Mellor, D.J., Moore, K.A., & Wall, C.L. (2012). Mapping the linear relationships among stressors, mediators, and coping. In K.A. Moore, P. Buchwald, & T. Ringeisen (Eds.) Stress and organisational factors. Berlin: Logos Publishers. ISBN 978-3-8325-2886-7
Linear Relationships among stressors, mediators, and coping David J. Mellor School Psychology Deakin University, Burwood 3196 Australia Kathleen A Moore School of Psychology Charles Darwin University, Darwin NT 0909 Australia Cindy Wall School Psychology Charles Darwin University, Darwin NT 0909 Australia Email:
[email protected] Abstract Occupational stress is a principal workplace concern because of the deleterious effects it can have both for employees and the organisation. Much research has investigated the impact of workplace demands on the physical and mental health of employees and related organizational outcomes, such as loss of productivity and intention to quit. Such studies often have included factors such as job control as a mediating variable and role demands as stressors. Despite sophisticated analyses, the relationships among these factors have not been elucidated clearly or consistently. It is the aim in this paper to explore the linear relationship among three distinct groups of factors previously identified in the literature, stressors: workplace demands, work to family conflict; mediators: job control and sense of challenge; and outcome variables: burnout, somatic symptoms, job satisfaction, professional efficacy and intention to quit in a sample of 126 call centre representatives (59% female; age M = 27.3 years, SD = 8.18) from 11 call centres in metropolitan Melbourne. The results of a Multidimensional Scaling Analysis indicate four clusters: work related variables including role ambiguity, excessive performance monitoring, thoughts of quitting, role conflict; personal outcomes: work-family conflict and somatic symptoms; job impact outcomes: depersonalisation and emotional exhaustion, to positive outcomes: professional efficacy and job satisfaction. These quadrants can be used to suggest a progressive relationship from stressors through job control, a sense of problemsolving to either positive or negative outcomes. While these results are cross-sectional and must be interpreted with caution, a pivotal point of the MDS map suggests that participants’ level of timing and method and attention demanded by their role might be factors which differentiate the two outcomes. Occupational stressors are a major workplace concern because of their deleterious effects on the physical and mental health of employees (Danna & Griffin, 1999; Lloyd, King & Chenoweth, 2002) and the flow-on to organizational outcomes (Firth, Mellor, Moore, & Loquet, 2004). Mental health outcomes, such as burnout, and somatic symptoms such as tension and poor sleep (Parker & Kulik, 1995), decreased employee efficiency and performance and can lead to absenteeism and quitting (Firth, et al., 2004) as well as loss of productivity. Various theorists have proposed factors which reduce stress outcomes, including Karasek (1979) in his Demand-Control Model. Using Karasek’s four-quadrant approach involving high/low control and high/low demands, it is the aim in this study to test neighbourhood relationships among a series of stressors, stress outcomes, as well as control over variables in a sample of call centre representatives (CCRs). This population was chosen as call centres generate highly demanding, repetitive and stressful work roles that result in high absenteeism and staff turnover rates (Australian Communications Association Research [ACA], 1998; Australian Services Union [ASU], 1999), and because it is possible, following Van der Doer and Maes (1999), to match the assessment of control variables to the workplace demands.
The study of call centre representatives (CCRs), whose primarily work roles are to answer incoming calls (inbound), make outgoing calls (outbound) or handle a mixture of both, in order to generate costefficient business is important as, in other countries, the Australian call centre industry has grown substantially over the last decade. Figures available for the period 1998 to 2001 show an increase in the workforce of 20% per year (ACA, 1998; Australian Council of Trade Unions [ACTU], 2001). However, in the same period, the average staff turnover rate in the industry rose from 18% to between 23.7% and 29.5%. (ACA, 1998; ACA, 2001). High turnover rates have multiple negative effects on an organization including lack of employee continuity, organizational instability, and an increase in the proportion of inexperienced CCRs (ACTU, 2001), which in turn has an impact on the standard of professional services provided to customers. Further, the recruitment, induction and training of replacement staff increase the operating costs for call centres. At the individual level, the average cost of losing a CCR in 1998 was estimated at $AUS 10,000 (ACA, 1998), while at the industry level the increases in staff turnover described above, led to an increase in staff turnover related costs from $AUS 100 million in 1998 to $AUS 330 million in 2001 (ACA, 1998; ACTU, 2001). In 2001 stress-related absenteeism added an additional costs of $AUS 7.5 million per annum (ACTU). Clearly it is desirable to reduce these high costs. In his Demand-Control Model Karasek (1979) argued that high demands combined with low control equal poor health outcomes while high control over the same demands can lead to adaptive outcomes. Van der Doef and Maes (1999) in a review of 63 studies found that only half of the studies which used demand and control as independent variables significantly predicted outcomes such as job satisfaction, burnout, and job or health wellbeing. Furthermore, of the studies which tested for an interaction between demand-control on outcome variables, less than half those studies had results which reached significance. The lack of significant results might be attributed to a lack of congruence in the measures of control with the types of demands. In this study we shall use common workplace demands as also experienced by CCRs as well as call centre relevant demands and control variables to ascertain the spatial relationship between these two sets of factors. Although not causal, these findings might also suggest directionality of the relationships. Workplace issues identified consistently as stressors include role ambiguity, role conflict and role overload (Firth et al., 2004; Tate, Whatley & Clugston, 1997; Wunder, Dougherty & Welsh, 1982). Role conflict is defined as the occurrence of incompatible sets of demands that need to be satisfied simultaneously (Khan, Wolfe, Quinn, & Snoek, 1964), while role ambiguity results when employees are uncertain of what is expected of them or when they have received inadequate feedback on their job performance so as to reduce their ambiguity (Australian Council of Trade Unions [ACTU], 2001; Posig & Kickul, 2003); role overload exemplifies the classic scenario of demands exceeding resources, most often the time available (Wunder et al., 1982). High levels of role conflict and role ambiguity have been associated with increased occupational stress and burnout (Lee & Ashforth, 1996; Low, Cravens, Grant & Moncrief, 2001) including emotional exhaustion (Babakus, Cravens, Johnston, & Moncrief, 1996). Some studies (e.g., Lee & Ashford, 1996; Posig & Kickul, 2003) have found that role overload contributes to depersonalization and emotional exhaustion among employees. Other studies (e.g., Mendelson, Catano & Kelloway, 2000) have shown that role overload increases work stress, and this may make employees more susceptible to physical illnesses such as fatigue, dizziness and nausea. The demands faced by employees in the workplace have been shown also to contribute to conflict in the family (Good, Sisler & Gentry, 1988; Greenhaus & Beutell, 1985) with some authors suggesting this conflict has a spillover effect from work to family while others suggest it is a cumulative effect. Researchers (e.g., Frone, Russell & Cooper, 1992; Netemeyer, Brashear-Alejandro, & Boles, 2004) have found that when work roles interfere with the family-role, there is an increase in stress responses. O’Driscoll, Poelmans, Spector, et al. (2003) also found that work to family conflict contributes to psychological strain, such as nervousness and irritability while Burke and Greenglass (2001) found that work to family conflict was associated with emotional exhaustion, depersonalization and diminished feelings of accomplishment among nurses. Aside from these demands which appear to be common across many workplace sectors, it is usual to find other stressors which are intrinsic to the organisation or to the particular job role (Moore & Cooper, 1998). While not unique to the call centre industry, it is interesting that the ACTU (2001) and Holman (2002; 2003) identified performance monitoring and lack of job opportunities or satisfaction as stressors that are highly relevant in that domain. Performance monitoring is where employee-customer interactions whether on a display screen or over the phone is monitored and recorded (ACTU, 2001). Supervisors use these data to assess employees’ phone ethics, customer service skills and call-durations. The immediacy of
feedback to CCRs, the use of constructive criticism and the clarity of rating criteria have been found to reduce role ambiguity, and to increase satisfaction with the monitoring system (Chlaykoff & Kochan, 1989) however, performance monitoring is often perceived by CCRs as pervasive and unpleasant (ACTU, 2001). When used excessively for punitive reasons or attached to narrow performance-based salary assessment, it has been found to increase stress reactions (ACTU, 2001), increase depression and decrease activity (Holman, 2002). It is not clear whether CCRs typically perceive performance monitoring as a stressor or not, or how their perceptions of performance monitoring could affect levels of burnout. A second aspect of job design, job demands, refers to unique organizational aspects of the CCR’s job that require a sustained physical and mental effort (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001) such as problem-solving demands and attention demands (Holman, 2002). In their review of demand-control studies, Van der Doef and Maes (1999) found that these demands predict adverse psychological outcomes, such as depression and burnout. Aronson and Rissler (1998) also found that increasing job demands predicts somatic symptomatology. Despite its intuitive good sense, application of the D-C model has often produced equivocal results (Van der Doef & Maes, 1999) or results of statistical yet little practical significance, which Van der Doef and Maes (1999) suggested reflects a lack of congruency between the measures of demand and those of control. The need for congruent measures can be thought of as the specificity hypothesis of which Wallston, Wallston and DeVellis (1978) and Bandura (1997) were earlier advocates with respect to the measurement of health locus of control and of self-efficacy, respectively. It seems prudent therefore, that when measuring the stressors within any one organisation researchers should also measure control as it relates to those conditions. Within call centres job control can be seen to represent the extent to which CCRs are given discretion over their work-task (Holman, 2002). Some call centres require CCRs to adhere to strict call durations (timing) and emphasize the use of scripts (method), while others allow employees to manage their own call durations when interacting with customers. A lack of job control, such as the requirement to adhere to scripts, has been found to correlate positively with emotional exhaustion (Deery, Iverson & Welsh, 2002; Lee & Ashford, 1996). Other studies (e.g., Beehr, Glaser, Canali & Wallwey, 2001) have found that job dissatisfaction is negatively related to demands and, according to them, is not an intervening factor between stressors and strains. It is the aim in this study to explore the spatial relationships among a series of workplace stressors, stress outcomes, as well as aspects of control and perceptions of demands as a challenge among CCRs. Method Participants Participants were 126 CCRs (58.7% female) recruited from 11 call centres based in metropolitan Melbourne. Respondents age ranged from 19 to 66 years (M = 27.3, SD = 8.18); education levels varied from completing some time at high school to completing a degree. Tenure was relatively short with 70.6 % of respondents having worked less than 2 years in their current position. The sample consisted of 53.2% inbound CCRs, 37.3% outbound CCRs, and 9.5% of participants in mixed roles. Measures Participants completed provided demographic data, and completed the following measures. Common organizational stressors were assessed using 9-items from Tate, et al.’s (1997) tri-nation study of retail sales people. Three items measured each of: role ambiguity (e.g., My job objectives are well-defined - R), role conflict (e.g., To satisfy some people at my job, I have to upset others), workoverload (e.g., It seems that I have more work at my job than I can handle) and work-family conflict (e.g., My job schedule interferes with my family life). Questions were answered on a five-point Likert scale, strongly agree to strongly disagree. In the current data alpha are all > .73. Call centre specific stressors were assessed using Jackson et al.’s (1993) 20-item scale to measure four components of job design: timing control (e.g., Can you decide when to answer or make a call?), method control (e.g., Can you vary how you talk with customers?), and attention demand (e.g., Do you have to concentrate all the time to watch for things going wrong?. Also from Jackson et al. was problem-solving demand (e.g., Are you required to deal with problems that are difficult to solve?). Items were reworded when necessary to reflect a call centre environment. Alpha in the current data were all α > .77. Performance monitoring was assessed by two-questions from Holman (2002). The first concerns the extent to which CCRs agree that they were being overly monitored; the second measures how much they agree with the statement: call monitoring is used to punish you rather than develop you. Participants
responded on a five-point scale, strongly disagree to strongly agree. Internal consistency in the current data was α =. 76. Job satisfaction was measured by a single item rated on a five-point scale ranging from very dissatisfied to very satisfied with advancement to a better position since commencing the job. Burnout was measured using the three factors of the Maslach Burnout Inventory-General Survey (MBI-GS) (Schaufeli et al., 1996): emotional exhaustion (e.g., Working all day is really a strain on me), depersonalization (e.g., I doubt the significance of my work), and diminished feelings of accomplishment (e.g., I have [not] accomplished many worthwhile things at work). Alpha in the current data are > .72). Somatic symptomatology was assessed using four items from Tate, et al. (1997) on anxiety and somatic complaints (e.g., Job-related problems make my stomach upset). Participants indicated on a six-point scale the degree to which they experience each of these symptoms. Alpha in the current data was .92. Procedure Following institutional ethics approval participants were recruited from call centres where managers had provided written approval for the promotion of the study and the distribution of questionnaires. These centres service a range of industry sectors, including financial services, telecommunication services, utility services, information technology services, newspaper companies, non-profit organizations, and travelling and airline services. Participants were provided with a Plain Language Statement, and completed the questionnaire in their own time. To ensure anonymity, participants were directed to return questionnaires via the replied-paid envelopes provided. Of the 280 questionnaires distributed, 126 questionnaires were returned, representing a response rate of 45%. Results Multidimensional Scaling, a technique which aims to identify the dimensions along which elements are perceived to vary in multidimensional space (Hair, Anderson, Tatham & Black, 1995), was used to explore the spatial relationships among the stressor, control and outcome variables. A dimensionality value of r2 .99 together with a Kruskall’s stress of .04, indicated a good fit between proximities and the mapped distances on the two-dimensions (Figure 1). These two dimensions can be labeled as health for the lower quadrants and job affect for the upper quadrants, and represent the largest distances (smallest similarities) between the constructs (Kruskall & Wish, 1981). On the health dimension, reading from right to left the variables are indicative of somatic symptoms, some work-family conflict, to the negative aspects of burnout: depersonalization and emotional exhaustion. On the job affect dimension, the variables read from demands through perceived control to job satisfaction and a sense of professional efficacy or accomplishment. Interestingly, intention to quit is located among the stressor variables.
Figure 1 Multidimensional scaling map.
The neighbourhood approach to interpretation, which focuses on the smallest distances (or largest similarities) between constructs, suggests that the variables cluster into stressors, comprised of acknowledged stressors: role ambiguity, role conflict, role overload, work-to family conflict, as well as performance monitoring and concurrent cognitive and behavioural effects of these stressors, namely, thinking about quitting and somatic symptoms. Using this approach, attention and problem-solving demands are lesser stressors, while control over the timing of one’s calls and method of calls forms a separate cluster where, if this were a temporal space, this perception of control might be argued to occur after the perception of demands. The outcome variables cluster in two groups: positive (a sense of professional efficacy or accomplishment and job satisfaction) and negative (emotional exhaustion and depersonalization) and it can also be seen by their positioning on the map’s axis that a perception of control, of timing and method, contributes positively to the variables in the positive cluster but inversely to those in the negative cluster. Discussion The aim in this study was to explore the spatial relationships among a series of stressors, control, and outcome variables using Multidimensional Scaling (MDS). MDS plots are constructed in terms of their similarities. In Figure 1, the stressors: role overload, role ambiguity, role overload, and performance monitoring and work to family conflict cluster together and intention to quit (a cognition which can be interpreted as a response to these demands, i.e., [I often] think about leaving my present job) is also part of this cluster. This position is contrary to most research which has used intention to quit as an outcome variable. This finding is particularly noteworthy as employees’ thoughts of quitting or intentions to quit are frequently not translated into behavioural outcomes, that is, actual quitting. It might be that they are more an indicator of the level of stress being experienced than an outcome of those stressors or, perhaps even a means of coping by distancing oneself, at least by some form of cognitive escape or defense from the issues. This proposition does of course require verification in future studies especially those of a longitudinal nature. Interestingly, somatic symptoms also lie close to the cluster of stressors and between the stressors and the negative outcome variables of depersonalization and emotional exhaustion suggesting that a physical response might occur following or in conjunction with the stressors and before the depth of emotional or psychological response to stressors suggested here by depersonalization and emotional exhaustion. Such an interpretation is congruent with an illness response to stressors and to related levels of absenteeism and suggests that a physical response to stressors might occur first and it, in conjunction with worker absenteeism, should perhaps be considered as an indicator of further future distress such as the components of burnout used here. An examination of these spatial relationships also raises the question does control precede or succeed perception of demands? While this is not a longitudinal study, the relational spaces in which the study variables are plotted can be used to suggest that a sense of control follows demands rather than preceeding them. The control variables (over method and timing) are plotted between the stressors and the positive outcome variable, that is, a sense of professional efficacy or accomplishment, and, conversely, they are opposed to the negative outcome variables. In other words, one perceives the demands and, using Lazarus and Folkman’s (1984) taxomony, would appraise them and decide that one has some control resulting in an increased sense of efficacy or accomplishment. Conversely, if there is no sense of control the outcome is negative. While these results are both exploratory and cross-sectional they do tentatively confirm the appraisal hypothesis of Lazarus and Folkman (1984) that these perceptions follow demands and clarify the role of demand-control as put forward by Karasek (1979). Clearly, Karasek’s four quadrant model may still be valid in terms of evaluating people or groups on differences, in terms of control and demands however, his model failed to answer the question of directionality. Clearly, the current findings pertain to a specific sample of CCRs however, one strength of this study is in the matching of demands and control variables as suggested by Van der Doer and Maes (1999), and in the use of MDS to allow the variables to ‘plot against each other’ in their own spaces. Replication of these findings is important as is a longitudinal study to determine if the suggested linear relationships are in fact valid. Only then can we validly apply these findings in workplace strategies. References
Aronsson, G., & Rissler, A. (1998). Psychophysiological stress reactions in female and male urban bus drivers. Journal of Occupational Health Psychology, 3, 122–129. Australian Communications Association Research [ACA] (1998), Call Centre Hang-ups. The Call Centre Agent Report, ACA Research Pty Ltd., Sydney. Australian Communications Association Research [ACA] (2001), The 2001 Australia and New Zealand call centre industry benchmark study, ACA Research Pty Ltd., Sydney. Australian Council of Trade Unions Call Centre Unions Group [ACTU]. (2001). On the line: The future of Australia’s call centre industry. Sydney: Author. Babakus, E., Cravens, D.W., Johnston, M., & Moncrief, W.C. (1996). Examining the role of organizational variables in the salesperson job satisfaction model. Journal of Personal Selling and Sales Management, 16, 33-46. Bandura, A. (1997). Self Efficacy: The Exercise of Control. Englewood Cliffs: Prentice Hall. Beehr, T.A., Glaser, K.M., Canali, K.G., & Wallwey, D.A. (2001). Back to the basics: Re examination of demand-control theory of occupational stress. Work and Stress, 15, 115-130. Burke, R.J., & Greenglass, E.R. (2001). Hospital restructuring, work-family conflict and psychological burnout among nursing staff. Psychology and Health, 16, 583-595. Chlaykoff, J., & Kochan, T. (1989). Computer-aided monitoring: its influence on employee job satisfaction and turnover. Personnel Psychology, 42, 807-834. Danna, K., & Griffin, R.W. (1999). Health and well-being in the workplace: A Review and synthesis of the literature. Journal of Management, 25, 357-376. Deery, S.J., Iverson, R.D., & Welsh, J.T. (2002). Work relations in telephone call centers: understanding emotional exhaustion and employee withdrawal. Journal of Management Studies, 39, 471-496. Demerouti, E., Bakker, A.B., Nachreiner, F., & Schaufeli, W.B. (2001). The job demands-resources model of burnout. Journal of Applied Psychology, 86, 499–512. Firth, L., Mellor, D.J., Moore, K.A., & Loquet, C. (2004). How can managers reduce employee intentions to quit. Journal of Managerial Psychology, 19, 170-187. Frone, M.R., Russell, M. & Cooper, M.L. (1992). Antecedents and outcomes of work-family conflict: testing a model of the work-family interface. Journal of Applied Psychology, 77, 65-78. Good, L.K., Sisler, G.F., & Gentry, J.W. (1988). Antecedents of turnover intentions among retail management personnel. Journal of Retailing, 64, 295-314. Greenhaus, J.H., & Beutell, N.J. (1985). Sources of conflict between work and family roles. Academy of Management Review, 10, 76-88. Hair, J.F., Anderson, R.E., Tatham, R.L. & Black, W.C. (1995). Multivariate data analysis (4th ed.). New Jersey: Prentice Hall. Holman, D. (2002). Employee wellbeing in call centres. Human Resource Management Journal, 12, 3551. Holman, D.J. (2003). Call centres. In D.J. Holman, T.D. Wall, C.W. Glegg, P. Sparrow, & A. Howard (Eds.), The new workplace: A guide to the human impact of modern working practices. Chichester, UK: Wiley. Kahn, R., Wolfe, D., Quinn, R., & Snoek. J. (1964). Organisational stress: studies in role conflict and ambiguity. New York: Wiley. Karasek, R.A. (1979). Job demands, job decision latitude, and mental strain: implications for job redesign. Administrative Science Quarterly, 24, 285-309. Kruskal, J. B., & Wish, M. (1981). Multidimensional scaling. Beverly Hills; London: Sage Publications. Lazarus, R.S. and Folkman, S. 1984. Stress, appraisal, and coping. New York: Springer. Lee, R.T., & Ashford, B.E. (1996). A meta-analytic examination of the correlates of the three dimensions of job burnout. Journal of Applied Psychology, 81, 123-133. Lloyd, C., King, R., & Chenoweth, L. (2002). Social work, stress and burnout: A review. Journal of Mental Health, 11, 255- 266. Low, S.L., Cravens, D.W., Grant, K., & Moncrief, W.C. (2001). Antecedents and consequences of salesperson burnout. European Journal of Marketing, 35, 581-611. Mendelson, M.B., Catano, V.M., Kelloway, K. (2000). The role of stress and social support in Sick Building Syndrome. Work and Stress, 14, 137-156. Moore, K.A., & Cooper, C.L. (1998). Models of occupational stress. In S. Hardy, J. Carson & B. Thomas (Eds.), Professional caring: Personal and professional approaches to occupational stress (pp. 3-17). Stanley Thornes (Publishers) Ltd: UK.
Netemeyer, R.G., Brashear-Alejandro, T., & Boles, J.S. (2001) A cross-national model of job-related outcomes of work role and family role variables: A retail sales context. Journal of the Academy of Marketing Science, 32, 49-60. O'Driscoll, M.P., Poelmans, S., Spector, P.E., Kalliath, T., Allen, T.D., Cooper, C.L., & Sanchez, J.I. (2003). Family-responsive interventions, perceived organizational and supervisor support, work-family conflict, and psychological strain. International Journal of Stress Management, 10, 326-344. Parker, P.A., & Kulik, J.A. (1995). Burnout, self- and supervisor-related job performance and absenteeism among nurses. Journal of Behavioral Medicine, 18, 581-599. Posig, M., & Kickul, J. (2003). Extending our understanding of burnout: Test of an integrated model in nonservice occupations. Journal of Occupational Health Psychology, 8, 3-19. Schaufeli, W.B., Leiter, M.P., Maslach, C., & Jackson, S.E. (1996). Maslach Burnout Inventory-General Survey. In C. Maslach, S. Jackson, & M. Leiter, (Eds.), Maslach Burnout Inventory Manual. (3rd ed.). Palo Alto, CA: Consulting Psychologist Press. Tate, U., Whatley, A., & Clugston, M. (1997). Sources and outcomes of job tension: A three-nation study. International Journal of Management, 3, 350-358. Van der Doef, M., & Maes, S. (1999). The job demand-control (-support) model and psychological well being: A review of 20 years of empirical research. Work and Stress, 13, 87-114. Wallston, K.A., Wallston B.S., & DeVellis, R. (1978). Development of the multidimensional health locus of control (MHLC) scales. Health Education Monographs, 6(2). 160-170. Wunder, R.S., Dougherty, T.W., & Welsh, M.A. (1982). A causal model of role stress and employee turnover. Academy of Management, 42, 297-301.
