters did not evaluate the man and the woman ratees significantly differently. Thus, H1a and H1b are supported.

The next two hypotheses predict that the same bias against the woman ratee will be observed in the perceived benefit derived from the expenditure incurred in carrying out the ratee's decision. This was also tested at two levels of performance--successful and unsuccessful. The dependent variable was measured on a scale anchored by `no, there was no benefit' (0) and `yes, there was a benefit' (100) (refer appendices A & B). The means (sd) are shown in table 3.

When the performance was successful, the overall mean (sd) for perceived benefit was 84.35 (14.99) (panel A, table 3). Two covariates had significant effects on the dependent measure (F = 6.62, p = 0.006 [raters' perception of the appropriateness of being held responsible for events outside of their control] and F = 3.04, p = 0.044 [the frequency of the former occurring]). Consistent with the performance evaluation measure, there was no significant main effect for rater sex (F = 1.12, p > 0.05). Inconsistent with hypothesis 2a, there was no significant main effect for ratee sex (F = 2.37, p = 0.065), although the means showed a difference of 6.82 points (mean scores [sd] were 87.82 [16.18] and 81.00 [13.17] respectively) (table 3). The interaction effect between rater sex and ratee sex was not significant (F = 0.14, p > 0.05). The results do not support H2a.

In the unsuccessful performance condition, the overall mean (sd) for perceived benefit was 47.74 (29.02). The covariate--frequency of being held responsible for events outside of the raters' control--had a significant effect on the dependent measure (F = 4.24, p = 0.022). The other covariates did not have significant effects on the dependent variable. There was no significant main effect for rater sex (F = 2.44, p = 0.062). There was also no significant main effect for ratee sex (F = 0.69, p > 0.05). The mean (sd) for the man ratee was 48.10 (27.88) and for the woman ratee it was 47.36 (31.04). The interaction effect was similarly not significant (F = 1.86, p > 0.05). The results support H2b.

5. Conclusion

In this study, managers (a man and a woman) made decisions to investigate a shortage of inventory and an unfavourable variance. In one case, expenditure was incurred in investigating an inventory shortage. The outcome of the investigation revealed a 50% probability that inventory levels should be increased. The manager made the decision to increase inventory levels. However, due to unforeseen circumstances, the expected demand did not materialise. This formed the unsuccessful performance condition. In the other case, an unfavourable variance was reported. There was a 50% chance that the system was out of control and required adjustment. The manager made the decision to incur expenses to investigate the variance. The investigation revealed an `out of control' system which was subsequently corrected, and this formed the successful performance condition. This study suggests that there may be a bias against the successful women managers in performance evaluation as successful male managers are evaluated higher relative to successful women managers. This difference in evaluation is observed in both men and women raters. No such difference is observed when unsuccessful male ratees are compared to unsuccessful female ratees.

The second dependent variable measured the perceived benefit that resulted from expenditure incurred in carrying out the ratee's decision. The purpose of this measure is to investigate the extent of biases against women by examining whether raters' mental accounts are affected by their biases against women. The means (sd) for the man and woman ratees were not significantly different (87.82 [16.18] and 81.00 [13.17] respectively), and Kahneman and Tversky's (1979) suggestion that people's biases extend to their mental accounts was not supported and require further research. However, in competitive situations, it is possible that someone with a performance score of 87.82 points could be promoted over someone with 81.00 points.

The results of this study have three important implications for senior management. First, the results suggest the existence of a gender-based bias against successful women managers in performance evaluation. To the extent this affects promotion opportunities, it may cause frustration and lead to the premature exit of successfully performing women, to the detriment of the employing organisations. Second, discrimination in the workplace can result in high litigation costs for the employer both in legal fees and possible punitive damages. The results of this study identify one such source of discrimination. To avoid the potential resulting costs, senior management should monitor workplace performance evaluations and ensure they do not exhibit the gender-related bias identified in this study. Third, the solution that some organisations have adopted to help mitigate biases against women is to involve more women in appointment, promotion and pay decisions. Such involvement may require women to participate in group or individual decisions. The results of this study, suggest that this solution, while useful in creating a perception of fairness in the minds of women, may not achieve the desired results

Appendix A

Inventory Decision Case

Assume that you are Deputy General Manager-Retail for a medium sized company. Your company uses a budgeted sales system with currently attainable standards. You have several sales managers working for you. Each one of them runs his/her own retail store.

Mary Brown is one of those sales managers. Below is information regarding an inventory decision that she faced this year.

Mary's store reported a shortage of some inventory items this year. She had to decide whether to investigate changing the level of inventory they currently hold for these items. The cost of such an investigation was estimated at $5,000. Other estimates stated that the interest and warehousing costs of increasing current inventory levels of these items would be $15,000. The present value of future lost sales as a result of inadequate inventory would be $27,000 (after tax). Past data show that the probability that these items would run out (given the present shortage of these inventory items) is 50%.

The company's policy states that potential cost savings of $5,000 or more should be investigated. The potential cost savings in this case is $7,000, indicating that the investigation should take place.

During the investigation, Mary found that the levels of those items where shortages were reported, needed to be increased. This was carried out. However, due to an unexpected change in consumption behaviour of customers, the expected sales did not materialized and the inventory remained in the warehouse.

You, as Deputy General Manager-Retail, received a report from Mary detailing all of the costs and estimates involved in the inventory decision, along with her decision.

You are required to evaluate Mary's performance for the year. Please give a judgment or opinion on the scale below. Although Mary is not held accountable for the sales variance, the interest and warehousing costs and investigation decisions are her responsibility.

Mary's performance this period was (make a cross [X] on the scale below):

   0                50               100

Very   Below     Average   Above     Very
Poor   Average             Average   Good

In the situation I presented to you, the company spend $5,000 on an inventory investigation. Do you feel that the firm got something in return for that money? Did the $5,000 expenditure result in some benefit to the company? Please make a cross (X) on the scale below to indicate your opinion.

        0       50       100

No, there was        Yes, there
 no benefit          was a benefit

Appendix B

Unfavourable Variance Case

Assume that you are Deputy General Manager-Manufacturing for a medium sized company. Your company uses a standard costing system with currently attainable standards. You have several production managers working for you. Each one runs his/her own production plant.

Matthew Smith is one of those production managers. Below is information regarding a variance investigation decision that he faced this year.

Matthew's plant had a large direct labour efficiency variance this year. He had to decide whether to investigate the variance further. The cost of such an investigation was estimated at $4,000. Other estimates stated that the cost of correcting a problem would be $12,000. The present value of future costs of an uncorrected out-of-control system would be $22,000. Past data show that the probability that the system is out-of-control (given this large reported variance) is 50%.

The company's policy states that variances greater than 7% of standard costs allowed should be investigated by the manager (while those below this should not). The observed direct labour efficiency variance was 8% (unfavourable), indicating that the investigation should take place.

During the investigation, Matthew found that the variance had been caused by a problem in the production system. This problem was not temporary and would continue until corrected. He, therefore, corrected the identified problem.

You, as Deputy General Manager-Manufacturing, received a report from Matthew detailing all of the costs and estimates involved in the investigation decision, along with his decision.

You are required to evaluate Matthew's performance for the period [you may wish to refer to the previous page]. Please give a judgment or opinion on the scale below. Although he is not held accountable for direct material or overhead variances, the direct labour variances and investigation decisions are his responsibility.

Matthew's performance this period was (make a cross [X] on the scale below):

   0                50               100

Very   Below     Average   Above     Very
Poor   Average             Average   Good

In the situation I presented to you, the company spent $4,000 on a variance investigation. Do you feel that the firm got something in return for that money? Did the $4,000 expenditure result in some benefit to the company? Please make a cross (X) on the scale below to indicate your opinion.

        0       50       100

No, there was        Yes, there
 no benefit          was a benefit
Table 1
Descriptive Data of Experimental Participants

        Variables               Men        Women     Minimum   Maximum
                             means(sd)   means(sd)

Age (years)                    33.61       32.67      20         51
                              (8.39)      (8.11)
Working Experience (years)     10.64       9.22        0.33      30
                              (7.89)      (6.97)
Senior Managers (n)               8           5         --        --
Middle Managers (n)              17          15         --        --
Entry Level Employees (n)         8           4         --        --
Table 2
Mean (SDs) of Performance Evaluation Between the Successful and
Unsuccessful Performance Conditions (a)

                                 Ratee Sex

                          Man                 Woman           Total

Panel A:
    Successful
    Performance
    Condition
  Mean (sd)              76.32                66.79           71.47
                        (16.98)              (22.97)         (20.64)
                        n = 28               n = 29          n = 57
                   (15 men, 13 women)   (18 men, 11 women)

Panel B:
    Unsuccessful
    Performance
    Condition
  Mean (sd)              58.93                62.96           60.91
                        (22.01)              (20.82)         (21.34)
                        n = 29               n = 28          n = 57
                   (18 men, 11 women)   (15 men, 13 women)

Note: (a) The performance evaluation scale was anchored by `very poor'
(0) and `very good' (100).
Table 3
Means (SDs) of Perceived Benefit (a) From the Expenditure Incurred
in Carrying Out the Ratees' Decision

                                                 Ratee Sex

                                              Man      Woman     Total

Panel A: Successful Performance Condition
Mean (sd)                                    87.82     81.00     84.35
                                            (16.18)   (13.17)   (14.99)

Panel B: Unsuccessful Performance
  Condition
Mean (sd)                                    48.10     47.36     47.74
                                            (27.48)   (31.04)   (29.02)

Note: (a) This scale was anchored by `no, there was no benefit' (0)
and `yes, there was a benefit' (100).

(1.) Equity in human resource decisions implies similar employment, promotion and pay outcomes for both male and female ratees who perform similarly.

(2.) In such societies, women are conditioned from a very young age to believe themselves as having lower social status compared to men (Steinke 1998). At this age, children begin to mold their processing and organising of information about themselves along culturally and socially accepted norms (Bem 1993). Their view of the world is then seen along these sex-related schemata. These schemata carry over into adulthood and may lead women to respond to other women in a manner that is consistent with societal standards (Steinke 1998), and to attribute societal stereotypes to themselves (Park, Hayes & Foster 1994).

(3.) ANCOVA analyses indicated that these two variables did not have significant effects on any of the results.

(4.) These five covariates were included in all four ANCOVA models.

(5.) All tests of significance were one-tailed. Levene's test showed that the assumption of equality of variance was not violated in all four ANCOVA tests of hypotheses.

(Date of receipt of final transcript: October 2001. Accepted by Greg Clinch, Area Editor.)

References

Almer, E.D., Hopper, J.R. & Kaplan, S.E. 1998, `The effect of diversity-related attributes on hiring, advancement and voluntary turnover judgments', Accounting Horizons, vol. 12, no. 1, pp. 1-17.

Bem, S.L. 1993, The Lenses of Gender: Transforming the Debate on Sexual Inequality, Yale University Press, London.

Brown, V. & Geis, F.L. 1984, `Turning lead into gold: Evaluations of men and women leaders and the alchemy of social consensus', Journal of Personality and Social Psychology, vol. 46, pp. 811-24.

Cejka, M.A. & Eagly, A.H. 1999, `Gender-stereotypic images of occupations correspond to the sex segregation of employment', Personality and Social Psychology Bulletin, vol. 25, no. 4, pp. 413-23.

Choudhury, N. 1986, `Responsibility accounting and controllability', Accounting and Business Research, Summer, pp. 189-98.

Deaux, K. & Emswiller, T. 1974, `Explanation of successful performance on sexlinked tasks: What is skill for the male is luck for the female', Journal of Personality and Social Psychology, vol. 29, pp. 80-5.

Decker, W.H. 1987. `Attributions based on managers' self-presentation, sex, and weight', Psychological Reports, vol. 61, pp. 175-81.

Dubno, P. 1985, `Attitudes toward women executives: A longitudinal approach', Academy of Management Journal, vol. 1, pp. 235-39.

Eagly, A.H., Makhijani, M.G. & Klonsky, B.G. 1992, `Gender and the evaluation of leaders: A meta-analysis', Psychological Bulletin, vol. 11, pp. 3-22.

Foddy, M. & Smithson, M. 1989, `Fuzzy sets and double standards: Modeling the process of ability inference', in Sociological Theories in Progress: New Formulations, ed. J. Berger, M. Zelditch Jr., & B. Anderson, Sage, Newbury Park, CA, pp. 73-99.

Foschi, M. 1989, `Status characteristics, standards, and attributions', in Sociological Theories in Progress: New Formulations, ed. J. Berger, M. Zelditch Jr., & B. Anderson, Sage, Newbury Park, CA, pp. 58-72.

Garland, H. & Price, K.H. 1977, `Attitudes toward women in management and attributions for their success and failure in a managerial position', Journal of Applied Psychology, vol. 62, no. 1, pp. 29-33.

Glick, P. 1991, `Trait-based and sex-based discrimination in occupational prestige, occupational salary, and hiring', Sex Roles, vol. 25, pp. 351-78

Hull, R.P. & Hicks, D.W. 1993, `Language and gender: The pervasiveness of sexism in accounting and business publications', Advances in Public Interest Accounting, vol. 5, pp. 121-43.

Hull, R.P. & Umansky, P.H. 1997, `An examination of gender stereotyping as an explanation for vertical job segregation in public accounting', Accounting, Organizations and Society, vol. 22, no. 6, pp. 507-28.

Johnson, E.N., Kaplan, S.E. & Reckers, P.M.J. 1998, `An examination of potential gender-based differences in audit managers' performance evaluation judgments', Behavioral Research in Accounting, vol. 10, pp. 47-75.

Judge, T.A. & Ferris, G.R. 1993, `Social context of performance evaluation decisions', Academy of Management Journal, vol. 36, pp. 80-105.

Kahneman, D. & Tversky, A. 1979, `Prospect theory: An analysis of decisions under risk', Econometrica, vol. 47, pp. 262-91.

Kahneman, D. & Tversky, A. 1984, `Choices, values and frames', American Psychologist, vol. 30, pp. 341-50.

Lipe, M.G. 1993, `Analyzing the variance investigation decision: The effects of outcomes, mental accounting, and framing', The Accounting Review, vol. 68, no. 4, pp. 748-64.

Morrison, A. & Von Clinow, M. 1990, `Woman and minorities in management', American Psychologist, February, pp. 200-08.

Nielsen, R.P. & Bartunek, J.M. 1996, `Opening narrow, routinized schemata to ethical stakeholder consciousness and action', Business & Society, vol. 35, no. 4, pp. 483-519.

Paludi, M.A. & Strayer, L.A., 1985, `What's in a name? Differential evaluations of performance as a function of author's name', Sex Roles, vol. 12, pp. 353-61.

Park, J.J., Hayes, R.S. & Foster S.F. 1994, `Men and women: Equal in accounting?', Journal of Education for Business, vol. 69, no. 6, pp. 349-54.

Ridgeway, C.L., 1993, `Gender, status and the social psychology of expectations', in Theory on Gender/Feminism on Theory, ed. P. England, Aldine, New York, pp. 175-97.

Ridgeway, C.L. & Johnson, B. 1990, `Expectations, legitimation, and dominance behavior in task groups', American Sociological Review, vol. 51, pp. 603-17.

Rosch, E. 1978, `Principles of categorization', in Cognition and Categorization, ed. E. Rosch & B.B. Lloyd, Erlbaum, Hillsdale, NJ, pp. 27-48.

Schein, V.E. 1973, `Relationships between sex roles stereotypes and requisite management characteristics among female managers', Journal of Applied Psychology, vol. 60, no. 3, pp. 340-44.

Schweitzer, M.E. 1999, `The construction of mental accounts in benefits decision making', Benefits Quarterly (First Quarterly), pp 52-5.

Shore, T.H. 1992, `Subtle gender bias in the assessment of managerial potential', Sex Roles, vol. 27, no. 9/10, pp. 499-515.

Steinke, J. 1998, `Connecting theory and practice: Women scientist role models in television programming', Journal of Broadcasting & Electronic Media, vol. 42, no. 1, pp. 142-51.

St. Pierre, R.N.M., Herendeen, D.S., Moore & Nagle, A.M. 1994, `Does occupational stereotyping still exist?' The Journal of Psychology, vol. 128, no. 5, pp. 589-98.

Swim, J. 1993, `In search of gender bias in evaluations and trait inferences: The role of diagnosticity and gender stereotypicality of behavioral information', Sex Roles, vol. 29, no. 3/4, pp. 213-37.

Todor, M.L. 1980, `The effect of sexual composition of a group on discrimination against women and sex-role attitudes', Psychology of Women Quarterly, vol. 5, pp. 292-310.

Tversky, A. & Kahneman. 1981, `The framing of decisions and the psychology of choice', Science, vol. 211, January, pp. 453-58.

Janne Chung, Schulich School of Business, York University, 4700 Keele St, Toronto ON M3P 1P3, Canada. Email: [e-mail]

The author thanks the editor, Vincent Chong, Denise Patterson, seminar participants at Edith Cowan University and attendees at the AAA Annual Meeting 1999 for their helpful comments.