Keywords

Introduction

Depression is a common mental disorder that is characterized by depressed mood, loss of interest or pleasure, feelings of guilt or low self-worth, disturbed sleep or appetite, low energy, and poor concentration. It is the leading cause of disability and the second biggest contributor to the global burden of disease. One in five people experiences depression at some stage in their life and the majority of those affected are working adults (WHO 2009). This represents a significant problem for organizations, with impacts including an average of 3–4 days off work per month for those diagnosed with depression, impaired job performance while attending work (presenteeism), and a greater vulnerability in the employee to other illnesses (Cocker et al. 2011; Haslam et al. 2005; Murphy et al. 2006; Caruso and Myette 2008).

These impacts can be reduced by effective treatment. Unfortunately, approximately 60 % of those with depression do not seek the help that is critical to recovery and management of their condition (Avey 2005). In the workplace, managers are well positioned to provide both task-related and emotional support to employees with depression. However, research shows that depressed employees are not likely to seek help from managers for fear of stigmatization and that those who do disclose their condition can experience a lack of understanding and support (Haslam et al. 2005). The stigma of mental illness is a major barrier to achieving effective management of mental health issues in workplaces (Szeto and Dobson 2010). In particular, stigmatizing attitudes among managers may limit the degree to which they engage in supportive behavior such as promoting help-seeking, negotiating job accommodations, and facilitating effective return to work following mental health-related absences. These issues are particularly complex, considering that depression can be considered an ‘invisible disability’ and the dynamics of disclosure and privacy are important (Martin and Fisher 2014). There is also significant potential for stigma to result in bias and discrimination in processes associated with performance appraisal and career development (Perez and Wilkerson 1998).

Despite a growing body of literature on mental health and the workplace, the critical role of managers in developing effective responses to the problems outlined above requires further research (Martin et al. 2015). In order to help inform more effective management training and organizational development strategies (Szeto and Dobson 2010), we investigate antecedents and outcomes of depression stigma, with a particular focus on the role of contextual cues in a simulated work environment, in relation to how an employee with depression is perceived.

Managers’ Stigmatizing Attitudes Toward Depressed Employees

Stigma broadly refers to “beliefs, attitudes, and behaviors that result in social rejection” (van Dorn et al. 2005, p. 153). In relation to mental illness, stigma can decrease the likelihood of workplace support (Barney et al. 2006). Martin (2010) examined managers’ stigma toward employees with depression, finding three forms of stigma: affective (emotional distance or a dislike toward them), cognitive (negative beliefs about them), and behavioral (intentions to behave in a discriminatory way toward them). She found that both individual and contextual factors were associated with managers’ self-reports of these forms of stigmatizing attitudes. Although there are implications of these results regarding individual manager characteristics for organizational support and training, the influence of contextual characteristics was noted as a priority area for broader, more pervasive organizational development. To date, how these prejudicial attitudes relate to managerial decision-making processes influencing performance appraisal and promotion decisions is yet to be examined empirically.

The current study continues investigation of the influence of individual characteristics upon stigma, and extends empirical investigation to investigation of the relationship between stigma and performance appraisal. Furthermore, the study improves on prior literature by examining the influence of the contextual environment on these relationships with an alternative methodology to the cross-sectional surveys that typify this area of research. We provide a brief theoretical rationale for a series of testable hypotheses below.

Organizational Support for Mental Health

Workplaces can foster effective mental health management strategies through the development and maintenance of a supportive organizational environment. In Martin’s (2010) study, an important aspect of organizational context associated with managers’ stigmatizing attitudes toward depressed employees was the presence of a mental health strategy (the extent to which the organization has mental health policies, procedures, and training programs). While mental health literacy programs show evidence that they can decrease stigmatizing attitudes among the working population (Kitchener and Jorm 2004) the broader influence of organizational contextual factors on the stigmatizing attitudes of managers toward depressed employees and their associated decision-making processes reflects a significant deficit in the literature.

The construct of organizational climate represents fertile ground for examining how context relates to attitudes toward employee mental health. Organizational climate reflects ‘shared understandings’ about priorities, procedures, and practices within an organization either generally or in relation to a specific ‘facet’ (Schneider 1990). In field studies, climate is usually operationalized as a high level of agreement among work group members about the facet of interest. Although it was an exploratory first attempt to study the phenomenon of interest, Martin’s (2010) study utilized a cross-sectional survey design and was thus subject to a potential validity threat related to common method variance (Podsakoff et al. 2003). Contextual characteristics that were termed “indicators of a mental health climate” were self-reported and reflected participants’ individual perceptions of varied organizational environments (managers from more than 200 different organizations). To address these limitations and extend theoretical development in this important new area of research, we employ a management simulation that allows context to be manipulated as an experimental variable and separate it from the measurement of other self-reported variables, allowing for potential causal processes concerned with context and individual attitudes and decision-making to be examined. However, as this does not allow us to examine climate as the ‘shared’ understanding of organizational members and as such, we herewith refer to the contextual variable in this study as Organizational Support for Mental Health (OSMH). Our approach aims to build evidence about whether organizations’ attempts to signal support for mental health by communicating elements of a strategy designed to proactively deal with mental health and well-being of their employee, may lead to a reduction in stigmatizing attitudes toward employees with depression among its managers.

Hypothesis 1

Individuals who receive contextual cues that their organization is unsupportive toward mental health will report higher levels of affective, cognitive, and behavioral stigma toward a depressed employee, than those who receive cues that their organization is supportive toward mental health.

Martin (2010) provided an extensive rationale for the relationships between individual differences and managers’ attitudes toward employee depression. Her study showed that less depression experience, an internal locus of control, higher levels of stress, and greater reticence to seek help for personal problems were associated with a greater propensity among managers to stigmatize depressed employees. She suggested that depression experience (gained either through their own experiences of depression or through contact with an employee or significant other), increases empathy and knowledge, resulting in lower stigma. Managers with an internal locus of control were seen as more likely to ‘blame’ the employee for their condition and those reporting higher stress may see employees with depression as an additional burden, causing them further problems. Managers who feel uncomfortable seeking help for personal issues often have stoic beliefs that one should cope with their problems without involving others and keep mental health issues to themselves, thereby judging others with mental health issues more harshly.

Hypothesis 2

Individuals who have less experience with depression, an internal locus of control, higher levels of stress, and greater help-seeking reticence will report higher levels of affective, cognitive, and behavioral stigma toward a depressed employee.

Martin (2010) encouraged researchers to help build theory in this area through systematic investigation. For example, she suggested examining the way individual and organizational characteristics might interact to influence the expression of such attitudes. The experimental design employed in the present study also allows further contribution to theoretical advances in our knowledge of depression stigma, by allowing an examination of mental health climate as a moderating situational factor rather than a simple direct effect variable. By examining the combined-interactive influence of individual characteristics and mental health climate on stigma, we apply an interactionist perspective (Terborg 1981). Hence, while a manager may be predisposed to have negative attitudes toward depressed employees because of a range of personal characteristics, the presence of a strong social context that is supportive of employee mental health may diminish the strength of these relationships and ultimately alter attitude–behavior relations.

Hypothesis 3

Organizational support for mental health (OSMH) will moderate the relationships between individual characteristics and managers’ stigma, such that the associations between these individual characteristics and stigma will be reduced when OSMH is supportive.

Depression Stigma and Performance Appraisal

Another area for research development identified by Martin (2010) was in relation to the criterion validity of the measure of managers’ stigmatizing attitudes such as whether outcomes of importance to individuals and organizations vary with the different dimensions of stigma. She suggested that managers with high levels of stigma might be more prone to exhibiting bias in processes such as performance appraisal resulting in lower performance appraisals of depressed employees. While empirical evidence on this is lacking, stereotype biases have been demonstrated in numerous other areas. Examples include age (Rupp et al. 2006), pregnancy (Halpert et al. 2006), and gender biases (Lyness and Heilman 2006), where supervisors rate employees more poorly based on these personal characteristics. It has also been observed that supervisors who hold prejudiced attitudes toward an employee give less weight to work behaviors than to personal characteristics in decision-making (Favero and Ilgen 1989). Fiske (1998) suggests that supervisors halt their search for performance-related information when presented with findings that confirm their expectations or stereotypes. For example, knowing that an employee is depressed may lead a supervisor to focus on situations in which performance has been poor, rather than considering all available performance and contextual information. This can lead to negative performance appraisals that are inconsistent with all available information, or do not take into consideration other factors that may be impacting on performance, such as the availability of resources needed to complete the job. Managers may make internal attributions for poor performance (e.g., because the individual is depressed) and external attributions for good performance (e.g., depressed individuals must have gained help from others). Such justifications and attributional biases may contribute to more negative appraisals of depressed employees (Martin and Fisher 2014).

Hypothesis 4

Stigma toward a depressed employee will be negatively associated with performance ratings, such that individuals who exhibit higher levels of affective, cognitive, and behavioral stigma will report lower performance ratings of a depressed employee than individuals with lower levels of stigma.

Context is also an important factor impacting performance ratings. Crandall and Eshleman (2009) model of prejudice proposes that individuals will express prejudicial responses when they feel that these responses are justified. Brief et al. (2000) showed this effect with black job applicants. Participants who received a business justification for discriminating against black job applicants were less likely to select black applicants, indicating that the presence of a justification is a significant predictor of prejudice. Similarly, we expect that a low OSMH context may further activate the influence of depression stigma on performance ratings, reducing the potential for leniency that may be associated with a more supportive or benevolent environment.

Hypothesis 5

Organizational support for mental health (OSMH) will moderate relationships between stigma and performance ratings, such that the associations between stigma and poorer performance ratings will be stronger when OSMH is lower.

Method

Participants

Three hundred and forty-eight undergraduate psychology students were recruited from an Australian university. The mean age of participants was 21.00 years (SD = 4.16, range 17–49 years). 71.2 % were female, 52.5 % were currently employed, and 33.2 % had previous management experience. Participants were self-selected and were remunerated with $10 (80.7 % of sample) or course credit (19.3 % of sample). There were no exclusion criteria and the recruitment method does not allow comparison of participants and nonparticipants.

Procedure

Participants entered the room and provided with an experiment code (e.g., 1, 2, 3) randomly allocating them to one of two conditions (organization = supportive or unsupportive, n = 174 in each group). The experimenter explained that the aim of the study was to investigate how managers make decisions, and that they were required to assume the role of a call center manager and complete three managerial tasks. All participants signed a consent form stating that the study had ethical clearance, participation was voluntary, responses were anonymous and participants could withdraw at any time.

Participants completed a pre-study questionnaire containing items regarding demographic information and measures of the independent variables in the study (personal characteristics). After participants completed the pre-study questionnaire, they received a document entitled ‘Job Description,’ and an accompanying instructional booklet of three managerial tasks, entitled ‘Job 1: Rating Job Applicants,’ ‘Job 2: Responding to a Memorandum,’ and ‘Job 3: Performance Appraisals.’ These tasks comprised a work sample of managerial jobs and served three purposes. First, it helped enhance the realism of the experiment as participants became immersed in realistic job activities. Second, it enabled the manipulation of the contextual cues to reflect OSMH. Finally, it enabled the assessment of mental health stigma toward an ‘employee’ within the ‘company’ and the assessment of their ‘performance.’ Participants were asked to assume the role of a manager and told that their documents contained all of the necessary information about the ‘company,’ their job history/responsibilities and the three tasks to be completed. Approximately 40 min were allocated to task completion. Self-pacing was encouraged as it would be consistent with the way a real manager would complete such jobs at work.

The first job, ‘Job 1: Rating Job Applicants,’ required participants to read résumés of three job applicants and rate the applicants’ employment suitability. This task was unrelated to the experiment other than to prime the participant into thinking about themselves as managers in the simulated context.

OSMH was manipulated in the second managerial job, ‘Job 2: Reading and Responding to a Memorandum.’ Participants read an interoffice memorandum received from the CEO, describing either that the company was performing well or badly in supporting the mental health and well-being of employees. To increase perceived objectivity, accuracy, and representativeness of the information, the CEO relayed summary results from an employee survey. In both conditions, the memorandum began by describing some of the positive results from the survey.

OSMH was manipulated in the subsequent paragraphs of the memorandum. These paragraphs relayed additional survey results that reflected employee perceptions that the company was either supportive or unsupportive toward mental health and well-being. In the supportive condition, employee opinion survey results showed strong agreement on a series of statements relevant to mental health in the organization (e.g., reflecting various high levels of agreement such as 85, 90, and 95 %). These statements were based on elements of employee welfare, mental health strategy and depression disclosure norms (Martin 2010), and reflected themes including how much the organization cares for employees, the clarity of procedures and supports available, how confident employees are discussing mental health issues with their supervisors, and how confident supervisors were in their mental health management skills (extracts of study materials are available upon request). In the unsupportive condition, the memorandum described the results of the survey as being negative in relation to these themes, with low levels of endorsement of those same statements (e.g., 15, 10, and 5 % agreement).

Additional information was then given to participants to further enhance OSMH manipulation. The statement in the supportive condition read “The results of the employee survey do not surprise you. At Aussie Mobile it is considered quite appropriate to discuss mental health problems like depression.” The corresponding statement in the unsupportive condition read “The results of the employee survey do not surprise you. At Aussie Mobile it is not considered appropriate to discuss mental health problems like depression.” Participants were then asked to summarize the memorandum in preparation for a future staff meeting, as a way of ensuring that they had read and thoroughly processed it.

The third job, ‘Job 3: Conducting Performance Appraisals,’ required participants to read short performance descriptions of three current employees, rate several aspects of their performance on a four-point scale, and write a summary of each employee’s performance. One of the employees, ‘John,’ was described as an employee who had been “recently diagnosed with depression.” This description was a modified version of a vignette used by Wolkenstein and Meyer (2008). After participants completed the third job, a post-study questionnaire was administered to gather manipulation and realism check data and they were debriefed.

Measures

OSMH Manipulation

Mental health strategy. The statements used to evaluate the success of the manipulation of mental health strategy were adapted from Martin (2010). Seven statements were rated on a scale from one (disagree strongly) to seven (agree strongly). An example statement is “Managers at Aussie Mobile would know what to do if an employee has a problem with depression.” The items made a reliable scale (α = 0.93).

Depression disclosure norms. The statements used to evaluate the success of the depression disclosure norms manipulation were adapted from Martin (2010). Eight statements were evaluated on a scale from one (disagree strongly) to seven (agree strongly). An example statement is “Employee depression is generally considered a suitable topic for discussion at Aussie Mobile.” The items made a reliable scale (α = 0.91).

Welfare dimension. The check for the welfare manipulation was adapted from the welfare subscale of Patterson et al. (2005) organizational climate measure. Participants rated four statements on a seven point scale, from one (disagree strongly) to seven (agree strongly) e.g., “Aussie Mobile cares about its employees.” The items made a reliable scale (α = 0.90).

Measures of experiment realism. Six items were developed by the study authors to measure the realism of the experiment. An example item is “I really felt as though I was a contact centre manager for Aussie Mobile.” The items made a reliable scale (α = 0.82).

Individual Characteristics

Depression experience. The items measuring experience with depression were developed by Martin (2010). One item measured personal experience with depression (“Have you ever been diagnosed with depression?”). Two items measured experience with a significant other experiencing depression (“Have any of your close friends or family ever been diagnosed with depression” and “Have you ever managed or supervised an employee who disclosed that they had been diagnosed with depression or who you seriously believed was suffering from depression?”). All three items were scored using the response options, yes (1) or no (0). Scores were coded as a dichotomous variable, indicating either no experience or some experience with depression.

Symptoms of stress were assessed with the 7-item stress subscale of the brief version of the Depression Anxiety and Stress Scale (Henry and Crawford 2005). The items used a scale ranging from one = “did not apply to me at all,” through to four = “applied to me most of the time” to assess the presence of symptoms such as “I find it difficult to relax.” The items made a reliable scale (α = 0.85).

Locus of control was assessed with the Valecha and Ostrom (1974) 11 item scale in which paired items are presented and one point is allocated for each item reflecting an internal locus of control, e.g., “people who can’t get others to like them don’t understand how to get along with others.” The scale was only marginally reliable (α = 0.65).

Help-seeking reticence. The measure for help-seeking reticence was based on the attitudes toward seeking professional help scale (Fischer and Turner 1970). Participants evaluated four items on a scale from one (disagree strongly) to seven (agree strongly). A sample item was “I find it difficult to talk about personal affairs with people such as doctors, teachers and clergymen.” The scale was only marginally reliable (α = 0.66).

Stigma

Martin’s (2010) measurement of managers’ affective, cognitive, and behavioral stigma was adapted so that participants rated their attitudes toward a specific employee, named John. Items were rated on a scale from one (disagree strongly) to seven (agree strongly). Sample items were: “It would make me feel awkward working alongside John” (affective); “It is John’s own fault that he is suffering from depression” (cognitive); and “I would be prepared to make temporary changes to the job to help John’s recovery process” (behavioral). All stigma scales had an acceptable level of reliability (α = 0.71; 0.83; 0.70, respectively).

Performance Rating

Participants rated the performance of all employees including ‘John’ using 5 items developed specifically for the current study based on an examination of call center industry performance evaluation standards such as “Meets benchmarks for call handling time.” Each of these competencies was rated on a scale where one = Must improve, and five = Exceptional. A composite performance appraisal score was created by taking the mean of these items. The items formed a reliable scale (a = 0.84).

Results

Preliminary Checks for Manipulation and Experimental Realism

Manipulation checks were performed on the data to determine if the manipulation of OSMH was successful. Independent groups t tests revealed a statistically significant difference between the supportive (M = 5.78, SD = 0.82) and unsupportive conditions (M = 3.08, SD = 1.26) on the mental health strategy check, t(221) = 18.96, p < 0.001. Similarly, there was a significant difference between the supportive (M = 5.64, SD = 0.82) and unsupportive conditions (M = 3.20, SD = 1.00) on the depression disclosure norm check, t(221) = 19.83, p < 0.001. The welfare norm manipulation check was also successful, t(221) = 13.22, p < 0.001, with those in the supportive condition (M = 6.05, SD = 0.79) scoring the organization higher on employee welfare than the unsupportive condition (M = 4.32, SD = 1.13). Participants agreed the experiment was realistic (M = 5.21, SD = 0.98). Participants in the supportive (M = 5.17, SD = 1.04) and unsupportive conditions (M = 5.24, SD = 0.92) found the experiment to be equally realistic, t(221) = −0.57, p = 0.567.

Data Analysis

Next, data were prepared for analysis and examined to ensure the statistical assumptions for regression were met. No violations of the assumptions were found. Descriptive statistics for the study variables are provided in Table 15.1. A review of the Pearson’s correlations indicated acceptable collinearity (<0.9) between the variables and only moderate correlations among the dependent variables indicating the discriminant validity and utility of separate regressions for the three dimensions of stigma. Prior to calculation of the product terms, all independent variables were centered (as recommended by Aitken and West 1991). According to our power calculations, the sample size was adequate for testing the proposed models (Tabachnick and Fidell 2007). There were no statistically significant differences in any of the study variables for the two experimental groups.

Table 15.1 Descriptive statistics and intercorrelations of the study variables
Full size table

Moderated multiple regression analyses were conducted with a hierarchical model in which control variables were entered (age and gender) along with the independent variables in the first block, and the cross products, or interaction effects were entered in the second block. To test Hypotheses 1, 2, and 3 the independent variables (personal characteristics and OSMH condition) and interactions among them were regressed on the 3 types of stigma. To test Hypotheses 4 and 5, the independent variables (3 stigma types and OSMH condition) and interactions among them were regressed on the performance ratings (Table 15.2).

Table 15.2 Hierarchical regression results for stigma: standardized beta (std error)
Full size table

The first regression model explained approximately 9 % of the variance in participants’ affective stigma toward the employee with depression. The results showed that age was significantly associated with affective stigma, with older participants reporting higher levels of affective stigma. Locus of control was also associated with affective stigma, with participants scoring higher on internal locus of control more likely to report stigmatizing attitudes. OSMH was not significantly associated with affective stigma. While an interaction between OSMH and depression experience explained an additional 2 % of the variance in affective stigma and showed a significant Beta, the second step of the model was not associated with a significant F change and was therefore not able to be interpreted.

The second regression model explained approximately 17 % of the variance in managers’ cognitive stigma toward the employee with depression. Sex was associated with cognitive stigma with women reporting lower levels of cognitive stigma than men. Main effects of locus of control, depression experience, help-seeking reticence, and mental health climate were observed, all in the hypothesized directions. A significant interaction between OSMH and help-seeking reticence was observed. The presence of this moderation effect was indicated by a significant change in r 2 as a result of the addition of the interaction term in the regression equation (3 %). Simple slopes analyses plotted in Fig. 15.1 enabled this effect to be interpreted. The slope was significant for the unsupportive OSMH condition with higher help-seeking reticence associated with more cognitive stigma, but the relationship between help-seeking reticence and cognitive stigma was not significant in the supportive condition.

Fig. 15.1
figure 1

Slope test results for help-seeking reticence, cognitive stigma and organizational support for mental health

Full size image

The third regression model explained approximately 12 % of the variance in behavioral stigma toward the employee with depression. Age and sex were positively associated with stigma, with women and younger people reporting lower levels of stigma. Main effects of depression experience and help-seeking reticence were observed in the direction hypothesized. OSMH, nor the interaction between the condition and any of the personal characteristics, were significant.

As shown in Table 15.3, the fourth regression model explained approximately 8 % of the variance in participants’ performance ratings of a depressed employee. Age was negatively associated with performance ratings, indicating that older participants rated the performance of the depressed employee more negatively. Affective stigma was significantly negatively associated with performance ratings but cognitive and behavioral stigma was not. While an interaction between OSMH and affective stigma explained an additional 2 % of the variance in performance and showed a significant Beta, the second step of the model was not associated with a significant F change and was therefore not able to be interpreted.

Table 15.3 Hierarchical regression results for performance appraisal
Full size table

Discussion

This study primarily aimed to provide a controlled examination of the effects of organizational context on relationships of interest regarding managerial stigmatizing of employees with depression. Partial support for four of the five hypotheses was found in the study. In relation to the first hypothesis, there was some evidence that OSMH was associated with cognitive stigma. This result highlights the importance of communicating OSMH to managers, for the positive benefits it has on reducing cognitive stigma toward depressed employees. The lack of an effect on the other stigma types was unexpected. As differences between the two OSMH conditions were only observed for cognitive stigma, it suggests that such signals may encourage beliefs about employees with depression to be modified, but not emotional reactions or behavioral intentions.

Some support was found for the second hypothesis regarding the relationships between individual characteristics and stigma. It was expected that depression experience would be negatively associated with stigma. The results showed a relationship between depression experience and cognitive and behavioral stigma, echoing Martin’s (2010) results for these same dimensions. This is consistent with the idea of gaining understanding about depression through ‘contact’ with someone who has direct experience of the condition, suggesting that it may be useful to design workplace training in ways that simulates this ‘experience.’ Also as predicted, individuals higher in help-seeking reticence reported more cognitive and behavioral stigma. Martin (2010) also found such an association for behavioral and cognitive stigma. These results suggest that those with a willingness seek help themselves may have more accurate thoughts about depression, and hence be more accepting, less judgemental, and more open to providing help to an employee with depression. Hence, normalizing help-seeking among managers may be an important mental health promotion strategy.

Hypothesis 3 predicted that OSMH would moderate the association between individual characteristics and affective, cognitive, and behavioral stigma, such that the individual predisposing characteristics would not be as strongly associated with stigma in the high OSMH condition. The results show partial support for this aspect of the third hypothesis regarding the moderating role of OSMH in the relationship between the participant characteristic of help-seeking reticence and their levels of cognitive stigma. This means that when OSMH was high, the extent to which participants were reticent to seek help for personal concerns was not associated with cognitive stigma. Conversely, when OSMH was low, managers’ personal characteristics played a role in influencing their beliefs about employees with depression. The failure to detect similar moderating effects for affective and behavioral stigma was unexpected and inconsistent with organization behavior theories that suggest that organizational climate has a moderating effect in relation to the relationship between individual characterizes and attitudes/behavior (Parker et al. 2003).

The cognitive stigma measure was more strongly associated with the other study variables. This finding mirrors Martin’s (2010) results showing that managers whose organizations had a clear mental health strategy reported lower levels of stigma but only for the cognitive dimension. Given our study aimed to provide a more stringent test of this relationship, we can have increased confidence in the importance of organizational context for the beliefs people hold about employees with depression. Behavioral stigma may be less likely to be reported in organizations given legal frameworks around discrimination, and hence there is less variance in this measure. Affective stigma may be more difficult to understand and target given the often subconscious nature of emotions and the fact that there is less research on affective stigma to draw upon. In general, the affective domain of attitudes has received far less attention in the research literature than the cognitive domain because it has been difficult to define and difficult to measure (Bolin et al. 2005).

In relation to the fourth hypothesis, there was some evidence that affective stigma was associated with lower performance ratings of the depressed employee. Research suggesting the role of affect and positive interpersonal relationships in performance appraisals (Lefkowitz 2000) may be one explanation for why affective stigma was found to be associated with performance ratings. The lack of a finding for the impact of cognitive and behavioral stigma on performance ratings was surprising given that previous research indicates that managers rate employees performance based on perceived negative personal characteristics (Halpert et al. 2006; Lyness and Heilman 2006; Rupp et al. 2006). No support was found for the fifth hypothesis, as OSMH, nor an interaction between it and the stigma types was associated with performance ratings. This finding is inconsistent with our theoretical predictions.

Implications for Workplace Mental Health Promotion

The findings highlight the importance of both individual characteristics and contextual factors in understanding stigma toward depressed employees and suggest that organizations need to be targeting both aspects in any mental health promotion interventions. In relation to individual differences, it was found that experience of depression is associated with less stigmatization of individuals with the illness. Watson and Corrigan (2005) suggest that while contact with persons who have experienced depression may be one of the best ways of reducing stigma, its implementation may prove more difficult than a widespread education program. There are ways however, to implement contact more easily in workplace training. For example, Crisp and Turner (2009) suggest that imagined contact is sufficient to reduce negative attitudes and Cameron and Rutland (2006) suggest that reading stories about disability can be helpful form of simulating ‘contact.’ Research evaluating the impact of different intervention strategies on the three types of stigma explored in the present study will be vital in ensuring research continues to inform practice.

The reluctance of participants to seek help for psychological issues was associated with how they perceived employees with depression. Consequently, if the willingness of managers to seek help was increased, this might lessen the stigma attached to employees with depression. This could be done through manager training on how to seek help, and realize that as a manager, it is acceptable (and indeed advantageous) for them to be seen by employees to seek help.

Another pertinent aspect of the findings is that OSMH was an important factor influencing cognitive stigma in both a direct and indirect manner. Organizations should aim to specifically to communicate this support by developing an integrated approach to mental health issues that includes prevention strategies (LaMontagne et al. 2014), assistance programs, mental health awareness training, and normalizing discussion of mental health issues. The importance of strong leadership of such initiatives is also critical and this has been recognized recently as central to creating ‘psychosocial safety climate’ (Dollard and Bakker 2010), a construct that is similar to our focus on OSMH.

Although not explicitly part of the hypotheses tested, participant age and gender included as control variables, displayed significant associations with components of stigma. Specifically, females reported less cognitive and behavioral stigma than did males and younger participants reported less affective and behavioral stigma than did older participants. This trend is consistent with previous research investigating depression stigma (Griffiths et al. 2008; Jorm et al. 1999; Martin 2010). These effects may have important implications in the workplace. Specifically, male or older managers may need more stigma reduction focussed training. The age and gender of the manager could also interact with the age and gender of the employee with depression. In the current study, the employee with depression was male, but stigma may differ if a female employee was described. For example, research has noted that male manager/male employee teams elicit more frustration than a female manager/male employee combination (McColl-Kennedy and Anderson 2005). In regard to depressed employees, this frustration may manifest through more stigmatizing attitudes.

Finally, the influence of affective stigma on performance ratings might indicate that this relatively unexplored form of stigma might be important to counter by better supporting managers in how to deal with performance issues among employees with depression (Martin et al. 2015). The complexity of factors influencing performance judgements is only further complicated by the presence of a mental health condition with known performance impacts (Haslam et al. 2005).

Limitations and Directions for Future Research

There are limitations to the study that need to be discussed. Although the quasi-experimental methodology was a strength in relation to theoretical development, it does limit the external generalizability of the findings. Although laboratory studies are common in organizational behavior and the simulation was designed to be as close to a manager’s role as possible to counteract threats to ecological validity, there are several limitations and potential problems with the methodology we employed. First, while the significant differences between the conditions in terms of the manipulation checks suggest the conditions were distinct, some of the ratings in the unsupportive climate condition are still close to or above the median point in the scale, and rather than a clearly dichotomous supportive versus unsupportive condition, the manipulation could be considered to have manipulated highly supportive versus neutral. There is also a possibility that the manipulation introduced cognitive priming or a mood induction rather than a true manipulation of context. An experimentally simulated contextual manipulation shares some similarity with the concept of mood induction where participants are exposed to various types of media, feedback, or instruction to imagine affective states and a manipulation check measuring the mood in question is used to determine the success of the mood induction (Polivy 1981). However, our manipulation of OSMH was not intended to induce mood. Rather we presented factual information regarding survey results that simulate environmental information participants could use to form cognitive impressions about what it would be like to work in this organization. This approach is similar to that used by Ziegert and Hanges (2005), who utilized a presidential memo as method of manipulating organizational context for racial bias. Our stimulus materials attempted to clearly signify very low levels of OSMH using very low percentage agreement figures for survey results in the memo.

Although the use of vignettes has a long tradition in social attitude research (Burstin et al. 1980), the organizational information we presented in the vignette was by no means identical to the long-term nature of an organizational environment, nor the relationships between managers and subordinates. In addition, conducting performance ratings for ‘paper people’ raises issues of generalizability given real performance is rated on many, many samples of behavior over a long period of time. In particular, the vignette used to prompt the assessment of depression stigma, the target employee was described as depressed and a number of symptoms associated with the condition were presented (including a drop in work performance). This vignette was based on a previously validated vignette about depression (Wolkenstein and Meyer 2008) and as such it was framed as a stigma study and not a performance appraisal study. Results may have been different if an employee with depression was described as maintaining strong performance—although the literature around presenteeism and depression tells us that reduced performance is a common outcome of depression and accounts for most of the economic costs borne by employers (Cocker et al. 2011).

Sample characteristics must also be noted. As the sample contained psychology students, it may be that lower levels of mental health stigma than in the general population may have contributed to the results. However, studies have shown that even qualified mental health professionals are not immune to holding stigmatizing attitudes (Jorm et al. 1999). Approximately 33 % had management experience and 52 % were employed. For those that were not experienced managers, we argue that vicarious experience offered through workplace, family, and community social learning opportunities would provide sufficient experience to help them ‘assume’ the role of a manager. Nonetheless, one should exercise caution in generalizing the findings to the workplace.

A minor limitation was the failure of two of the scales to reach acceptable levels of reliability (Locus of control and Help-seeking reticence were both marginally below 0.70). More substantive limitations relate to the research design itself.

Future research is urgently needed exploring the links between employee depression and performance appraisal/performance management. Examining the role of justification, or ‘excuses,’ for performance appraisal decisions may also be illuminating. Justifications for racial bias that have been studied are ‘even innocuous’ past mistakes (Knight et al. 2003) and ‘business justifications’ regarding racial homogeneity of customers and staff (Brief et al. 2000). Such justifications may interact with personal biases to contribute to more negative appraisals of depressed employees.

While our results provide some indication that OSMH may lead to a reduction in cognitive stigma, it should be recognized that to completely evaluate a causal relationship, a longitudinal panel-designed intervention-based field study is required. Such a study would allow for the reverse causal relationship (i.e., widespread stigma can cause poor OSMH) also to be evaluated.

Multi-level studies examining the variables of interest at the level of employees, managers and organizations would be a useful future research objective. In this type of investigation, it would be possible to obtain measures of OSMH from groups of employees to examine the extent to which these perceptions are shared. Managerial stigma could also be measured from the perspective of employees. Research could also examine any variation in employee outcomes associated with the different types of stigma such as perceived discrimination, relationship quality, and career advancement. Similarly, at the organizational level, various facets of climate could be associated with mental health attitudes and relevant outcomes.

Finally, regarding our findings on age, gender, and stigma, we suggest future research could counterbalance vignettes with a male and a female with depression to investigate such an interaction effect. Similarly, different age combinations in managers and subordinates could also be important given research on age discrimination. The assessment of depression experience could also be further improved by including more information about how and when this ‘experience’ was gained (e.g., current, recent, treatment history, etc.).

Despite the evidence that our attempt to manipulate of the extent to which an organization was supportive of mental health was successful and realistic, future research should attempt to confirm these findings in the field. We suggest that building theory from both experimental and field studies incrementally will allow us to systematically explore this novel research agenda.

Conclusion

Employee depression has many challenges and costs, some of which are amplified if the work environment is not supportive. It is vital that we understand the antecedents of depression stigma in the workplace so that better support can be provided to employees with depression and barriers to effective treatment, employee retention, and productivity can be reduced. The quasi-experimental design of the study afforded examination of a potentially causal relationship between the level of organizational support for mental health and stigma. The results of this study have implications for workplace depression awareness training and organization development strategies that foster a work environment that emphasizes supportive management of employee mental health problems.