Email: pri@po.state.ct.us www.cga.state.ct.us/pri

DIGEST OF FINDINGS AND RECOMMENDATIONS

A. Results of Statistical Tests on Number of Custody Staff

B. Results of Statistical Tests on Incidents of Inmate Misconduct

C. Correction Officer Survey Form

D. Correction Officer Survey Results

E. Agency Response

The objective of this study was to determine if the current Connecticut Department of Correction (DOC) custodial staffing levels are sufficient for the safe and efficient management of the state's prison population.

In fiscal year 2002, the Connecticut Department of Correction operated 18 correctional facilities and expended $533.6 million, of which all but $20.3 million came from the state's General Fund. The average daily inmate population was 18,288 and the department had 6,587 filled staff positions on June 30, 2002.

As shown in Figure I-1, $482.8 million, or 94 percent of the budget, were spent on the care and custody program that supports the department's security and treatment services. A distant second among the department's three major budget categories was community services. It accounted for 5 percent of all expenditures. One percent of the budget was spent on management services.

Sixty-four percent of the Department of Correction's FY 02 expenditures were for personal services. As shown in Figure I-2, the distribution of staff by program closely parallels the department's expenditure pattern.

The study, approved by the program review committee in February 2003, was prompted by concerns raised by some of the department's employees, their union representatives, and the Senate president pro tempore about the adequacy of the staff resources devoted to maintaining a secure environment within the state's correctional facilities.

Many factors were considered in analyzing whether the staffing levels at the state's prisons have led to the creation of a safe or unsafe environment. Among those factors were the:

· number of inmates being supervised;

· operating philosophy of the department's administration;

· characteristics of the department's 18 correctional facilities;

· department's process for assigning staff; and

· number of staff responsible for security.

Data on the number and nature of incidents involving inmates were used as the primary measure of the safety of the state's prisons. The reported incidents were separated into categories including inmate on staff assaults, inmate on inmate assaults, and disciplinary reports. Annual data were collected as well as data covering a few specific points in time. The incidents were examined in the aggregate and by facility.

Data on staff, inmates, and facilities were obtained from the department and analyzed in terms of their relationship to reported incidents.

Written materials provided by the department and interviews of its staff were used to obtain information on the operating philosophy of the Department of Correction and the process followed in determining the number of staff to be assigned to a given facility.

The perspective of correction officers was obtained through interviews with their union representatives and an opinion survey mailed directly to 500 officers selected at random.

Finally, the committee held a public hearing on the subject on September 10, 2003.

It is the policy of the Legislative Program Review and Investigations Committee to provide the agencies subject to a study with an opportunity to review and comment on the recommendations prior to publication of the final report. Appendix E contains the response from the Department of Correction.

Table II-1 shows the inmate population at each of the 18 facilities run by the Department of Correction as of March 2003. The numbers include both sentenced and unsentenced individuals. Excluding the 500 inmates being housed in Virginia, the total inmate population in March 2003 was 18,385.

As stated in the Department of Correction's FY 04 budget request, the department operates under the philosophy that "correctional facilities serve as a means to punish and deter criminal behavior." This is an outgrowth of a mission statement adopted in FY 95 that directs the department to:

· protect the public;

· protect staff; and

· ensure a secure, safe, and humane environment for offenders in a climate that promotes professionalism, respect, integrity, dignity, and excellence.

The department's 18 facilities vary in the elements commonly thought to relate to prison staffing. These elements include the inmate population number, design, population density (inmate population divided by permanent beds), status of inmates (unsentenced vs. sentenced), and security classification level. Table II-2 contains data on four of these elements. (Inmate population numbers are provided in Table II-1 above.)

Facility design. As shown in Table II-2, seven facilities house inmates in cells, five use dormitories, and six employ a combination of the two. A few older facilities such as York-Niantic and Gates also provide cottage-style housing.

Within a facility, there can be multiple individual housing units. Some are reserved for inmates who present disciplinary problems, while others are for those requiring protection from the general inmate population or with special needs such as medical conditions.

Inmate status. Seven of the department's 18 facilities house both sentenced and unsentenced inmates. In only three facilities were more than one percent of the inmates classified as unsentenced: Bridgeport, 2.3 percent; Hartford, 7.4 percent; and New Haven, 3.2 percent. These facilities were designed to include pretrial inmates.

Density. A facility's density was calculated by dividing the average daily number of inmates reported for March 2003 by the number of permanent beds reported for the same point in time. (It should be noted both numbers fluctuate within a small range on an almost daily basis.) Using this measure, six facilities were overcapacity and 12 were under capacity.¹

Security level. The security level of each facility is designated by the Department of Correction. The department uses five security classifications:

· Level 5 - maximum security;

· Level 4 - high security;

· Level 3 - medium security;

· Level 2 - minimum security; and

· Level 1 - eligible for community release.

Most facilities contain inmates at two or more classification levels with their overall ratings based on the highest classification of inmates housed.

Figure II-1 shows the inmate population by security level for the department as a whole. The data are from June 30 of each year from 1998 through 2003.

Level. As shown in Figure III-1, increases in the number of correction officers over the last seven years have not kept pace with increases in the inmate population. The inmate population has grown at an average annual rate of about 4 percent compared to an average of about 1 percent for custody staff.

The differential growth rates have resulted in the ratio of inmates to correction officers increasing from 4.1 to 1 in 1997 to 4.7 to 1 in 2003. Figure III-2 compares Connecticut's ratio to the other New England states and the national average in 2000, the most recent year for which comparative data are available.

The figure shows that in 2000, Connecticut had a lower inmate to staff ratio than the national average, but a higher rate than all of the other New England states except New Hampshire.

Allocation. Table III-1 shows the number of staff at each of the department's 18 facilities as of March 2003 and the percentage responsible for custody, programs, and other duties. Nearly 70 percent of the facility staff were classified as custodial, 14 percent as program related, and 17 percent as "other". ²

However, the percentage of custody staff is misleading because the security function is an around-the-clock (24/7) responsibility, while the other two categories generally are not. Using only data from the first shift -- the most likely period when all three classifications of staff are working simultaneously -- the proportion of custody staff falls to 52 percent. Similarly, the inmate to custody staff ratio rises from 4.7 inmates per correction officer to 11.3 per officer when only first shift data are used in the calculation.

It is important to note, however, that many staff classified as "program" or "other" are trained correction officers available to respond in emergency situations. This was demonstrated for the committee in a videotape of a recent incident at one of the state's prisons. The tape, reviewed during a committee tour of DOC facilities, showed non-custody staff responding to an emergency right along with custody staff.

Span of control. In addition to the ratio of inmates to custody staff, the committee also examined the ratio of correction officers to lieutenants. In the Department of Correction, lieutenants function as supervisors and their numeric relationship to line staff -- referred to as the span of control -- is a key factor in the supervision of correction officers in both routine and emergency settings.

Figure III-3 shows an upward trend in the ratio of correction officers to lieutenants. Over the period shown, the number of correction officers per lieutenant rose from 13.6 to 1 in 1997, to 16.5 to 1 in 2003.

Inmate population. As expected, the committee found a strong relationship between the number of inmates at a facility and the number of custody staff. However, when facilities were compared to each other, there was a significant difference among them in terms of their inmate to custody staff ratios. (See Appendix A.)

This difference is shown in Figure III-4. The range is from 1.97 inmates for every custody officer at Northern (a maximum security celled facility) to 7.90 inmates per custody officer at Willard (a minimum security dormitory style facility). This suggests factors in addition to the size of the inmate population are influencing the distribution of custody staff among the department's facilities.

Facility characteristics. Among the factors analyzed to determine if they were related to differences in custody staff size among the 18 facilities were facility design, security designation, density, and sentencing status. The committee's analysis found each of these factors is linked to the number of inmates per custody officer at the department's facilities. The factors and their relationships with the inmate to custody staff ratio are summarized in Table III-2.

The committee also examined the relationship between combinations of factors and the inmate to custody staff ratio. The results from data pairing the facility design factor with the security level factor are shown in the matrix below.

The non-shaded area indicates the average number of inmates per custody staff for each design type (row averages) and security risk level (column averages) when the two factors are looked at separately. The shaded area shows the average inmate to custody staff ratio for facilities with the combined characteristics of the two factors identified by the row and column headings. For example, dormitory facilities with a Level 3 risk rating have 7.2 inmates per custody officer, whereas celled facilities designated risk Level 5 have an inmate to custody staff ratio of 2.0 to 1.

However, neither the combination of variables represented in the above matrix (facility design crossed with security level) nor any of the other factor combinations produced a sufficient number of observations to allow testing of the interactive effects for statistical significance.

Summary of findings on factors impacting staffing levels

¬ The number of inmates per custody staff is increasing.

¬ The number of custody staff per lieutenant (supervisor) is increasing.

¬ The number of inmates per custody staff varies significantly between facilities.

¬ Variation in the inmate to custody staff ratio is related to many factors including inmate population, facility design, facility security level, facility population density, and sentencing status of inmates at a facility.

¬ As the size of the inmate population increases at a facility, so does the number of custody staff.

¬ As facility design becomes more open, the number of inmates per custody staff increases.

¬ As the security level designation of a facility becomes higher, the number of inmates per custody staff decreases.

Process. Formulating the number of custody staff needed begins with identifying how many posts (i.e., work stations) at a specific facility must be staffed each shift. In addition to facility characteristics, elements emphasized in determining the number of posts are:

· safety;

· ability to respond to emergencies;

· analysis of incident data;

· feedback from staff; and

· observations.

Some posts require staffing seven days a week, others only five days. Posts are classified into three types:

· fixed post - must be staffed for an entire shift for safety and security reasons;

· pull post - may be vacant for up to three hours a shift, while a correction officer is reassigned; and

· shutdown post - may be vacant for part or all of a shift.

The Department of Correction reports it reviews staffing plans at least annually for each facility. The department indicated it monitors such things as inmate grievances, incident reports (formal and informal), and assaults, and will redo staffing plans at any time when warranted by this type of information.

Number of posts. In 2003, the department had approximately 2,200 defined posts spread over 18 facilities and three shifts. Figure III-5 shows the distribution of posts by type and shift. As expected, the first shift (days) has the most posts, followed in order by the second (evenings) and third (midnights) shifts. Of the 2,200 posts, 42 percent were defined as fixed, 36 percent pull, and 22 percent shutdown.

The committee found the allocation of posts by security level is generally consistent with the distribution of inmates. In 2003, approximately 52 percent of the posts were assigned to security level 4 facilities, which collectively house about 47 percent of the inmates. Similar relationships between the number of posts and the inmate population exist among the other security levels.

An analysis of post changes between 2001 and 2003 found the number of posts increased by 66 or 3.1 percent at a time when the inmate population increased by about 8.2 percent. However, economies of scale such as those connected to posts assigned to electronic control rooms and perimeter security means the number of posts would not necessarily have to track the inmate population.

The committee could not determine if either the current number of posts or the rate of change in the number of posts is correct. The committee lacked the information detail, time, and resources necessary to draw such definitive conclusions on these two issues.

Shift relief factor (SRF). Once the number of posts for a facility has been set, a shift relief factor is calculated to enable DOC to compute how many correction officers are actually needed to cover each post. The SRF is a formula based on staff absences for vacations, holidays, personal days, sick leave, workers' compensation, training days, military leave, disciplinary action, and other matters. Always greater than one, it currently is 2.2 for seven-day posts and 1.8 for five-day posts for each shift. Multiplying each post by its SRF determines the number of correction officers needed to cover each post.

Staff need. Table III-3 shows the projected number of custody staff needed based on the SRF, the actual number of staff reported by the Department of Correction on March 2003, and the percent difference between the two. As indicated in the table, using the shift relief factor the department was about 18 percent short of projected staff needs in March 2003. Covering the staff shortfall requires the department to use overtime and reassign staff from pull and shutdown posts.

Meeting staff need. Figure III-6 illustrates the overtime hours worked by correction officers and custody supervisors from 1998 through 2002. The unusual jump in 2000 is due in large part to the fact there were 27 pay periods in FY 00 as opposed to the normal 26 pay periods.

Since the average correction officer works about 1,350 hours per year -- net hours after deducting days off, holidays, vacation time, sick time, and personal leave time -- the 712 staff deficiency estimated using the SRF methodology represents nearly a million staff hours on an annual basis.

Based on FY 02 overtime data and assuming all of the estimated additional staff hours needed were worked, approximately 87 percent of the staff shortfall was covered by overtime. By default, the remaining 13 percent of the staff hours needed must have come from pull and shutdown posts.

Summary of findings related to determining staff levels

¬ The Department of Correction is about 700 staff short of meeting the optimum number of staff called for by the department's staffing plan.

¬ On a daily basis staff shortages are made up through the use of overtime (87 percent in 2002) and reassigning staff from one post to another.

Reporting and processing inmate misconduct. For purposes of this study the safety level at the state's correctional facilities is measured in general terms by the number of disciplinary reports generated by inmate misconduct. In specific terms, safety is measured by the number of inmate on staff and inmate on inmate assaults.

The Department of Correction Code of Penal Discipline defines acts of inmate misconduct, the process for handling allegations of misconduct, and the sanctions that can be imposed on inmates.

Misconduct. There are three levels of inmate misconduct under the code. Among a group of 28 actions falling into the most serious offense category are assaults on inmates and staff, the two variables selected by the committee as the key measures of the safety level at the department's facilities.

Processing procedure. Figure III-7 outlines the procedure for reporting and handling incidents occurring within a facility.

Punishment. Sanctions may include punitive segregation, forfeiture of good time, and loss of various privileges such as recreation, telephone, and visitation.

It should be noted that misconduct that also violates state criminal statutes is reported to the State Police for investigation, but the code of discipline is separate and unaffected by the actions of authorities external to the department.

Levels of inmate misconduct. As shown in Figure III-8, formal reports of inmate misconduct (disciplinary reports) have declined steadily since 1994. Figure III-8 also shows direct measures of safety within the department's facilities -- assaults on inmates and staff -- have fluctuated within a fairly narrow range since declining dramatically between 1994 and 1995.

Workers' compensation. With the limitations noted below, workers' compensation claim numbers can serve as a broad indicator of prison safety. Claims are not a precise measure for several reasons. First, not all claims are directly connected to prison incidents; some are caused by routine risks present in all work environments. Also, data provided to the program review committee by the Department of Administrative Services do not necessarily relate lost workdays and payments to the year in which each claim was initiated.

However, if it is assumed these limitations occur equally in each year, then any dramatic swing in the numbers could signal a real change in prison safety. With this in mind, the program review committee examined workers' compensation disability payment data, the only data available to the committee.

Figure III-9 shows a modest growth in claims paid over the last three years. In percentage terms, the increase was 1 percent from 2001 to 2002 and 2 percent from 2002 to 2003. On a claims paid per employee basis, the numbers were 0.89 in 2001, 0.89 in 2002, and 0.93 in 2003.

Summary of findings on staff safety

¬ The safety climate within the state's prison system as a whole has improved since the early 1990s.

¬ The rate of inmate incidents has declined despite an increase in the prison population.

¬ Workers' compensation claims resulting in disability payments have increased slightly in each of the last two fiscal years.

Analysis of annual incident data. The committee analyzed incident data beginning with 1994. As shown in Table III-4, the committee found negative relationships between measures of prison misconduct and the number of inmates and custody staff. While the decrease in the number of incidents associated with an increase in custody staff seems logical, a decrease in incidents associated with an increase in the inmate population seems counterintuitive.

As one possible explanation for the latter, a certain level of incidents may be caused by a small core of inmates prone to violence. Over a period of time, opportunities for these inmates to create problems are diminished as they are put in more secure settings, segregated from the general inmate population, and subjected to behavior modification techniques such as denial of various privileges.

Also contributing to this negative relationship may be a change in the type of inmate being sentenced to prison. The decrease in the number of individuals committing and being sentenced for violent crime means a larger number of new inmates are less violent than those that preceded them. Thus, while the overall number of inmates has risen, the percentage of inmates with a tendency to commit assaults may have declined, resulting in fewer incidents per inmate.

Another possible explanation for the increase in inmate population and decrease in incidents of inmate misconduct is the effect of policy changes instituted by the Department of Correction in the mid-1990s. In discussions with committee staff, department officials indicated these changes reduced inmate movement and greatly reduced opportunities for inmates to engage in assaults and other types of misconduct.

Analysis of facility-based data. In addition to analyzing inmate misconduct over a period of years, the committee also examined differences in the number of misconduct incidents among facilities. The intent was to assess the contribution of selected factors to prison safety. The data used in this analysis covered the 12-month period beginning February 1, 2001.

Table III-5 shows, by facility, the incidents per 1,000 inmates for the three variables selected as proxies for facility safety -- disciplinary reports, inmate assaults on staff, and inmate assaults on inmates. As with most of the previous comparisons, the variation among facilities is large.

For example, disciplinary reports ranged from a low of 441 per 1,000 inmates at Webster to a high of 4,323 per 1,000 inmates at Northern. Assaults on staff per 1,000 inmates ranged from a low of 0 at five different facilities to a high of 111 at Northern. Northern also led in assaults on inmates with 277 per 1,000 inmates compared to 9 per 1,000 at Cheshire.

Further examination of facility data found security level and density were the only factors of the six analyzed by the committee³ significant in differentiating facilities in terms of incidents of inmate misconduct. (See Appendix B.)

In the case of the security level factor, this means the higher the risk rating of inmates assigned to a facility the greater the number of incidents per 1,000 inmates. Of course, this is the expected result and indicates the importance of the security level rating system.

With respect to density, there is a significant negative relationship between the percent of capacity at which a facility is operating and the number of incidents per 1,000 inmates. In other words, the higher the density level, the lower the number of incidents. This is probably driven by the fact low-density facilities tend to have higher security level ratings. (See Table 2.)

Of note among the other factors analyzed was the positive relationship between custody staff numbers and incidents. Although the relationship was not particularly strong (see Appendix B), it supports the appropriateness of the relatively low inmate to staff ratios at facilities with high-risk ratings and conversely the relatively higher ratios at low-risk rated facilities (see matrix on page 10).

Summary of findings on factors associated with prison safety

¬ There is significant variation among the state's prison facilities with respect to measures of safety.

¬ In general, the higher the security level rating of a facility, the greater the rate of inmate incidents.

¬ In general, the higher the density level of a facility, the lower the rate of inmate incidents.

¬ In terms of safety, the population density of a facility is not as important as the risk level of the inmates it houses.

¬ Taken alone, facility design did not appear to be as important a factor in the rate of incidents as the facility's security level.

The committee conducted an opinion survey of 500 randomly selected correction officers. The 13-item questionnaire was mailed directly to the home address of each of the selected officers along with a stamped return envelope. (See Appendix C.) Sixteen surveys were returned as undeliverable, while 150 valid responses were received (31 percent of the total possible returns). Although the distribution of respondents by facility closely matched the distribution of correction officers, the low response rate means the results must be viewed with caution.

The tabulated responses to the nine opinion questions in the survey are shown in Appendix D, which displays the results grouped by the security level of the facility where the respondents were assigned.

Highlights of the responses

¬ Approximately 55 percent of the respondents indicated facility leaders have the biggest impact on prison safety, 19 percent believed it was the governor's office, 12 percent named the legislature, and only 9 percent cited the central office of the Department of Correction.

¬ Most respondents gave a neutral or negative job rating to the performance of everyone -- facility leaders, department central office staff, governor's office, and legislature -- except correction officers, who were viewed positively by 80 percent of the respondents.

¬ Nearly 70 percent of the respondents indicated they did not believe the Department of Correction's incident data accurately reflected the safety of facilities. About 76 percent of the Level 2 and 3 respondents expressed this view, compared to about 63 percent of those working in Level 4 and 5 facilities.

¬ Almost 80 percent of all respondents did not believe the Department of Correction was doing a good job in identifying the number of staff needed to assure a safe environment. The number holding this view was lowest among Level 5 workers, but still high at 68 percent.

¬ Nearly 50 percent of the respondents indicated their facility was safer now than in the mid-90s, but only 37 percent found the current level of safety acceptable. In both instances, respondents working in Level 3 facilities were more negative than those in Levels 2, 4, and 5.

¬ Almost 55 percent of the respondents indicated more staff was needed to improve safety. The only other option scoring more than 10 percent was "new department policies," which was favored by just under 20 percent.

Based on the analysis and findings outlined in Chapter III, the program review committee reached the following conclusions.

¬ The Department of Correction is about 700 correction officers short of the number needed to fully staff the department's custody staffing plan. The shortage is covered almost exclusively by the use of overtime.

¬ There is no objective method for setting an overall custody staff level or inmate to custody staff ratio due to facility variation, making doing it by statute inadvisable.

¬ There is significant variation among the Department of Correction's facilities in terms of the number of inmates per custody officer and measures of safety.

¬ The Department of Correction's procedures for determining staffing needs are consistent with nationally recognized standards.

¬ Correction officers are generally distrustful of the Department of Correction's incident data and the ability of the department to determine the number of custody staff needed to assure safety.

¬ Correction officers generally hold the belief prison safety is better now than in the mid-90s, but not safe enough.

¬ There is no objective method for establishing an acceptable level of safety for either the entire department or individual facilities.

¬ There is inadequate data on the relationship between staff injuries as measured by workers' compensation claims and overtime.

In an effort to address issues associated with some of the conclusions cited, the program review committee offers the following recommendations.

1. An overall custody staff level or inmate to custody staff ratio should not be set in statute.

2. Changes in the number of custody staff at the Department of Correction should be based on changes in objective measures of prison safety including but not limited to disciplinary reports, inmate on staff assaults, inmate on inmate assaults, and the security risk level of the inmate population being supervised.

3. Beginning no later than one month after the close of the first quarter of the 2005 state fiscal year and continuing one month after the close of every quarter thereafter, the Department of Correction shall submit to the governor and the General Assembly's committees of cognizance quarterly reports on the number of: disciplinary reports; inmate on staff assaults; inmate on inmate assaults; and workers' compensation claims by custody staff.

4. If the number of disciplinary reports, inmate on staff assaults, inmate on inmate assaults, or workers' compensation claims by custody staff increase by more than 5 percent over the previous quarter or 5 percent over the same quarter of the previous year, the Department of Correction shall provide a written explanation for the increase and a general outline of the measures the department will undertake to deal with the increase.

5. The Department of Correction should establish a system for handling disciplinary reports similar to the systems used by law enforcement agencies. The system should include pre-numbered blank disciplinary forms or some other means of assuring all reports can be audited and missing report forms can be tracked to a specific facility and location within the facility.

6. Committees of the General Assembly receiving the Department of Correction quarterly safety status report may hold a hearing on the report.

7. The Department of Correction should do a cost benefit analysis on its use of overtime to meet staff shortages. The study should consider as a cost the emotional and physical impact of overtime on staff.

8. The Department of Correction should undertake a study of the relationship between workers' compensation claims and the use of overtime. At a minimum, the study should determine if the incident generating the claim originated while the claimant was working overtime or had worked overtime within 72 hours immediately preceding the incident responsible for the claim. The results of the study should be reported to the governor and the General Assembly no later than January 1, 2005.

Custody staff r = .78 p = .01*

Custody staff to inmate ratio r = .01 p = .94

Custody staff per 1,000 inmates r = -.24 p = .34

Custody staff r = -.32 p = .19

Custody staff to inmate ratio r = .64 p = .01*

Custody staff per 1,000 inmates r = .75 p = .01*

Custody staff F = 5.95 p = .01*

Custody staff to inmate ratio F = 11.43 p = .00*

Custody staff per 1,000 inmates F = 4.08 p = .04

Custody staff F = 2.52 p = .10

Custody staff to inmate ratio F = 11.79 p = .00*

Custody staff per 1,000 inmates F = 8.11 p = .04

Custody staff F = 5.89 p = .03*

Custody staff to inmate ratio F = 4.57 p = .05*

Custody staff per 1,000 inmates F = 0.89 p = .40

Disciplinary reports r = -.23 p = .35

Assaults on staff per 1,000 inmates r = -.26 p = .30

Assaults on inmates per 1,000 inmates r = -.26 p = .30

Disciplinary reports r = .25 p = .32

Assaults on staff per 1,000 inmates r = .19 p = .46

Assaults on inmates per 1,000 inmates r = .20 p = .43

Disciplinary reports r = -.76 p = .00*

Assaults on staff per 1,000 inmates r = -.73 p = .00*

Assaults on inmates per 1,000 inmates r = -.62 p = .00*

Disciplinary reports F = 3.63 p = .06

Assaults on staff per 1,000 inmates F = 1.37 p = .28

Assaults on inmates per 1,000 inmates F = 1.08 p = .36

Disciplinary reports F = 7.31 p = .00*

Assaults on staff per 1,000 inmates F = 4.31 p = .02*

Assaults on inmates per 1,000 inmates F = 4.03 p = .03*

Disciplinary reports F = 1.06 p = .32

Assaults on staff per 1,000 inmates F = 0.12 p = .73

Assaults on inmates per 1,000 inmates F = 0.39 p = .54

___ New department policies for handling and moving inmates

___ More programs to help inmates cope with their problems

___ More custodial staff

___ Better designed facilities

___ Reduction in the number of inmates in the facility where you work