Counselor emotional exhaustion and turnover intention in therapeutic communities

Journal of Substance Abuse Treatment 31 (2006) 173 180 Regular article Counselor emotional exhaustion and turnover intention in therapeutic communities Hannah K. Knudsen, (Ph.D.) a, 4, Lori J. Ducharme, (Ph.D.) a, Paul M. Roman, (Ph.D.) a,b a Center for Research on Behavioral Health and Human Services Delivery, The University of Georgia, 101 Barrow Hall, Athens, GA 30602-2401, USA b Department of Sociology, 119 Baldwin Hall, The University of Georgia, Athens, GA 30602-1611, USA Received 14 November 2005; received in revised form 28 March 2006; accepted 17 April 2006 Abstract Counselor turnover is a significant problem facing substance abuse treatment agencies. Understanding the role of organizational culture in predicting burnout and turnover intention may yield important information on how to address turnover in treatment organizations. Using data collected from 817 counselors employed in a national sample of 253 therapeutic communities (TCs), structural equation modeling was used to estimate the associations between emotional exhaustion, turnover intention, and three measures of organizational culture: centralized decision making, distributive justice, and procedural justice. The model controlled for counselor demographics, credentials, and earnings. Counselors emotional exhaustion scores were higher in TCs with greater centralized decision making ( p b.01) but lower in TCs where greater distributive justice ( p b.05) and procedural justice ( p b.001) were reported. Likewise, turnover intention was positively associated with centralized decision making ( p b.05) and inversely associated with the workplace justice measures ( p b.001). These data suggest that management practices in TCs and perhaps in other types of substance abuse treatment facilities likely play a substantial role in counselors well-being and in their decisions to leave their jobs. Because these practices are not structural features of organizations, they may be targeted for intervention and change. D 2006 Elsevier Inc. All rights reserved. Keywords: Turnover intention; Emotional exhaustion; Counselor burnout; Workplace justice 1. Introduction Although voluntary employee turnover is a concern for most types of organizations, the loss of counseling staff in substance abuse treatment facilities has especially strong impacts. Loss or change in counselors has implications for both the quality of care delivered and the program management. The loss of these employees disrupts patient care (Barak, Nissly, & Levin, 2001; Lum, Kervin, Clark, Reid, & Sirola, 1998) and increases costs for organizations that must then attract, hire, and train additional counseling staff. The rate of voluntary turnover is notably high among substance abuse treatment counselors (McLellan, Carise, & Kleber, 2003). These negative consequences of counselor 4 Corresponding author. Tel.: +1 706 542 6090; fax: +1 706 542 6436. E-mail address: hknudsen@uga.edu (H.K. Knudsen). turnover as well as their high rates have generated considerable interest among program managers in understanding its causes and reducing its incidence (Gallon, Gabriel, & Knudsen, 2003; Hser, 1995; Laundergan, Flynn, & Gaboury, 1986). Data from both treatment settings and other types of human service organizations suggest that emotional exhaustion, a key component of burnout, and turnover intention are key precursors of actual turnover. Therefore, understanding the predictors of emotional exhaustion and turnover intention may offer important insights into how treatment organizations can reduce counselor turnover. Drawing on theoretical principles of social exchange and using data from 817 counselors employed by therapeutic communities (TCs), this research examines the associations between emotional exhaustion, turnover intention, and three aspects of organizational culture: centralized decision making, procedural justice, and distributive justice. The TC setting 0740-5472/06/$ see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jsat.2006.04.003

174 H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 is particularly appropriate for testing these ideas because it involves substantial intensity of interaction between counselors and clients and often is involved in treating clients with a broad range of deeply seated problems in addition to their substance abuse (De Leon, 2000). TCs represent one of the five major substance abuse treatment modalities in the United States (Prendergast, Podus, Change, & Urada, 2002). Given their significance in the delivery of substance abuse treatment services in the United States, it is important that the parameters of their operation be documented. Emotional exhaustion is defined as the perception that one s emotional resources have been completely expended (Cordes & Dougherty, 1993; Maslach & Jackson, 1981). It is one of the three dimensions in Maslach and Jackson s (1981) classic model of employee burnout. The present research focuses on this dimension because emotional exhaustion may better encompass the true nature of burnout than the other two dimensions of depersonalization and diminished personal accomplishment (Cropanzano, Rupp, & Byrne, 2003; Shirom, 1989; Wright & Cropanzano, 1998). Further, a focus upon emotional exhaustion builds on a prior set of literature that has documented the positive association between this dimension and turnover intention (Barak et al., 2001; Blankertz & Robinson, 1997; Lee & Ashforth, 1996). The negative consequences of emotional exhaustion are wide reaching. Negative organizational outcomes include poor job performance (Cropanzano, Rupp, & Byrne, 2003; Wright & Cropanzano, 1998) and reduced client satisfaction with the services they have received (Garman, Corrigan, & Morris, 2002; Vahey, Aiken, Sloane, Clarke, & Vargas, 2004). Emotional exhaustion also has negative consequences in terms of illness, fatigue, and depression for the employees themselves (Burke & Deszca, 1986; Cherniss, 1980; Kahill, 1988; Pines & Maslach, 1978). Individual-level emotional exhaustion is consequential for other employees who interact with the burned-out individual, as there is a bsocial contagionq effect where burnout spreads among organizational members (Bakker, Schaufeli, Sixma, & Bosveld, 2001; Halbesleben & Buckley, 2002). Finally, longitudinal studies have demonstrated that emotional exhaustion is significantly associated with voluntary turnover (Maslach & Jackson, 1986; Wright & Cropanzano, 1998). Given the constraints of cross-sectional research methods, it is difficult to study directly the process of employee turnover; hence, turnover intention has been identified as a useful proximal measure (Farkas & Tetrick, 1989). Selfreported intentions to quit have repeatedly been demonstrated to predict actual turnover (Tekleab, Takeuchi, & Taylor, 2005). The validity of this approach is further supported by the meta-analytic work of Griffeth, Hom, and Gaertner (2000) who demonstrated that the effect size of intention to quit on actual turnover is considerably larger than the predictive power of other job-related factors on turnover behavior. Thus, understanding the aspects of work and organizational culture that are associated with turnover intention is critical if the larger issue of counselor turnover is to be addressed. 1.1. Conceptualizing predictors of emotional exhaustion and turnover intention An extensive literature suggests that stressful experiences in the workplace have implications for the well-being of employees (Barling, Kelloway, & Frone, 2005). Substance abuse treatment counselors face a wide range of stressors that result from organizational conditions and the social interactions they have with their clients. Certain stressors within the substance abuse treatment profession are difficult to mitigate (Shoptaw, Stein, & Rawson, 2000). For example, substance abuse treatment agencies face an increasingly turbulent environment where they are expected to deliver more and higher quality services with fewer resources (Lamb, Greenlick, & McCarty, 1998). In particular, TCs have been impacted by a newly demanding regulatory environment as they have become increasingly mainstreamed within the substance abuse treatment system (De Leon, 1999, 1995). Melnick and De Leon (1999) note that TCs have been impacted by the managed care environment that has sought to reduce treatment costs. In this environment, most programs cannot simply increase tangible rewards such as wages to increase the retention of counselors. Similarly, counseling involves substantial bemotional laborq (Hochschild, 1983) in building therapeutic relationships with clients, many of whom have not voluntarily entered treatment (Hiller, Knight, Broome, & Simpson, 1998). Additionally, these clients are often recidivists because of the chronic, relapsing nature of addiction (McLellan, Lewis, O Brien, & Kleber, 2000). Further, they typically have complex mental and physical needs that require counselors to coordinate the delivery of additional health services (Kessler, Chiu, Demler, & Walters, 2005; Sacks et al., 1998). These clinical complexities are likely to be highly prevalent in TC settings, as clients in residential programs tend to report more severe problems (Melnick, De Leon, Thomas, & Kressel, 2001). As Cordes and Dougherty (1993) observe, human service occupations such as counseling are at high risk of burnout bbecause they are constantly dealing with other people and their problemsq (p. 628). These challenges that are based on therapeutic relationship are intrinsic to the occupation and, therefore, are not as amenable to intervention as other organizational factors. In contrast, certain management practices within treatment organizations have the potential to reduce employee exhaustion and turnover intention through changes in the social interactions between management and counselors. The primary advantage of an emphasis on change within an organization s culture is that it does not require infusion of costly resources. Instead, these management practices, facilitated within the social interactions of everyday organizational life, develop perceptions among employees

H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 175 about the distribution of power and the presence of workplace fairness. When these perceptions are positive, they are likely associated with reduced emotional exhaustion and turnover intention among counselors. Organizations can be described as having centralized, hierarchical power structures or decentralized, participatory power structures. In the centralized form, employees are not free to make relevant decisions about their work; nearly all decisions must be referred up a bchain of commandq to supervisors for approval (Hage & Aiken, 1967; Hall, 1963). Participatory structures place significant amounts of decision making in the hands of employees (Hodson, 2001; Osterman, 2000). These participatory forms of organization have been shown to increase job satisfaction (Finlay, Martin, Blum, & Roman, 1995; Mortimer & Lorence, 1995). However, in organizations where power is centralized, employees likely feel greater stress due to having little control over their work; a lack of job control is well established as a significant work stressor (Karasek, 1979). Thus, we hypothesize that centralization is positively associated with counselor burnout and turnover intention. Strong emphasis on rigid hierarchical control within the treatment process is part of the TC tradition (De Leon, 2000) and would be expected to spill over into patterns of organizational management. Related to, but distinct from, the distribution of power is the concept of workplace justice. The theoretical underpinning of justice s likely effects on employee well-being is drawn from equity theory. According to equity theory, feelings of disadvantage in social relationships translate into psychological distress (Adams, 1965; Van Dierendonck, Schaufeli, & Buunk, 2001). Recent work on organizational justice suggests that experiencing injustice in the workplace can be a significant source of stress and, hence, may be linked to a variety of negative outcomes (Cropanzano et al., 2005; Elovainio, Kivimaki, & Vahtera, 2002). Organizational justice is multifaceted; two aspects that have been studied in a variety of different types of organizations are distributive justice and procedural justice. Distributive justice refers to perceptions about how fairly the workload and rewards are distributed across members of the organization (Adams, 1965). These perceptions of distributive justice may matter more than objective levels of such justice (Bloom & Michel, 2002). Recent reviews of the literature have suggested that perceived distributive injustice is positively associated with employee withdrawal behaviors including burnout (Halbesleben & Buckley, 2004) as well as intention to quit (Cohen-Charash & Spector, 2001; Colquitt, Conlon, Wesson, Porter, & Ng, 2001). These associations have also been demonstrated in samples of workers in human service occupations (Schaufeli, Van Dierendonck, & Van Gorp, 1996; Van Dierendonck et al., 2001). Procedural justice describes the extent to which the processes through which organizational decisions are made are perceived to be fair (Leventhal, 1980). Procedural justice is typically the consequence of collecting relevant information from employees before decisions are made and then applying these decisions fairly across employees (Niehoff & Moorman, 1993). Recent meta-analyses found that perceptions of procedural justice were associated with a constellation of positive work attitudes, including job satisfaction, organizational commitment, and intention to quit (Cohen-Charash & Spector, 2001; Colquitt et al., 2001). To summarize, we hypothesized that centralized decision making, distributive justice, and procedural justice are significantly associated with emotional exhaustion and turnover intention. It is also expected that there is a positive correlation between exhaustion and turnover intention. These hypotheses are examined in a large sample of counselors employed in TC settings. 2. Materials and methods 2.1. Sample A nationally representative sample of 380 self-identifying TCs was drawn in 2002 2004. A two-stage sampling process was utilized. First, U.S. counties were allocated to 1 of 10 strata based on population, and counties were randomly sampled within strata. All substance abuse treatment facilities within the sampled counties were enumerated using published directories, including federal and state provider listings. From these lists, treatment organizations were then randomly selected proportionate to the total number of centers in the sampled counties. A brief telephone interview was conducted to assess eligibility for the study. Facilities that were ineligible or refused to participate were replaced with organizations randomly selected from the same geographic stratum. The resulting sample of TCs was located in 42 states, including urban, suburban, and rural areas. Three criteria determined eligibility for the study, and this information was collected during the brief telephone screening interview. First, the organization was required to provide substance abuse treatment. Second, the facility must be community based; facilities that operated within the Veteran s Administration Health System or in correctional settings were not eligible for the study. Third, the center needed to self-identify as a TC. Although data such as membership in Therapeutic Communities of America and adherence to the classic TC model developed by De Leon (2000), Melnick and De Leon (1999), and Melnick, De Leon, Hiller, and Knight (2000) were collected during the in-depth interviews, such criteria were not employed during sample selection. This approach allowed for measurement of the variation among facilities that self-identified as TCs, thereby capturing the diversity of TCs within the United States. The 380 TCs included in these analyses represent 86% of the facilities that met these eligibility criteria and agreed to participate.

176 H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 2.2. Data collection Data collection utilized two methods: face-to-face interviews and mailed questionnaires. Organizational-level data about the TC were collected during face-to-face interviews with the TC s administrator, clinical director, or both. These interviews averaged about 2 hours in length, and participating TCs received a US$100 honorarium. At the conclusion of the interview, the respondent was asked to provide a list of counselors at the facility who provide services to substance-abusing clients. All listed counselors were mailed a questionnaire and an informed consent form. Counselors who returned a completed questionnaire received US$40 for their participation. These questionnaires were mailed to 1,779 counselors, and 1,053 completed questionnaires were received for a response rate of 59.2%. This rate of participation was similar to recent data collection from counselors at public and private treatment facilities (Knudsen, Ducharme, Roman, & Link, 2005) as well as other workforce studies (Forman, Bovasso, & Woody, 2001; Mark, Kranzler, & Song, 2003; Thomas, Wallack, Lee, McCarty, & Swift, 2003). At least one counselor questionnaire was received from 253 (66.6%) of the TCs in the total sample. Complete data on all indicators for the present analyses were available from 817 counselors. 2.3. Measures Two dependent variables were examined: emotional exhaustion and turnover intention. Emotional exhaustion was measured by nine items from the well-established scale developed by Maslach and Jackson (1986). Each item used a Likert response format where 1 represented not at all true and 7 represented very true. Turnover intention was measured by three items, adapted from Walsh, Ashford, and Hill (1985). Counselors were asked to indicate their agreement using a Likert response format where 1 represented strongly disagree and 7 represented strongly agree. The text for these items and all other scales used in the analyses appear in Table 1. The three independent variables of interest were centralized decision making, distributive justice, and procedural justice. Centralized decision making was measured by five items; this classic organizational measure was developed by Hage and Aiken (1967). Counselors were asked how true each statement was with regard to the TC where they were employed; responses ranged from 1 (not at all true) to 7(very true). The five indicators of procedural justice and four measures of distributive justice were adapted from Niehoff and Moorman (1993). Counselors were asked to rate their agreement with each item on a scale that ranged from 1 (strongly disagree) to 7(strongly agree). Eight control variables are included in the model of counselor burnout and turnover intention. These sociodemographic measures include the following: gender (1 = female, 0 = male), age (in years), and race (1 = racial/ethnic minority, 0 = White). Five additional control variables assessed other individual-level characteristics associated with counselors occupation, including certification status (1 = certified addictions counselor, 0 = not certified), personal recovery status (1 = in recovery, 0 = not in recovery), master s-level degree or higher (1 = yes, 0 = no), years in the substance abuse treatment field, and earnings. This measure of earnings consisted of one of nine categories (bus$15,000, US$15,000 20,000, US$20,000 25,000, etc.); the midpoint of each Table 1 Confirmatory factor model Item wording Factor loading Emotional exhaustion I feel emotionally drained from my work..748 I feel fatigued when I get up in the morning.754 and have to face another day on the job. Working with people all day is really a strain for me..596 I feel burned out from my work..820 Working directly with people puts.637 too much stress on me. I feel frustrated with my job..835 I feel used up at the end of the workday..809 I feel like I am working too hard on my job..731 I feel like I am at the end of my rope..753 Turnover intention As soon as I can find a better job,.915 I will leave this TC. I am actively looking for a job at another TC..725 I am seriously thinking of quitting my job..880 Centralized decision making There can be little action taken here until.714 a supervisor approves a decision. People who want to make their own.733 decisions would be quickly discouraged. Even small matters have to be referred.892 to someone higher up for approval. Employees have to ask their supervisors.914 before doing almost anything. Any decisions employees make must.854 have their boss s approval. Distributive justice Where you work, the amount of pay.702 employees receive is distributed fairly. Employees receive an amount of.663 fringe benefits that is fair. The workload at this center is fairly distributed..751 The overall rewards workers.837 receive are fairly distributed. Procedural justice The TC s management makes sure that employee.853 concerns are heard before decisions are made. Job decisions are applied consistently across.820 all affected employees. Employees are allowed to challenge or appeal.748 job decisions that are made by managers. When decisions are made, all the people who.836 will be affected are asked for their ideas. The TC s management clarifies decisions and provides.822 additional information when requested by employees.

H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 177 category, and the method of Parker and Fenwick (1983) for estimating the value of the final open-ended category (NUS$50,000), were used for analytic purposes. The resulting earnings measure is expressed in thousands of U.S. dollars. 2.4. Data analysis The relationships between the variables were estimated using Mplus version 3.01, a structural equation modeling software package (Muthen & Muthen, 2004). This software allows for the estimation of a measurement model and a structural model of hypothesized relationships. The measurement model was developed based on an exploratory factor analysis that produced a five-factor solution (not shown). This model (of the two dependent variables and three independent variables) produced five factors with eigenvalues over 1, and none of the items cross-loaded. The next step was to conduct the analysis of the confirmatory factor and structural models, which were estimated simultaneously using the Mplus software. The confirmatory factor analysis yields a measurement model of latent variables from the shared variance between individual items, which has the advantage of parceling out the error components from the items. The result is an unobserved measure that is more reliable and valid. Some residual errors within constructs were allowed to be correlated to improve model fit. Using these unobserved measures, Mplus estimates the structural model of hypothesized relationships between latent variables. This software produces measures of overall model fit, estimates of the hypothesized associations (unstandardized and standardized coefficients, standard errors, and t tests), and measures of the proportion of variance explained for each dependent variable. 3. Results 3.1. Preliminary analyses Although 1,053 TC counselors responded to the survey, only 817 counselors provided complete data on all variables of interest. Therefore, it was important to consider the issue of response bias before conducting the structural equation model of emotional exhaustion and turnover intention. To examine possible response bias, we compared the demographic characteristics of the 817 counselors with administrator reports regarding the composition of their TC s counseling staff. The counselor questionnaire data revealed that most TC counselors (59.1%) were female and 43.1% reported of having a racial/ethnic minority background. The average age was 43 years (M = 43.17, SD = 10.55), and the average TC counselor had spent about 7 years in the treatment field (M = 7.08, SD = 6.23). More than half (57.4%) were personally recovering from substance abuse. About 46.1% of the sample were certified in addiction treatment and about one quarter (26.8%) had attained a master s-level degree or higher. The average salary was about 27.12 or roughly US$27,000 per year (SD = 9.30). Notably, these descriptive statistics were remarkably consistent with the reports of administrators who were asked to describe the characteristics of their counseling staff. The averages for these TC-level counselor characteristics were as follows: 56.7% were female, 44.0% were racial/ ethnic minorities, 57.1% were personally in recovery, 46.9% were certified, and 29.0% had a master s-level degree or higher. The average counselor salary, as reported by administrators, was about US$28,400. This high congruence between the characteristics of the counselor respondents and the reports of TC administrators suggests that response bias was not a significant problem in the counselor-level data. Table 2 Structural model of counselor emotional exhaustion and turnover intention (N = 817) Emotional exhaustion Turnover intention Variables b (SE) b b (SE) b Centralized decision making.118 (.038)44.127.098 (.048)4.078 Distributive justice.115 (.055)4.116.338 (.070)444.251 Procedural justice.267 (.053)444.302.419 (.067)444.347 Gender (1 = female).192 (.092)4.139.139 (.116).074 Age in years.023 (.005)444.177.009 (.006).051 Race (1 = non-white).269 (.090)44.196.162 (.113).087 Certified in addictions.069 (.092).050.105 (.116).056 Recovering.088 (.097).064.155 (.123).083 Master s degree or higher.241 (.112)4.176.449 (.141)44.240 Years in the treatment field.007 (.008).032.008 (.010).028 Earnings.009 (.005).063.003 (.007).015 Association between emotional exhaustion.635 (.073)444.247 and turnover intention Latent variable, R 2.298.378 4 p b.05 (two tailed). 44 p b.01 (two tailed). 444 p b.001 (two tailed).

178 H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 3.2. Structural equation model of emotional exhaustion and turnover intention The confirmatory factor model for the measures of organizational culture, emotional exhaustion, and turnover intention appears in Table 1. All items had loadings exceeding.50 and loaded significantly on their intended factors. Overall, the model fit the data very well. Hu and Bentler (1999) suggest that the Tucker Lewis Index and the Comparative Fit Index should exceed.95. For this model, the values were.952 and.960, respectively. In addition, the root mean squared error of approximation (RMSEA) was.040 and the standardized root mean squared residual (SRMR) was.039. Excellent model fit is indicated by RMSEA and SRMR values under.05 (Kelloway, 1998). The structural model supported the hypothesized relationships between the variables. First, there was a positive association between centralized decision making and emotional exhaustion (b =.127, p b.01). Counselors who reported working in TCs where decisions were required to go through a hierarchical chain of command were significantly more emotionally exhausted than counselors working in TCs where decision making was less centralized. In addition, there was a significant positive association between centralized decision making and turnover intention (b =.078, p b.05), such that counselors indicated greater intentions to quit when the TC placed greater emphasis on a hierarchical model of decision making (Table 2). Second, distributive justice was significantly associated with emotional exhaustion and turnover intention in the expected directions. The negative association between distributive justice and emotional exhaustion (b =.116, p b.05) suggests that when counselors perceive that the workload and rewards within the TC are more fairly distributed among its employees, they are less likely to report feelings of emotional exhaustion. Additionally, there was a significant negative association between this form of justice and intentions to quit (b =.251, p b.001), such that greater perceived distributive justice was associated with lower levels of turnover intention. Procedural justice was significantly associated with both dependent variables. Counselors who perceived that workplace procedures were fair reported significantly less emotional exhaustion than counselors who reported lower levels of procedural justice (b =.302, p b.001). Turnover intention was also negatively associated with procedural justice; higher counselor reports of procedural justice were associated with lower levels of turnover intention (b =.347, p b.001). It was hypothesized that there would be a significant positive correlation between emotional exhaustion and turnover intention. As expected, there was a significant positive association between these dependent variables (b =.247, p b.001). The control variables were more predictive of emotional exhaustion than turnover intention. Four variables were associated with emotional exhaustion. Greater emotional exhaustion was reported by counselors who were male, were Caucasian, were younger, and held a master s-level degree or higher. Only one of the control variables was associated with turnover intention. Counselors with at least a master s-level degree reported significantly greater intentions to quit. Altogether, these variables explained a considerable amount of the variance in both emotional exhaustion and turnover intention. Nearly 30% of the variance in the latent variable of emotional exhaustion was explained by centralized decision making, distributive justice, procedural justice, and the control variables. For turnover intention, about 38% of the variance was explained by the model. 4. Discussion Although turnover among counselors in substance abuse treatment organizations has been identified as a significant issue by program managers, research examining the precursors of turnover has been rare. Using theoretical propositions derived from studies conducted in other types of organizations, this research applied the concepts of centralized decision making, distributive justice, and procedural justice to self-reported emotional exhaustion and turnover intention among counselors employed by TCs. This study of emotional exhaustion and turnover intention is the first to be conducted among counselors working in a national sample of TC settings. Data collected from 817 TC counselors provided support for the hypothesized relationships between the variables. As expected, centralized decision making was positively associated with emotional exhaustion and turnover intention, whereas the two forms of organizational justice were negatively associated with these dependent variables. In addition, there was a significant positive correlation between exhaustion and intention to quit. Several limitations associated with this research should be noted. First, the data are cross-sectional in nature, making it difficult to establish causal relationships between the variables. Second, the data are only representative of one modality within the American substance abuse treatment system. However, the consistency of these results with management research conducted in other types of organizations suggests that these findings may have applicability to other treatment organizations. Some caution regarding the extent to which these results will generalize due to rates of response may be warranted. However, it is notable that the characteristics of the counselors from whom complete data were available are highly consistent with the description of the TC workforce provided by administrators. Finally, the models explain a considerable amount of the variance in emotional exhaustion and turnover intention, but future research should continue to explore other variables that contribute to these outcomes to build a more thorough understanding of the burnout and turnover processes.

H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 179 Although all three independent variables were significantly associated with emotional exhaustion and intention to quit, there were noticeable differences in the magnitudes of the associations. The standardized coefficients reveal that procedural justice was more strongly associated with emotional exhaustion than centralized decision making and distributive justice. This finding is consistent with other research that has shown that procedural justice is more important than distributive justice in predicting burnout (Brotheridge, 2003; Schmike, Ambrose, & Cropanzano, 2000). The standardized coefficients for the two forms of justice were closer in magnitude when predicting turnover intention, although procedural justice remains larger. Although centralized decision making was significantly associated with intention to quit, the association was quite small compared with the measures of organizational justice. This may also be reflective of the fact of historically institutional hierarchy within the TC therapeutic model (De Leon, 2000). These findings suggest that program managers may gain substantial benefits from examining perceptions of procedural justice within their organizations. Perceptions among counselors that managers will listen to their opinions and will apply decisions fairly across workers are strongly associated with emotional exhaustion and intentions to quit. Program managers may be able to address, in part, problems of high turnover within their organizations through social interactions with their counseling staff. Further attention to these documented sources of organizational instability should lead to relatively straightforward and inexpensive interventions to reduce their prevalence. Acknowledgment The authors gratefully acknowledge the research support of the National Institute on Drug Abuse (Grant No. R01DA14976). References Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.) Advances in experimental social psychology (vol. 2, pp. 267 299). New York7 Academic Press. Bakker, A. B., Schaufeli, W., Sixma, H. J., & Bosveld, W. (2001). Burnout contagion among general practitioners. Journal of Social and Clinical Psychology, 20, 82 98. Barak, M. E. M., Nissly, J. A., & Levin, A. (2001). Antecedents to retention and turnover among child welfare, social work, and other human service employees: What can we learn from past research? A review and metaanalysis. Social Service Review, 75, 625 661. Barling, J., Kelloway, E. K., & Frone, M. R. (Eds.). (2005). Handbook of work stress (pp. 545 546). Thousand Oaks, CA7 Sage. Blankertz, L. E., & Robinson, S. E. (1997). Turnover intention of community mental health workers in psychosocial rehabilitation services. Community Mental Health Journal, 33, 517 529. Bloom, M., & Michel, J. G. (2002). The relationship between organizational context, pay dispersion, and managerial turnover. Academy of Management Journal, 45, 33 42. Brotheridge, C. M. (2003). The role of fairness in mediating the effects of voice and justification on stress and other outcomes in a climate of organizational change. International Journal of Stress Management, 10, 253 268. Burke, R. J., & Deszca, E. (1986). Correlates of psychological burnout phases among police officers. Human Relations, 39, 487 502. Cherniss, C. (1980). Professional burnout in human service organizations. New York7 Praeger. Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in organizations: A meta-analysis. Organizational Behavior and Human Decision Processes, 86, 278 321. Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O. L. H., & Ng, K. Y. (2001). Justice at the millennium: A meta-analytic review of 25 years of organizational justice research. Journal of Applied Psychology, 86, 425 445. Cordes, C. L., & Dougherty, T. W. (1993). A review and integration of research on job burnout. Academy of Management Review, 18, 621 656. Cropanzano, R., Goldman, B. M., & Benson, L. (2005). Organizational justice. In J. Barling, E. K. Kelloway, & M. R. Frone (Eds.), Handbook of work stress (pp. 63 87). Thousand Oaks, CA7 Sage. Cropanzano, R., Rupp, D. E., & Byrne, Z. S. (2003). The relationship of emotional exhaustion to work attitudes, job performance, and organizational citizenship behaviors. Journal of Applied Psychology, 88, 160 169. De Leon, G. (1995). Residential therapeutic communities in the mainstream: Diversity and issues. Journal of Psychoactive Drugs, 27, 3 15. De Leon, G. (1999). Therapeutic communities. In M. Galanter, & H. Kleber (Eds.), The American Psychiatric Press textbook of substance abuse treatment (2nd ed., pp. 447 462). Washington7 American Psychiatric Press. De Leon, G. (2000). The therapeutic community: Theory, model, and method. New York7 Springer Publishing Company. Elovainio, M., Kivimaki, M., & Vahtera, J. (2002). Organizational justice: Evidence of a new psychosocial predictor of health. American Journal of Public Health, 92, 105 108. Farkas, A. J., & Tetrick, L. E. (1989). A three-wave longitudinal analysis of the causal ordering of satisfaction and commitment on turnover decisions. Journal of Applied Psychology, 74, 855 868. Finlay, W., Martin, J. K., Blum, T. C., & Roman, P. M. (1995). Organizational structure and job satisfaction: Do bureaucratized organizations produce more satisfied employees? Administration and Society, 27, 427 450. Forman, R. F., Bovasso, G., & Woody, G. (2001). Staff beliefs about addiction treatment. Journal of Substance Abuse Treatment, 21, 1 9. Gallon, S. L., Gabriel, R. M., & Knudsen, J. R. W. (2003). The toughest job you ll ever love: A pacific northwest treatment workforce survey. Journal of Substance Abuse, 24, 183 196. Garman, A. N., Corrigan, P. W., & Morris, S. (2002). Staff burnout and patient satisfaction: Evidence of relationships at the care unit level. Journal of Occupational Health Psychology, 7, 235 241. Griffeth, R. W., Hom, P. W., & Gaertner, S. (2000). A meta-analysis of antecedents and correlates of employee turnover: Update, moderator tests, and research implications for the next millennium. Journal of Management, 26, 463 488. Hage, J., & Aiken, M. (1967). Program change and organizational properties: A comparative analysis. American Journal of Sociology, 72, 503 519. Halbesleben, J. R. B., & Buckley, M. R. (2004). Burnout in organizational life. Journal of Management, 30, 859 879. Hall, R. (1963). The concept of bureaucracy: An empirical assessment. American Journal of Sociology, 69, 32 40. Hiller, M., Knight, K., Broome, K., & Simpson, D. (1998). Legal pressure and treatment retention in a national sample of long-term residential programs. Criminal Justice and Behavior, 25, 463 481. Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley, CA7 University of California Press.

180 H.K. Knudsen et al. / Journal of Substance Abuse Treatment 31 (2006) 173 180 Hodson, R. (2001). Dignity at work. New York7 Cambridge University Press. Hser, Y. I. (1995). Drug treatment counselor practices and effectiveness: An examination of literature and relevant issues in a multilevel framework. Evaluation Review, 19, 389 408. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1 55. Kahill, S. (1988). Symptoms of professional burnout: A review of the empirical evidence. Canadian Psychology, 29, 284 297. Karasek, R. A. (1979). Job demands, job decision-latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285 308. Kelloway, E. K. (1998). Using LISREL for structural equation modeling: A researcher s guide. Thousand Oaks, CA7 Sage. Kessler, R. C., Chiu, W. T., Demler, O., & Walters, E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the national comorbidity survey replication. Archives of General Psychiatry, 62, 617 627. Knudsen, H. K., Ducharme, L. J., Roman, P. M., & Link, T. (2005). Buprenorphine diffusion: The attitudes of substance abuse treatment counselors. Journal of Substance Abuse Treatment, 29, 95 106. Lamb, S., Greenlick, M. R., & McCarty, D. (Eds.). (1998). Bridging the gap between practice and research: Forging partnerships with community-based drug and alcohol treatment (pp. 1 15). Washington, DC7 National Academy Press. Laundergan, J. C., Flynn, D., & Gaboury, J. D. (1986). An alcohol and drug counselor training program: Hazelden Foundation s trainee characteristics and outcomes. Journal of Drug Education, 16, 167 179. Lee, R. T., & Ashforth, B. E. (1996). A meta-analytic examination of the correlates of the three dimensions of job burnout. Journal of Applied Psychology, 81, 123 133. Leventhal, G. S. (1980). What should be done with equity theory? In K. J. Gergen, S. Greenberg, & R. H. Willis (Eds.), Social exchange: Advances in theory and research (pp. 25 55). New York7 Plenum. Lum, L., Kervin, J., Clark, K., Reid, F., & Sirola, W. (1998). Explaining nursing turnover intent: Job satisfaction, pay satisfaction or organizational commitment? Journal of Organizational Behavior, 19, 305 320. Mark, T., Kranzler, H. R., & Song, X. (2003). Understanding US addiction physicians low rate of naltrexone prescription. Drug and Alcohol Dependence, 71, 219 228. Maslach, C., & Jackson, S. E. (1981). The measurement of experienced burnout. Journal of Occupational Behavior, 2, 99 113. Maslach, C., & Jackson, S. E. (1986). Maslach burnout inventory manual. (2nd ed.). Palo Alto, CA7 Consulting Psychologists Press. McLellan, A. T., Carise, D., & Kleber, H. D. (2003). Can the national addiction treatment infrastructure support the public s demand for quality care? Journal of Substance Abuse Treatment, 25, 117 121. McLellan, A. T., Lewis, D., O Brien, C., & Kleber, H. (2000). Drug dependence, a chronic medical illness: Implications for treatment, insurance, and outcome evaluation. Journal of the American Medical Association, 284, 1689 1695. Melnick, G., & De Leon, G. (1999). Clarifying the nature of therapeutic community treatment: The survey of essential elements questionnaire (SEEQ). Journal of Substance Abuse Treatment, 16, 307 313. Melnick, G., De Leon, G., Thomas, G., & Kressel, D. (2001). A clienttreatment matching protocol for therapeutic communities: First report. Journal of Substance Abuse Treatment, 21, 119 128. Melnick, G., De Leon, G., Hiller, M. L., & Knight, K. (2000). Therapeutic communities: Diversity in treatment elements. Substance Use and Misuse, 35, 1819 1847. Mortimer, J. T., & Lorence, J. (1995). Social psychology of work. In K. S. Cook, G. A. Fine, & J. S. House (Eds.), Sociological perspectives on social psychology (pp. 497 523). Boston, MA7 Allyn and Bacon. Muthen, L. K., & Muthen, B. O. (2004). Mplus user s guide. (3rd ed). Los Angeles7 Muthen & Muthen. Niehoff, B. P., & Moorman, R. H. (1993). Justice as a mediator of the relationship between methods of monitoring and organizational citizenship behavior. Academy of Management Journal, 36, 527 556. Osterman, P. (2000). Work reorganization in an era of restructuring: Trends in diffusion and effects on employee welfare. Industrial and Labor Relations Review, 55, 179 188. Parker, R. N., & Fenwick, R. (1983). The Pareto curve and its utility for open-ended income distributions in survey research. Social Forces, 61, 872 885. Pines, A., & Maslach, C. (1978). Characteristics of staff burnout in mental health settings. Hospital and Community Psychiatry, 29, 233 237. Prendergast, M. L., Podus, D., Chang, E., & Urada, D. (2002). The effectiveness of drug treatment: A meta-analysis of comparison group studies. Drug and Alcohol Dependence, 67, 53 72. Sacks, S., De Leon, G., Balistreri, E., Liberty, H. J., McKendrick, K., Sacks, J., et al. (1998). Modified therapeutic community treatment for homeless mentally ill chemical abusers: Sociodemographic and psychological profiles. Journal of Substance Abuse Treatment, 15, 545 554. Schaufeli, W. B., Van Dierendonck, D., & Van Gorp, K. (1996). Burnout and reciprocity: Towards a dual-level social exchange model. Work & Stress, 10, 225 237. Schmike, M., Ambrose, M. L., & Cropanzano, R. (2000). The effect of organizational structure on perceptions of procedural fairness. Journal of Applied Psychology, 85, 294 304. Shirom, A. (1989). Burnout in work organizations. In C. L. Cooper, & I. Robertson (Eds.), International review of industrial organizational psychology (pp. 25 48). New York7 Wiley. Shoptaw, S., Stein, J. A., & Rawson, R. A. (2000). Burnout in substance abuse counselors: Impact of environment, attitudes, and clients with HIV. Journal of Substance Abuse Treatment, 19, 117 126. Tekleab, A., Takeuchi, R., & Taylor, M. S. (2005). Extending the chain of relationships among organizational justice, social exchange, and employee reactions: The role of contract violations. Academy of Management Journal, 48, 146 157. Thomas, C. P., Wallack, S. S., Lee, S., McCarty, D., & Swift, R. (2003). Research to practice: Adoption of naltrexone in alcoholism treatment. Journal of Substance Abuse Treatment, 24, 1 11. Vahey, D. C., Aiken, L. H., Sloane, D. M., Clarke, S. P., & Vargas, D. (2004). Nurse burnout and patient satisfaction. Medical Care, 42 (2, Suppl), 57 66. Van Dierendonck, D., Schaufeli, W. B., & Buunk, B. P. (2001). Burnout and inequity among human service professionals: A longitudinal study. Journal of Occupational Health Psychology, 6, 43 52. Walsh, J. P., Ashford, S. J., & Hill, T. E. (1985). Feedback obstruction: The influence of the information environment on employee turnover intentions. Human Relations, 38, 23 36. Wright, T. A., & Cropanzano, R. (1998). Emotional exhaustion as a predictor of job performance and voluntary turnover. Journal of Applied Psychology, 83, 486 493.