Testing the Conceptual Path to Correctional Staff Safety: A Study of the Implementation of Unit Management in Two Medium Security State Institutions in the USA

Abstract

Unit management is the preferred management model in Federal, state and private correctional institutions in the United States and is a model used in Australia, New Zealand and South Africa. It has also been seriously considered in Slovenia. The opportunity for staff participation in decision-making that this model specifies has been so lauded that unit management has become almost synonymous in practice and in the corrections literature with decentralization and direct supervision. The model has been credited for reducing inmate violence, maintaining/regaining staff control, and improving safety for inmates and staff. The uniqueness of the research here is that it presents a comparative case study of staff perceptions of their safety and the structural differences at two prisons that both have decentralized unit management staff and centralized hierarchal staff. It was predicted that unit management staff, compared to non-unit management staff, would feel safer and would perceive more delegation of authority and more opportunities to participate in decision-making. This study reports how the path to staff safety, as theorized by the unit management model, is not always as predicted. Staff perceptions were compared on a safety index that was created by factor analysis of responses to organizational climate surveys and to indices of decentralization and communication. Multiple regression analyses were then run on the staff safety index. The results indicate that the delegation of authority for decision-making can have inverse effects on the perception of staff safety; it depends on the way unit management is implemented. [PUBLICATION ABSTRACT]

Headnote

Abstract

Unit management is the preferred management model in Federal, state and private correctional institutions in the United States and is a model used in Australia, New Zealand and South Africa. It has also been seriously considered in Slovenia. The opportunity for staff participation in decision-making that this model specifies has been so lauded that unit management has become almost synonymous in practice and in the corrections literature with decentralization and direct supervision. The model has been credited for reducing inmate violence, maintaining/regaining staff control, and improving safety for inmates and staff. The uniqueness of the research here is that it presents a comparative case study of staff perceptions of their safety and the structural differences at two prisons that both have decentralized unit management staff and centralized hierarchal staff. It was predicted that unit management staff, compared to non-unit management staff, would feel safer and would perceive more delegation of authority and more opportunities to participate in decision-making. This study reports how the path to staff safety, as theorized by the unit management model, is not always as predicted. Staff perceptions were compared on a safety index that was created by factor analysis of responses to organizational climate surveys and to indices of decentralization and communication. Multiple regression analyses were then run on the staff safety index. The results indicate that the delegation of authority for decision-making can have inverse effects on the perception of staff safety; it depends on the way unit management is implemented.

Keywords: Correctional personnel, Unit management, Direct supervision, Staffsafety, Delegation of authority, Decentralized management.

Introduction

Those researching correctional management (see Antonio & Young, 2011; Blevens, Cullen & Sundt, 2007; Brookes, Smith & Bennett, 2008; Lambert, Hogan & Allen, 2006) have argued that the prison environment, including structure and policy, leads to staffattitudes and behaviors, both positive and negative. Investigators (i.e., Blevins, Cullen & Sundt, 2007: Lambert & Paoline, 2008; Maahs & Pratt, 2001) refer to this conceptual flow as the "deprivation" or "prisonization" penal model. This study of the implementation and effects of unit management on staffperceptions of their safety and security follows this respected tradition.

Unit management was introduced into correctional facilities around 1973 (Wener, 2005) and is conceptually based on the prison literature and the array of penal policies. In comparison to the authoritative penal policy and organizational structure of the 1950s that Sykes (1958) saw as causing prison unrest, unit management has been argued to be a policy that increases staffcommitment, lessens staffalienation, and results in control over inmates without the need to overpower them (Fihla, 2001; Wener, 2005). Unit management also addresses the problem of poor communication among staff, and between staffand inmates (Yocum, Anderson, DaVigo & Lee, 2006; Gerard, 1991; Levinson, 1999), that Cressey (1961) attributed to the separate hierarchies for the custody and treatment branches of the prison organization. Hobbs & Dear (2000), however, raise some questions about the validity of the effectiveness of unit management in improving communication between staffand inmates to the level that inmates are comfortable discussing personal problems, like depression and self-harm, with the staff.

Unit management is a correctional policy that specifies a specific structure. For example, multi-disciplinary unit teams, housing small numbers of inmates together permanently, placing officers in the inmate housing units, and decentralized organizational hierarchy (Hobbs & Dear, 2000; Wener, 2005). The implementation of unit management has been reported to positively change various processes such as the opportunities for individual members of the team to participate in decision-making, and the proactive intervention that heightens security and prevents incidents (Wener, 2005; Yocum et al., 2006). These processes, in turn, affect the attitudes and behaviors of both staffand inmates. This structural model, as depicted in Figure 1, serves as a guide for the present study.

Unit management is thought to be one way of separating the inmate population into smaller groups to bring about the fairness and safety specified by the retribution policy of the 1970s and carried through to the present. Unit management today also seems to be compatible with both the "punishment as retribution" and "re-entry" sides of the current debate over penal policy. With authority decentralized and stafflocated in inmate living quarters for direct supervision, unit management teams are in a position to be immediate reinforcers of positive inmate behavior and observers of potentially dangerous issues that jeopardize staffand inmates. Since these unit management teams are multi-disciplinary, more choices (i.e., classes, counseling, and recreation) should be visible and available to inmates. Inmates should have the option, then, of making more positive use of their incarceration period, a hope expressed by Fihla (2004) in South Africa, Hobbs and Dear (2000) in Australia, Houston and Stefanoviae (1996) in Slovenia and many researchers in the United States.

The basic hypothesis for this present study was based on the presumed and documented success of the unit management structure and process (Figure 1) as it relates to staffbeliefs about their work environment, specifically their own safety. It was hypothesized that unit management staff, compared to non-unit management staff, would feel safer and would perceive more delegation of authority and more opportunities to participate in decisions regarding their work environment and issues relating to the inmates.

Literature Review

Unit Management

Most of the research has shown that the implementation of unit management is inextricably connected to and frequently referred to as direct supervision (Edwards, 2007; Wener, 2005) and results in improved staffattitudes towards their work and their department of corrections (Lambert, Hogan, Moore, Tucker, Jenkins, Stevenson & Jiang, 2009). It also leads to improved safety for staffand inmates (Edwards, 2007; Edwards, 2011; Nelson & Davis, 1995). This change is attributed to the increase in daily collaborations between staffand increased familiarity with inmate concerns (Cooksey & Carlson, 1997; Fihla, 2004, Wener, 2005).

Staffat prisons and jails that are decentralized (more delegation of authority and opportunities to participate in decision-making) have reported more favorable attitudes toward either the prison or their jobs (Guay, Senecal, Gauthier, & Fernet, 2003; Wener, 2005). Conversely, staffbecome more stressed when they perceive a poor communication, a lack of information input (Dowden & Tellier, 2004; Hogan, Lambert, Jenkins & Wambold, 2006; Lambert, Hogan & Allen, 2006) and little control over their environment (Yocum et al., 2006).

Support of the prison system and/or departments of corrections has also resulted from unit management. Edwards (2007) found that while the direct supervision afforded by unit management was criticized and viewed with trepidation 30 years ago, it is held in the highest regard now and is seen as the catalyst for more staffcontrol and a safer, more secure, correctional environment that has less violence. Wener (2005) also reported this history. Citing early penitentiary studies, he recalls how the increased commitment by staffto the system was a function of the continuous communication effort administrators undertook during the transition to unit management and after initial implementation.

In general, the decreased tension and increased safety resulting from unit management has led American staffand inmates to report more favorable attitudes towards prison work and prison living (Edwards, 2011; Freeman, 1997; Levinson, 1999; Wener, 2005). The same has been reported in Australia (Robson, 1989) and South Africa (Fihla, 2004). On the other hand, studies of prisons and jails without unit management or some form of staffparticipation in decision-making have consistently shown negative staffattitudes towards their jobs and low commitment toward either the prison administration or the department of corrections (Lambert, et al., 2006; Tewksbury & Higgins, 2006).

In South Africa, Fihla (2004) reported that a new Constitution called for prisons previously designed and managed for punishment should be converted to the humanitarian unit management model. He cited all the advantages and successes of unit management in the United States, Australia and New Zealand. Among the advantages were increased staffsafety, more positive relations between staffand inmates, better control and more direct supervision of inmates and the reduction of incidents related to overcrowding. The Robson (1989) study referenced by Fihla (2004) was Robson's inquiry into an Australian prison, where the improvements in prison conditions were attributed to the increased frequency of contact and communication among staffand between inmate and staff. This was positively related to the degree of authority delegated to the staffunder the unit management model.

Applegate and Paoline (2007), however, found something different when they studied the perceptions of staffin a traditionally centralized facility and the perceptions of staffworking in a newer and decentralized unit facility. They had hypothesized that the staffin direct-supervision inmate units would report more job satisfaction, more autonomy and control over inmates, fewer inmate attacks on staffand a safer environment. They discovered the opposite of what they had predicted. The jail officers in the directsupervision units perceived being less safe and thought they had less decision-making authority over the inmates. The speculation was that the stafffelt their jobs were more dangerous because they were required to walk around in the inmate units and were not protected by physical barriers. The direct-supervision staffalso did not believe their communication with inmates had improved.

Organizational Theory

Decentralized organizations, like prisons utilizing unit management, tend to be characterized by greater communication between upper and lower management and an increased capacity to respond to the environment (Abu-Jarad, Yusof & Nikbin, 2010; Hennestad, 2000). The authority that is delegated often goes to a team, or work group, rather than to an individual. The delegation of authority implies the delegation of decision-making (Ismail, Mohamed, Sulaiman, Mohamad & Yusuf, 2011), but the amount of delegated authority and the opportunities to participate in decision-making can vary. There can be, for example, little delegated authority, but widespread participation in decision-making. Lower level involvement can range from consultation to the manager, to limited input in the decision, to full responsibility for the choice of action to be taken.

It has been recognized that decentralized prisons and corporations can better meet the social and psychological needs of their members than can a larger, more impersonal bureaucratic systems (Boone & Hendricks, 2009). It has also been found that empowerment, because of the opportunities it offers for participation in decision-making, can lead to improved group member satisfaction (Boone & Hendricks, 2009; Hatvany & Gladstein, 1982; Hennestad, 2000; Lambert et al., 2006; Wright, Saylor, Gilman & Camp, 1997). Some researchers (Dowden & Tellier, 2004; Slate, Vogel & Johnson, 2001) have reported that prison staffare less stressed when empowered to make decisions. In addition, empowerment has produced a more positive social climate in foreign countries and companies (Galle & Leahy, 2009; Gill, Mathur, Sharma, & Bhutani, 2011; McLaurin, 2008; Vasugi, Kaviatha, Fabiyola & Prema, 2011). Mutual trust has been increased between the management and the worker (Akbar, Yousaf, Haq, & Hunjra, 2011; Arnetz & Blomkvist, 2007).

There is considerable agreement in the behavioral science literature that staffempowerment can benefit the participants as well as the organization that adopts this policy. Being a valued group member not only results in a sense of power, but job satisfaction (Stanley, 2011; Vasugi et al., 2011; Yao & Cui, 2010), security and friendship (Raquib, Lama, Anantharaman, Eze, & Murad, 2010). The organization benefits from the resultant increase in group member motivation and individual and group productivity, as well as from increased problem-solving ability and commitment to the organization (Tjosvold & Sun, 2006).

Staffempowerment, however, is by no means a panacea for all management problems. In fact, empowerment can create new problems. It has been found that failures with staffempowerment frequently occur because of the way it is implemented. If employees are pressured into participating when they have not interest in doing so, officers can become resentful towards their supervisor (Galle & Leahy, 2009). In addition, participants in the joint decision-making process often flounder when there is no clear direction (Chacar & Suryekar, 2009), or become discouraged if the group leader dominates the decisionmaking process (Siddique, Khan & Fatima, 2011).

Staffadjustment to empowerment can also be fraught with problems. Siddique et al. (2011) found the tension between new participative systems and traditional organizational hierarchies if extra funds did not follow the new empowerment model. Researchers (Chacar & Suryekar, 2009; Stevenson & Gumpert, 1985) have cited the difficulty of accommodating staffempowerment within an organization used to stability and control. Another study (Muczyk & Reimann, 1987) suggests that the model has been oversold at the expense of the follow-up. It was found that employees in small organizations, who were given increased opportunities to participate in decision-making, often lacked the initiative and experience necessary to execute the decisions. Managers who supervised the decision-making group often had to step in and follow up to ensure that decisions were carried out.

The idea that staffempowerment is effective only under certain conditions has support in traditional organizational research (Chacar & Suryekar, 2009). In addressing psychological contingencies, Vroom & Yetton (1973) found that efforts to generate participative involvement failed if the organization members resisted participation. Chacar & Suryekar (2009) show clearly that production and morale can benefit from "less close" supervision, but that efforts to increase participative involvement will not likely succeed if employees have low trust in managers who are seen as autocratic. Nor is success likely if supervisors resist employee involvement (Klein, 1984).

Staffempowerment can also fail if the people in control of the organization have an insincere commitment to the philosophy (Siddique et al., 2011). McGregor (1960) noted, 'The not infrequent failure of such ideas as these to work as well as expected is often attributable to the fact that a management has 'bought the idea' but applied it within the framework of Theory X and its assumptions' (p. 277).

Methodology

Prior to the collection of data for this study of staffsafety, it was reasonable to expect that the staffdesignated as "unit management" would perceive more delegation of authority and more opportunities to participate in decision-making than would "non-unit management" staff. It was also predicted that "unit management" staff, compared to "nonunit management" staff, would evaluate more highly the security of their environment (i.e., being safe, control over inmates, controlling for contraband, likelihood of being assaulted, and access to the inmate pipeline).

Research Sites

Both correctional institutions are medium security and overcrowded. Both have a formal decentralized management organization and written policies that resemble Federal Bureau of Prisons standard specifications. These two facilities were chosen to minimize the confounding effects of such factors as security level and type of inmate on measures of social climate. There are three unit management teams at Glenway. Each of these teams included a group manager, a sergeant, a correctional counselor, at least two correctional officers, and several relief personnel. There was an average of 10 unit management staffand 246 male inmates assigned to each unit management team. The ratio of unit management staffto inmates under their supervision is 1 to 25. At the time of this study, Glenway housed about 737 inmates and employed about 310 staff.

At Senoia, the single unit management team consisted of 20 staff, including a team manager, one sergeant, five correctional officers, four correctional counselors, and several relief personnel. They were all assigned to the east side of the facility, which included four separate housing units with a total of 150 male inmates. The Senoia unit management staffto inmate ratio was 1 to 7.5, one third the average Glenway unit management staffto inmate ratio. Senoia housed about 644 inmates and employed about 350 staff.

Participants

The participants were 57 staff(26 unit management and 31 non-unit management) from Glenway and 92 (16 unit management and 76 non-unit management) from Senoia correctional institutions (pseudonyms are used). Respondent's participation was completely voluntary and was determined by saturation sampling. Different days and times during all work shifts were arranged for the volunteers to be released from work to complete the survey. The median respondent age was 34.2. There were 112 (75 percent) males and 37 (25 percent) females. The staffhad worked at their respective prisons for an average of almost six years. Their median education was two years of college, 24 percent had a bachelor's degree, and only two percent of them had not completed high school. Almost 60 percent of the staffrespondents were security personnel, and 40 percent were treatment personnel. Stafffrom all shifts participated.

Survey Instrument

To analyze the effects of the unit management specified structural changes on the correctional institutions, a questionnaire was adapted with permission from the Saylor, Gaes and Vanyur (1987) "Prison Social Climate Survey - StaffVersion." This questionnaire has been used by the Federal Bureau of Prisons to track climate changes in the Federal system. The survey used for this study is retrospective and is limited to issues related to staffsafety items. One addition made to the questionnaire instructions was to limit the respondents to a six month time reference. This limitation of the period of reflection was to maximize reference to the "same" climate while providing a time span large enough to collect sufficient data for comparison.

Analytical Sequence

Several indices of the delegation of authority and the staffsafety/security environment were created by factor analysis of responses of Glenway and Senoia staffto the survey adapted from Saylor et al. (1987). With unit management (staffstatus was determined by upper management) as the independent variable and these indices as the dependent variables, several hypotheses were tested by bivariate analysis to determine if the unit management stafffaired better than the non-unit management staffon the delegation of authority and safety issues. Then several of these indices were regressed on the major safety/security index to test the conceptualized path to staffsafety.

Methods of Index Development

Unit management calls for the delegation of authority and opportunities to participate in decision-making. These two concepts were measured separately to distinguish authority from participation. The items comprising the two indices (Table 1) were chosen on face validity to determine the extent to which these two elements existed. In addition to these indices, three more were created to measure the communication level between staffand inmates. The first of these was frequency of staff/inmate contact that consisted of only one question, "How often do you interact with inmates?" The second index, access to inmate pipeline, asked, "How much access do you have to the inmate information network." And the third, knowing inmates, asked "How well do you personally know the inmates you are responsible for?"

A factor analysis was then employed on the staffresponses to all questions in the Saylor et al. (1987) survey relating to the staffsafety and security. Principal components analysis (to explain the observed variables' variance) with orthogonal varimax rotation produced six factor indices that accounted for 72.6 percent of the total variance among the 21 items entered. The factor loadings of the index items, a measure of reliability (Cronbach's alpha) and percents of factor variance are provided in Table 2. In keeping with the justifications of Nunnally (1967, p. 24-26), multiple regression was used for this purpose, despite the fact that the scales used are all ordinal and do not meet the necessary interval assumptions. Moreover, in keeping with acceptable standards (Johnson & Wichern, 1982; Nunnally, 1967), only factor items with loadings of .40 or higher were used to create factor scores.

Six indices were constructed by simply summing the standardized scores on each item. Following Kachigan (1982, p. 254), it was decided to retain the high negatively loaded items because, despite the loading, the variables contributed to the interpretation of the factors. This means that a staffmember scoring high on the negatively loaded item will score lower on this factor. To construct factor scores for those items with negatively loaded items, the researcher followed Johnson and Wichern's (1982, p. 436) advice and added the standardized scores in the factor according to the sign of the loading. In other words, negatively loaded items were summed as negative.

The first factor, safety in numbers, was be used to measure an interesting combination of attitudes towards the degree of inmate freedom to move and whether there are enough staffto keep staffsafe. It is apparent from the high negative loadings that there is an inverse correlation between the staffperception of inmate freedom to move and staffperception of staffsafety. In other words, if staffscored lower on the freedom to move the item, they would feel safer.

The second index, control over inmates, is a combination of questions that measure staffperception of their control over the inmates during all three work shifts, including the shifts the respondents do not work. The third index, safe from injury, measures the extent to which there is danger of physical injury to male and female staff. There was a .70 correlation between the first two items relating to injury. The item regarding weapons correlated from .39 to .43 with those.

The fourth index, likelihood of assault, combines several items measuring the possibility of inmate assaults on staff, but differs from the previous index in that there is no notion of injury. Also, this index includes the degree to which staffis bothered by potential assaults. The fifth index, elimination of contraband, is composed of items that address the frequency in which different methods are used to eliminate drugs, weapons and other contraband which are common causes of arguments and assaults. A high negative loading showed up on the third item that factored here: whether or not a staffmember had been assaulted by an inmate. Yet, neither of the shakedown questions correlate with the assault question (.07 and .05). As one might expect, however, the shakedown items correlate .56 with each other. It could well be that the commonality that caused these three items to factor together is the association between shakedowns and assaults. Shakedowns are viewed as a way of preventing assaults, yet there is always a danger of being assaulted while shaking down. In any event, only 10 (6.7 percent) out of the 149 staffrespondents reported having been assaulted, so this item was eliminated from the index and treated separately.

The sixth index, contact with AIDS inmates (contact with inmates known to have, or suspected of having, AIDS), measures a major safety concern for staff. The only AIDS questions in the questionnaire loaded together.

Testing of Decentralization and Safety Indices

After creating the appropriate indices, and in final preparation for the regression testing of the conceptual path for staffsafety, a number of tests were run to determine if there was a statistical difference between how unit management and non-unit management compared at Glenway and Senoia. The results of these tests are in Table 3.

As can be seen, Glenway unit management staffreported significantly less delegation of authority and fewer opportunities to participate in decision-making than did the non-unit management staffthere. This was particularly odd and cast doubt as to whether the unit management model at Glenway was being practiced as the standard model specifies. Although Glenway had a flatter organizational structure, as did Senoia, the Glenway nonunit management staffreported more decentralization than the unit management staffreported. At Senoia, the unit management staffreported significantly more delegation of authority, but their decision-making opportunities, while in the right direction, was not significantly different than the non-unit management staff.

Although the primary purpose of this study is not to focus on these bivariate results, it is worth noting that for Glenway the unit management staff, compared to the non-unit management staff, reported significantly more favorable perceptions of safety in numbers, eliminating contraband and the frequency of staff/inmate contact. At Senoia, the unit management staff, compared to the non-unit management staff, reported significantly less favorable perceptions of eliminating contraband and contact with AIDS inmates.

What was unexpected in these tests at Glenway was that the unit management staff(compared to the non-unit management staff), reported being safe and having more contact with inmates, but at the same time perceiving having less authority and fewer chances to influence their environment. Also unpredicted were the non-significant results when the perceptions of the unit management and non-unit management staffat Glenway and Senoia were compared on such indices as control over inmates, feeling safe from injury, likelihood of assault, access to the inmate pipeline and knowing inmates.

Conceptual Path to StaffSafety

The next step in the research process was to conceptualize the path to staffsafety. As shown in Figure 2, it was conceptualized that unit management would first lead to more opportunities to participate in decision-making. This index, which measured the respondent's actual experience, would lead to the reporting of more delegation of authority, which was the respondent's perception of the more general practice. The delegation of authority would lead to greater frequency of staff/inmate contact because unit management staffwould have direct responsibility for the supervision of inmates. Greater frequency of staff/inmate contact would result in the staffbetter knowing the inmates. Conceptually, higher scores on knowing inmates would result in the staffhaving greater access to the inmate pipeline.

With greater access to this inmate communication network, the staffshould have more information about inmate plans for riots, fights and other behavior that threatens staffsafety and security. Staff, knowing they can take action to prevent this threatening behavior, should feel safer. The index used to measure staffsafety and security was safety in numbers. This index accounted for 25.5 percent of the total variance of all items entered in the safety and the security factor analysis.

Findings of Multiple Regressions on StaffSafety

Several categories of staffand inmates (independent variables) were used in the regression equations as dummy variables. In each of the equations for staff, there were four categories that designated unit management status (Glenway unit management, Glenway non-unit management, Senoia unit management and Senoia non-unit management, and two categories that designated security status (security and treatment). The Senoia nonunit management staffand the treatment staffwere leftout of the staffequations. These variables that were leftout were the control groups, against which the other variables in those categories were measured.

Controlling for the effects of unit management status and security jobs (versus treatment jobs), safety in numbers was regressed on access to inmate pipeline, knowing inmates, frequency of staff/inmate contact, delegation of authority, and opportunities to participate in decision-making. In general, it was found that most of the relationships between the indices were in the predicted direction, but only part of the conceptualized path was significant. Table 4 shows the results of the multiple regressions.

The first regression equation (adjusted R square = .31) (Table 4) revealed that both the unit management and non-unit management staffat Glenway, compared to the non-unit management staffat Senoia, reported being safer and more secure, as measured by the safety in numbers index. The beta coefficients for the unit management staffat Glenway and Senoia were .31 and .34, respectively, and they were both significant at p<.001. The security staff, compared to the treatment staff, was also significantly safer and more secure. The delegation of authority, however, had a significant negative direct effect on safety and security.

A possible explanation for this is that the perceived delegation of authority is over security issues, and the more one perceives a loss of authority through delegation, the more one perceives a loss of control over those issues, and the less secure one feels. Knowing inmates and having opportunities to participate in decision-making also had a negative direct effect on safety and security, although the strength of the effect was not significant.

Knowing inmates had a significant positive direct effect (with a beta coefficient of .32) on access to inmate pipeline in the second equation (adjusted R square =.14), as conceptualized. Opportunities to participate in decision-making had a negative direct effect on access to inmate pipeline. This relationship was weak (p<.10) and not as predicted. The Senoia unit management, compared to the Senoia non-unit management staffreported significantly more access to inmate pipeline. The other variables contributed very little in this equation.

In the third equation (adjusted R square = .12), the best predictor of knowing inmates was the frequency of staff/inmate contact, with a beta coefficient of .35. This was significant at the .001 level. The other variables contributed very little to the explanation. The fourth equation (adjusted R square = .08) failed to support the predicted positive direct effect of the delegation of authority on the frequency of staff/inmate contact. Instead, it was revealed that the delegation of authority had a slight, although not significant, negative effect. In the fifth equation (adjusted R square = .35), opportunities to participate in decision-making was the best, and only, a significant predictor of the delegation of authority. The beta coefficient was .56 and was significant at the .001 level. The direction of this relationship was as predicted.

In the sixth equation (adjusted R square = .10), the best predictor of opportunities to participate in decision-making was the security staff, compared to the treatment staff(control group). The beta weight was -.21, which was significant at p<.01. The security staffwas reporting significantly fewer opportunities to participate in decision-making. Another good predictor of opportunities to participate in decision-making was the nonunit management staffat Glenway, compared to the Senoia non-unit management staff(control group). The beta coefficient was .17, which was significant at p<.05. The unit management staffat Senoia came close to being a significant predictor (p<.10) of opportunities to participate in decision-making.

Figure 3 shows the path model based on these results. It had been conceptualized that the unit management staffat both Senoia and Glenway, compared to the non-unit management staffat their respective prisons, would report more opportunities to participate in decision-making. In this regression, only the Senoia unit management staff, compared to the Senoia non-unit management staff, reported significantly more opportunities to participate in decision-making (at p<.10). As shown in Figure 3, opportunities to participate in decision-making, as conceptualized, had a direct positive effect on the delegation of authority.

A portion of the conceptualized path to staffsafety was significant, however. The frequency of staff/inmate contact had the conceptualized direct positive effect on knowing ights reserved. Under a creative commons Attribution-Noncommercial-Share Alike 2.5 India License inmates, and knowing inmates had the predicted direct positive effect on access to inmate pipeline. Access to inmate pipeline, however, was not significantly related to safety in numbers. Despite this, the unit management and non-unit management staffat Glenway, as compared to the unit management and the non-unit management (control group) staffat Senoia, and the security staff, as compared to the treatment staff, felt safer and more secure. It could be that the Glenway staffreported being safer than the Senoia staffbecause 63 percent of the Glenway sample was security, compared to 54 percent of the Senoia sample, and security stafffelt safer than treatment staff.

It is interesting to look at the indirect effects of the variables (Figure 3) on staffsafety. As previously mentioned, the security staffreported fewer opportunities to participate in decision-making, which was a significant predictor of the delegation of authority, and the delegation of authority had a significant direct negative effect on safety. But because there are two significant negative relationships in the indirect path from security to safety, what is found is that security staffreally has a positive indirect effect on safety.

It is also interesting to note that the Senoia unit management staff, through opportunities to participate in decision-making and delegation of authority, report being less safe than the Senoia non-unit management staff. This is probably due to the fact that the unit management staffsample at Senoia has more treatment personnel in it than the non-unit management staffsample there, and treatment staff, compared to security staff, report being less safe.

Discussion

The unit management staffrespondents at Senoia reported significantly more delegation of authority than the Senoia non-unit management staff. While not significantly different, the Senoia unit management staffalso reported more opportunities to participate in decision-making. Thus, the first expected outcome was, in fact, found at Senoia.

The expected results were not found at Glenway, however. The staffat Glenway who were said to have been "unit management" reported significantly less delegation of authority and fewer opportunities to participate in decision-making than did the non-unit management staff, although Glenway unit management staffhad a flatter organizational structure than did the Glenway non-unit management staff. Follow-up interviews at Glenway revealed that the managers of unit management teams held few team-decision meetings and often made the decisions themselves. Although their subordinate staffacknowledged being empowered by top management, they did not feel they were members of a team and they felt they had no actual power.

Possible explanations for the Glenway non-unit management staffreporting more decentralization of authority are both structural and procedural. All of the inmates at Glenway (which was operating at 401 percent of bed capacity) were under the supervision of one of three unit management teams. The unit management staffto unit management inmate ratio was 1 to 25 (compared to 1 to 7.5 at Senoia), and the Glenway unit management staffwere responsible for the supervision of so many inmates (in overcrowded living quarters) that there was little time for team consultations and decisions. Glenway unit management staff, in follow-up interviews, claimed that the three unit managers often made decisions by themselves, without consulting the other team members, with whom they meet formally only once a week. The non-unit management staffunder the traditional centralized organizational structure, on the other hand, has decisions and authority delegated to them more frequently, because they have less direct responsibility for inmates.

At Senoia, unlike at Glenway, it was the unit management staff, compared to the nonunit management staff, who reported better scores on the decentralization of authority indices, the delegation of authority and opportunities to participate in decision making. Senoia, as compared to Glenway, had both unit management and non-unit management inmates and was operating at 323 percent of bed capacity. Glenway was operating at 401 percent capacity. There was only one unit management team at Senoia, and the unit management staffto unit management inmate ratio was 1 to 7.5, as compared to three unit management teams and a unit management staffto unit management inmate ratio of 1 to 25 at Glenway. Half of the unit management sample at Senoia were correctional counselors. Some of these correctional counselors claimed, in follow-up interviews, that there was so much delegation of authority and so many decisions to be made about inmates that they had difficulty in getting the paperwork done when needed.

Conclusion

In tough economic times, politicians and correctional administrators are compelled to consider deep budget cuts. Some state correctional institutions have already abandoned unit management because maintaining the model-specified unit manager positions could not be fiscally justified. Future research into the unit management/direct supervision model should focus on facilities that have cut this humanistic decentralized management technique. Have there been increased incidents of inmate/inmate and inmate /staffassaults? Do staffreport feeling less safe? Has there been increased staffnegativity towards their prison or their department of correction?

What was found in this case study of two medium security state institutions was that unit management staffcompared to non-unit management staffreported feeling safer and having more access to the inmate pipeline. Security staff, in particular, perceived themselves safer than did treatment staff. The lower scores on decentralization by the Glenway unit management staff, as compared to Glenway non-unit management staffand the Senoia unit management staff, provides evidence that unit management was not properly managed at Glenway. The "unit management" staffwas under a shortened bureaucratic hierarchy, but the staff-to-inmate ratio (1:25) was way too large and the staffwere not reporting the decision-making opportunities and delegated authority that was to accompany the structural change. This might follow from some of the research that has reported too little direction and follow through by management to be a major factor in failed unit management efforts.

It is evident that for more of the lauded and desired results of unit management to be obtained, the staff-to-inmate ratio must be kept low and the unit management staffshould have good relations with, and support from, upper management. Researchers have predicted the failure of unit management if this were not the case. Unit teams should be given the delegation of authority, and the necessary resources, time and training. The training should be in the areas of staffmanagement, team communication, commitment to a shared correctional mission, and methods of modern data collection to evaluate the effectiveness of this model. The continued use of unit management should be supported, but great care should be taken in its implementation.

The direct supervision of inmates through the delegation of authority to unit teams operating in the inmate housing areas, although often more stressful for correctional staff, can potentially improve staff/inmate and staff/staffcommunication, heighten vigilance over safety and security issues and, thus, maintain the humane working and living environment called for by government constitutions, bureaus of prisons, state departments of correction and municipal jails.

References

References

Abu-Jarad, I., Yusof, N., & Nikbin, D. (2010). A review paper on organizational culture and organizational performance. International Journal of Business and Social Science, 1(3), 26-46.

Akbar, S., Yousaf, M., Haq, N., & Hunjra, A. (2011). Impact of employee empowerment on job satisfaction: an empirical analysis of Pakistani service industry. Interdisciplinary Journal of Contemporary Research in Business, 2(11), 680-685.

Antonio, M., & Young, J. (2011). The effects of tenure on staffapathy and treatment orientation: a comparison of respondent characteristics and environment factors. American Journal 0f Criminal Justice, 36, 1-16.

Applegate, B. & Paoline, E. (2007). Jail officers' perceptions of the work environment in traditional versus new generation facilities. American Journal of Criminal Justice, 32(2), 64-80, 4-5.

Arnetz, B., & Blomkvist, V. (2007). Leadership, mental health, and organizational efficacy in health care organizations. Psychotherapy and Psychosomatics, 76, 242-248.

Blevins, K., Cullen, F. & Sundt, J. (2007). The correctional orientation of "child savers": support for rehabilitation and custody among juvenile correctional workers. Journal of Offender Rehabilitation, 45, 47-83.

Boone, C., & Hendricks, W. (2009). Top management team diversity and firm performance: moderators of functional-background and locus-of-control diversity. Management Sciences, 55(2), 165-180.

Brooks, S., Smith, K. & Bennett, J. (2008). The role of middle and first-line managers. In: J. Bennett, B. Crewe & A. Wahidin (eds.) Understanding prison staff. Willan, Portland, OR, 262-276.

Chacar, A. S., & Surysekar, K. (2009). A survey of new product development: Can decentralization alone deliver? The Review of Business Information Systems, 13(4), 79-87.

Cooksey, M. & Carlson, P. (1997). Emergency preparedness: A BOP follow-up report. Corrections Today, 59(1), 34-36.

Cressey, D. (1961). The prison: studies in institutional organization and change. New York: Rinehart, Winston.

Dowden, C., & Tellier, C. (2004). Predicting work-related stress in correctional officers: a meta-analysis. Journal of Criminal Justice, 19(2), 31-47.

Edwards, E. (2007). A constant commitment to enhanced safety and security. Nashville, TN: Corrections Corporation of America.

Edwards, E. (2011). Efficient, effective and evidence-based: Unit management with Corrections Corporation of America. Nashville, TN: Corrections Corporation of America.

Fihla, N. (2004). Unit management in prisons: briefing by national commissioner. PMG. South Africa: Parliamentary Monitoring Group.

Freeman, R. (1997). Management and administrative issues. In J. Pollack (Ed.), Prisons Today and Tomorrow (p. 300). Gaithersburg, MD: Aspen.

Galle, B., & Leahy, J. (2009). Laboratories of Democracy? Policy Innovation in Decentralized Governments. Emory Law Review, 58(6), 1333-1400.

Gerard, R. (1991). The ten commandments of unit management. Corrections Today, 53(2), 32-36.

Gill, A., Mathur, N., Sharma, S., & Bhutani, S. (2011). The effects of empowerment and transformational leadership on employee intentions to quit: A study of restaurant workers in India. International Journal of Management, 28(1), 217-229.

Guay, F., Senecal, C., Gauthier, L., & Fernet, C. (2003). Predicting career indecision: A self-determination theory perspective. Journal of Counseling Psychology, 50, 165-177.

Hatvany, N. G., & Gladstein, D. (1982). A perspective on group decision making. In D. A. Nadler, M. L. Tushman & N. G. Hatvany (eds.), Managing Organizations. Boston: Little, Brown and Company.

Hennestad, B. (2000). Implementing participative management: Transition issues from the field. The Journal of Applied Behavioral Science, 36(3), 314-335.

Hobbs, G. & Dear, G. (2000). Prisoners' perceptions of prison officers as sources of support. Journal of Offender Rehabilitation, 31(1/2), 127-142.

Hogan, N., Lambert, E., Jenkins, M., & Wambold, S. (2006). The impact of occupational stressors on correctional stafforganizational commitment: A preliminary study. Journal of Contemporary Criminal Justice, 22, 44-62.

Houston, J., & Stefanoviae, D. (1996). Corrections in a new light: Developing a prison system for a democratic society. Policing in Central and Eastern Europe: Comparing firsthand knowledge with experience from the west. Slovenia: College of Police and Security Studies.

Ismail, A., Mohamed, H., Sulaiman, A., Mohamad, M., & Yusuf, M. (2011). An empirical study of the relationship between transformational leadership, empowerment and organizational commitment. Business and Economics Research Journal, 2(1), 89-107.

Johnson, R. A., & Wichern, D. (1982). Applied Multivariate Statistical Analysis. Englewood Cliff, NJ: Prentice Hall.

Kachigan, S. (1982). Multivariate statistical analysis: a conceptual introduction. New York: Radius Press.

Klein, J. A. (1984). The human costs of manufacturing reform. Harvard Business Review, 67, 60-66.

Lambert, E., Hogan, N. & Allen, R. (2006). Correlates of correctional officer job stress: The impact of organizational structure. American Journal of Criminal Justice, 30(2), 227- IV.

Lambert, E., Hogan, N., Moore, B., Tucker, K., Jenkins, M., Stevenson, M., & Jiang, S. (2009). The impact of the work environment on prison staff: Tissue of consideration, structure, job variety, and training. American Journal of Criminal Justice, 34, 166-180.

Lambert, E. & Paoline, E. (2008). The influence of individual, job, and organizational characteristics on correctional staffjob stress, job satisfaction, and organizational commitment. Criminal Justice Review, 33(4), 541-564.

Levenson, R. (1999). Unit management in prisons and jails. Lanham, Md: American Correctional Association.

Maahs, J. & Pratt, T. (2001) Uncovering the predictors of correctional officers' attitudes and behaviors: a meta-analysis. Corrections Management Quarterly, 5, 13-19.

McGregor, D. (1960). The Human Side of Enterprise. New York: McGraw-Hill, Inc.

McLaurin, J. R. (2008). Teamwork in the city of gold: An Arabian adventure of modern management capabilities in the 21st century continues. Proceedings of the Academy for Studies in International Business, 8(1), 51-55.

Muczyk, J. P., & Reimann, B. (1987). Has participative management been oversold? Personnel, May, 52-56.

Nelson, W. & Davis, R. (1995). Podular direct supervision: The first twenty years. American Jails, 9(3), 11-22.

Nunnally, J. (1967). Psychometric Theory. New York: McGraw, Inc.

Raquib, A., Anantharaman, R. N., Eze, U. C., & Murad., W. (2010). Empowerment practices and performance in Malaysia - An empirical study. International Journal of Business and Management, 5(1), 123-148.

Robson, R. (1989). Managing the long term prisoner: a report on an Australian innovation in unit management. The Howard Journal of Criminal Justice, 28, 187-203.

Saylor, W. G., Gaes, G. G., & Vanyur, S. D. (1987). Prison Social Climate Survey" U.S. Department of Justice. Federal Bureau of Prisons.

Siddique, A., Khan, M, & Fatima, U. (2011). Impact of academic leadership on faculty's motivation and organizational effectiveness in higher education system. International Journal of Business and Social Science, 2(8), 184-191.

Slate, R., Vogel, R., Johnson, W. (2001). To quit or not to quit: perceptions in correctional decision-making and the impact on organizational stress. Corrections Management Quarterly, 5, 68-78.

Stanley, M. (2011). Why do we find it so hard to delegate? Contract Management, 12-13.

Stevenson, H., & Gumpert, D. (1985). The heart of entrepreneurship. Harvard Business Review, 63, 85-94.

Sykes, G. (1956). The society of captives. New York: Random House.

Tewksbury, R., & Higgins, G. (2006). Prison staffand work stress: The role of organizational and emotional influences. American Journal of Criminal Justice, 30(2), 247- 270.

Tjosvold, D., & Sun, H. (2006). Effects of power concepts and employee performance on managers' empowering. Leadership & Organizational Development Journal, 27(3), 217- 234.

Vasugi, M. S. P., Kaviatha, S., Fabiyola, S. & Prema, R. (2011). An empirical investigation on employee empowerment practices in Indian software industries. International Journal of Contemporary Research in Business, 2(11), 668-674.

Vroom, V. H. & Yetton, P. W. (1973). Leadership and Decision Making. Pittsburg, PA: University of Pittsburg Press.

Wener, R. (2005). The invention of direct supervision. Corrections Compendium, 30(2), 4- 32.

Wright, K., Saylor, W., Gilman, E., & Camp, S. (1997). Job control and occupational outcomes among prison workers. Justice Quarterly, 14, 525-546.

Yao, K., & Cui, X. (2010). Study on the moderating effect of the employee psychological empowerment on the enterprise employee turnover tendency: Taking small and middle enterprises in Jinan as the example. International Business Research, 3(3), 21-31.

Yocum, R., Anderson, J., DaVigo, T., & Lee, S. (2006) Direct-supervision and remotesupervision jails: A comparative study of psychological factors. Journal of Applied Social Psychology, 37, 1790-1812.