Academic educational programming in prisons has existed since prisons were first developed in the United States (Gerber & Fritsch, 1995; Lawrence, 1994). Walnut Street Jail, the first facility to house felons during the latter part of the 1700's, included "...instruction in reading, writing, and math" (Welch, 1996, p. 173). Such programs have expanded to the present time, so that few prisons are lacking in some type of educational endeavor (Office of Correctional Education, Survey of state correctional educational systems: Analysis of data from 1992 field test, 1996). According to this survey, in which 42 state responses were used, "...almost all facilities offer Adult Basic Education and GED Preparation Courses" (p. 2).

According to the "Survey of State Prison Inmates, 1991," "...nearly half of all inmates have received academic education" (Bureau of Justice Statistics, 1993, p. 27). It has been reported that 400,000 state-held inmates "...have participated in some type of correctional education program" (Office of Correctional Education, Using correctional education data: Issues & strategies, 1997, p. 1).

Despite the widespread use of educational programming in prisons, some jurisdictions in recent years have cut funding and eliminated education programs (see Tewksbury & Taylor, 1996). The most recent illustration is the elimination of Pell Grants for inmates pursuing post-secondary education (Tewksbury & Taylor, 1996; Batiuk, Moke, & Rountree, 1997).

Educational programs--in the Oklahoma Department of Corrections (ODOC) and elsewhere--are justified on the basis of extreme inmate academic deficiencies. If not illiterate, many offenders function at several grade levels below where they are expected to be based on age or grade completed (Snarr, 1996; Clear & Cole, 1994; Egan, 1993). Specifically, sixty-six percent of inmates in ODOC have not completed high school and over 40% read at or below 8th grade level (Oklahoma Department of Corrections, Highest grade completed, URL http://www.doc.state.ok.us/docs/Lstgrade.htm; Reading level, URL http://www.doc.state.ok.us/docs/educate.htm). Among all states, only 34% of 1991 state inmates had obtained a high school diploma (Bureau of Justice Statistics, Survey of state prison inmates, 1991, 1993).

Significant proportions of inmates are handicapped, have learning disabilities, and are mentally retarded (Platt, Bohac & Barnes, 1993). Cognitive functions--such as problem-solving abilities and creative thinking--are limited (Yeonopolus, 1994). Numerous inmates are dropouts (Egan, 1993; Menninger, 1968), while some others are young enough so that they have not completed their education.

In addition to the benefits derived in the area of offender management (see Marquart et al., 1994) and normalization of the prison environment intended to address "prisonization" (Harer, 1994), educational programming is intended to achieve long-term consequences in addressing the visible and real "needs" of inmates. These include productive living and pro-social behavior while incarcerated, and upon release, employability (from both vocational and academic experiences), and ultimately reduction in recidivism (Jones, 1993; Platt et al., 1993).

That education ultimately can or inevitably does affect the recidivism rate (i.e., rate of return to prison after release), while widely believed, is still open to question. Research findings, while often encouraging in the demonstration of a link between education and lower recidivism rates (see Egan, 1993; Snarr, 1996; Gerber & Fritsch, 1995; Harer, 1994), have also evoked numerous questions, skepticism, and if nothing else, caution in the interpretation of the findings (see Batiuk, et al., 1997; Office of Correctional Education, Using correctional education data: Issues & strategies, 1997; Tewksbury & Taylor, 1996; Clear & Cole, 1994).

This section will examine (1) research on the effects of educational programming, (2) recidivism studies, and (4) ODOC educational programming.

Gerber and Fritsch (1995) provide a recent and thorough review of the literature. According to their review, most studies have shown a correlation between pre-college education and lower recidivism rates, post-release employment, and post-release educational participation. In contrast, fewer studies have shown "...no correlation between prison education and recidivism" (1995, p. 126). Glaser’s 1961 study found no relationship between these variables (in Clear & Cole, 1994). It was postulated that with the acquisition of the GED came unrealistic expectations of post-release life and employment. In addition to the numerous methodological questions--lack of random assignment to groups, absence of control groups, etc.--raised about these studies which have shown positive effects of prison education (Batiuk, et al., 1997; Gerber & Fritsch, 1995), other issues of a theoretical nature are articulated. Specifically, it is asserted that "...there is a lack of clear specification about the process by which correctional education and recidivism are linked..." (Batiuk, et al., 1997, p. 169).

Batiuk and associates (1997), in an effort to address this problem, studied the relationship between post-secondary education and recidivism. They concluded that "...college education [in prison] does reduce the likelihood of recidivism, but this effect works principally through post release employment" (1997, p. 175). While focusing on recidivism, one Oklahoma study by Langenbach, North, Aagaard, and Chown (1990) found a "...significant negative relationship between participation [in instruction by television] and recidivism" (as cited in Lawrence, 1994, p. 57).

Despite the methodological concerns identified above, Gerber and Fritsch (1995) conclude that prison education programs have positive outcomes. They state that the programs with the greatest degree of success include those (1) which have the more extensive program, (2) separate the inmates from the rest of the prisoners, (3) provide follow-up after release, and (4) place only inmates in the program meeting the program criteria (1995, pp. 135, 136).

The recidivism literature will be briefly examined. While recidivism studies have focused on numerous explanatory variables, typically education-related variables are included. In the study of 1979 state prison admissions, level of education is included in the analysis, although differences between recidivists and first time offenders do not appear significant (although no test of differences was presented) (Bureau of Justice Statistics, Examining recidivism, 1985). A 1987 study of 1978 parolees from 22 states found recidivism ranked in order of highest to lowest: (1) less than high school education, (2) high school graduate, and (3) some college (Bureau of Justice Statistics, Recidivism of young parolees, 1987).

In another study of prison releasees from 11 states in 1983, it was found that "prisoners who had graduated from high school or had some college education had somewhat lower rates of rearrest, reconviction, and reincarceration than those who failed to complete high school" (Bureau of Justice Statistics, Recidivism of prisoners released in 1983, 1989, p. 5). Education, however, was not found to be a meaningful predictor of risk of rearrest, although other variables were, including age at release and prior incarceration. Significantly, the "Salient Factor Score," used to determine parole risks for federal inmates since the early 1970's, does not incorporate any reference to education, although it includes age and prior incarcerations (see Hoffman, 1994).

Several recidivism studies published during the 1990's also included education. In a study of Driving While Impaired (DWI), 1986-1987 offenders exposed to treatment which included group therapy, Kernodle and associates found a relationship between educational level and lower recidivism (1995). Lower recidivism rates for those with more education have been found among parolees (Sims & Jones, 1997; Morgan, 1995).

The Harer (1994) study of federal prison releasees during 1987 provides useful information regarding (1) educational characteristics and (2) participation in prison education programs. He found that the highest recidivism rates occurred among those with less than a high school education, and much lower recidivism rates among those with "some college" or "college graduate." And, it was found that offenders had lower recidivism rates with the more educational courses in which they participated. Special benefits for prison programming were found for those with "...an eighth grade education or less at admission who participate in ABE and GED courses..." (P. 26).

The Gendreau, Little, and Goggin (1996) review of 131 recidivism studies adds a discussion of intellectual functioning and IQ, although most research tends to exclude any direct consideration of IQ. The Chown and Davis (1986) study of recidivism within ODOC included an examination of IQ. They found higher odds of recidivism for offenders with an IQ of 75 or below than offenders with an IQ above 75.

The ODOC has long had a commitment to prison education.

Today, the following academic educational programs are available in ODOC facilities: (1) Learning Disability Education, (2) Literacy programs, (3) Intellectually Challenged programs, (4) Adult Basic Education, (5) Chapter 1, (6) General Education Development (GED), (7) High School Diploma, (8) Post Secondary Education, and (9) College courses (Oklahoma Department of Corrections, Program description guide, 1995).

The GED program "...provides instruction in reading, math, language arts, social studies, spelling, and science for students functioning above the eighth grade level. Inmates passing the state GED exam are rewarded with a high school equivalency certificate" (Oklahoma Department of Corrections, Program description guide, 1995, no page number).

Within ODOC, 5,564 inmates took part in "academic" (i.e., non-vocational) programs during FY 1996 ("7,035 inmates participate in education programs," 1997). During this year, the passing rate for the GED test was 85% (4,729 offenders). The ODOC education program has a staff of 112 FTE and a budget of approximately $5 million dollars (including grants) (Oklahoma Department of Corrections, Oklahoma Department of Corrections FY 97 Budget, n.d./1997). Educational programming is available in all ODOC facilities (Oklahoma Department of Corrections, Program description guide, 1995).

The purpose of this research was to examine the effectiveness of educational programming within ODOC. The research question stated: Does the acquisition of education while incarcerated reduce the likelihood of recidivism among prisoners? The research was designed to specifically study those who obtained the GED while incarcerated within ODOC through the use of a recidivism measure.

Inmates who complete a GED while in ODOC custody would seem to recognize their educational deficiency (perhaps with overt "encouragement" from the courts or ODOC), be willing to work to make up that deficiency, and with the GED completion and the corresponding motivation, would be less likely than others to return to prison. Thus, it was hypothesized that those obtaining a GED while incarcerated would have lower recidivism rates than offenders released with a high school diploma gained prior to incarceration and those released with neither a high school diploma nor a GED.

The data set, acquired from the ODOC Research and Evaluation Unit, consisted of 24,916 offenders released from "prison jurisdiction" between January 1, 1990, and December 31,1994. Cases were then excluded if not Oklahoma jurisdictional inmates and if missing a release date.

Next, three mutually exclusive groups were formed. The first group refers to those who entered the correctional system with a high school diploma. The second group consists of those who received a GED while incarcerated. And the third group represents those who entered the correctional system without a high school diploma and who did not obtain a GED while incarcerated. Thus, the study is based on the resulting 23,879 cases.

This release time frame provided a minimum of 3 years after release for all but 519 of those released (representing 2.1% of the population), who were released between January 1, 1994 and July 1, 1994. Consequently, this provided us an adequate time frame after the offender’s release for measuring recidivism.

The research methodology makes use of comparison groups in a quasi-experimental design (see Babbie, 1995; Maxfield & Babbie, 1995). The research design is subject to the same limitations as any other research with this design, namely lack of random assignment to groups. While random assignment of offenders to the three groups is not possible, the population is being studied, so results will reflect the current population’s recidivism and characteristics.

The variables for the data analysis included:

1. Dates of birth, ODOC reception, ODOC release, and ODOC return (if applicable),

2. Demographic variables--including sex, race, and martial status at reception,

3. Offense-related information--including offense, sentence length, and county of origin,

Groups 1 and 2 were created directly from the variables of "high school diploma at ODOC reception" and "obtained GED during incarceration." Group 3 was determined by combining those who did not have a high school diploma at ODOC reception and did not

get a GED while incarcerated.

Recidivism serves as the dependent variable, primarily based on availability of the educational and correctional data. Recidivism is defined as the return to ODOC custody after release (via a "return date" in ODOC inmate records) during the research follow-up period—minimum of three years. This research does not distinguish between returns for technical violations and those from a new sentence. Many would argue that using only those who return to the Oklahoma Department of Corrections is not a accurate measure of recidivism, it does provide a measure of the recidivism for the state. Data on re-arrest or re-arrest and non-prosecution is not available from existing ODOC records.

Chi-square was used to measure the extent of association and statistical significance of relationships between independent variables and the dependent variable, Alpha=.05.

The findings section will be divided into a description of the population, which will be followed by a description of the 3 comparison groups. Then, recidivism analysis will be carried out.

There were 23,879 inmates released from ODOC custody between January 1, 1990 and December 31, 1994. Table 2 indicates that the percentages of the population released for the 5 years ranged from 17.9% to 23.6%, with a low of 4,273 in 1991 and a high of 5,625 in 1994. It is worthwhile to note that the numbers released each year increased with one exception--from 1990 to 1991. Relatively few inmates--8.2%--were released from maximum and medium security facilities. Almost equal percentages were released from minimum/community security levels and probation and parole (47.5% and 44.3%, respectively). Forty-three percent of offenders were discharged to the streets without any supervision. The remainder were nearly equally divided between probation--30.5%--and parole--25.6%. Over fifty percent of offenders cases originated in Oklahoma and Tulsa counties (28.6% and 23.1%, respectively). Comanche County accounts for 5.7% with the other 74 counties providing the remaining 42.5%. Approximately seventy-three percent had sentences of between 1 and 5.99 years. Nine percent had a sentence of less than a year, while 18.1% had a sentence longer than 6 years.

We created offense categories by combining offenses, the results of which are found in Table 3. Assault with a deadly weapon, shooting with intent to kill, murder first degree, murder second degree, manslaughter first degree, and manslaughter second degree were combined to create the "homicide/assault category." "Robbery" consists of robbery first degree, robbery second degree, robbery with a weapon, and robbery. "Sex offenses" includes sodomy, indecent exposure, rape, rape first degree, rape second degree, and lewd or indecent proposals/acts to child. Burglary first degree and burglary second degree are combined to form the "burglary" offense. "Larceny" includes grand larceny, receiving/possessing/concealing stolen property, and larceny of an auto, aircraft, or other motor vehicle. "DUI" consists of DUI-liquor or drugs/APCV (i.e., after a prior conviction). Illegal distribution of controlled substances and illegal possession of controlled substances are combined to create the offense category, "Illegal Poss/Dist CS."

The drug-related offenses combined with DUI represent approximately one-third of the cases. Larceny and burglary number almost 6,000, which is nearly 25% of the population. The remaining offense groupings stand individually at about 5% or less.

Females account for 12.7% of the population. Whites represent 59%, Blacks--30.7%, Indian--7.1%, and Hispanic--2.9%. Of those reporting marital status, forty-two percent were identified as single at ODOC reception, while 29.5% were married and 15.3% were divorced. Over half—54%—were age 29 or younger at ODOC reception, while 29.8% were between 30 and 39. The remainder--15.1% were 40 years of age or older at reception.

Incarceration-related data are presented next. Over three-fourths of the offenders had no previous incarcerations. For those with a prior incarceration, 17.1% had one previous incarceration and 6.7% had two or more. A majority of offenders--57.4%--had no misconducts during incarceration. Almost 16% had one misconduct, with nearly 16% having 2-4 misconducts. Offenders with 5 or more misconducts totaled 11.4%. Most--82.1%--did not file grievances while in custody. Nearly equal percentages file only one or two or more grievances(9.3% and 8.5%, respectively).

The recidivism rate for the population was 32.8% (Table 4). Thus, 67.2% of the offenders released during 1990-1994 did not recidivate during a follow-up period of at least 3

years (with a small number of exceptions for those released during the last 6 months of 1994), and some were followed for up to 7 years. For example, the 1994 releasees were followed for half of 1997.

The two comparison groups--High School Diploma and No High School Diploma--are statistically less likely to be male than female (p<.001). Almost 90% of the GED group are male. Table 5 provides the data by sex and comparison group.

Table 6 presents the breakdown by age. The GED group is noticeably younger than Groups 1 and 3. Almost 75% of the GED group is under the age of 29. These differences are significant at the p<.0000 level.

Racial categories are presented in Table 7. GED completers are much more likely to be white. These racial differences are significant at the p<.0000 level.

There are significant differences between the comparison groups by martial status (p<.0000). GED completers are more likely to be single (57.2%) than married or divorced. This finding is to be understood along with the age variable. Younger persons would be expected to more likely be single than divorced or married.

Offense related findings will now be presented. Examining county of origin, GED completers are more likely to be from the 74 counties (labeled as "Other") (accounting for 47.7%) and less likely to be from Comanche County. These comparison group differences are significant at the p<.0000 level.

Table 8 provides the analysis by offense. The GED group is more likely to be incarcerated for burglary and larceny (31.1% combined) and sex offenses, and less likely to be doing time for DUI (5.3% compared to 14.3% for Group 1 and 11.6% for Group 3) than the other groups. These differences are significant at the p<.0000 level.

Sentence length is broken down by comparison group in Table 11, with differences significant at the p<.0000 level. GED completers were more likely to have a sentence of less than a year and from 3-5.99 years and less likely to have a sentence of 1-2.99 years and 6-9.99 years than the comparison groups.

At release, the GED group was more likely to be located at a minimum security facility (37.7% compared to 25.1% and 26.4%), and much less likely to be in probation and parole (31.7% compared to 47.4% and 42.2%). These differences were significant at the p<.0000 level.

GED completers were more likely to be released to probation (39% compared to 29.9% and 29.6%) and less likely to be paroled (20% compared to 26.8% and 25%). These differences were significant at the p<.0000 level.

The GED group has fewer previous incarcerations. The data show that over 8 of 10 have no previous incarcerations (82.8% compared to 77.1% and 73.6%). These differences were significant at the p<.0000 level.

GED completers were more likely to have received misconducts. For example, 16% of the GED group had 5 or more misconducts, compared to 9% for Group 1 and 13% for Group 3. Likewise, the GED group was much less likely to have no misconducts (46.9% compared to 61.4% and 55.1%). These differences were significant at the p<.0000 level. Similarly, the GED completers were less likely to have filed no grievances (77.7% compared to 84.4% and 80.1%). Thus, they were more likely to have one filed (11.3% compared to 8.4% and 10%) and two or more filed (11.1% compared to 7.1% and 9.9%). These differences were significant at the p<.0000 level.

Significant differences between the GED group and the High School Diploma Group and the No High School Diploma were found with every comparison. The GED group consist of offenders who are more likely to be males than females, younger than older, white than member of another race/minority, single than married or divorced. The GED completers were more likely to have their case originating in one of the 74 counties and less likely to be from Comanche County. GED completers were more likely to be involved in burglary, larceny, and sex offenses, and less like to be convicted of DUI. They were more likely to be serving less than a year or between 3 and 5.99 years. This group was more likely at a minimum security facility when released and released to probation. They are less likely to have previous incarcerations, yet when incarcerated have more misconducts and file more grievances.

Table 9 provides the recidivism data broken down by comparison groups. Those completing a GED while in ODOC custody fared better than the other two groups. The recidivism rate for the GED completers was 29%, while those with a high school diploma at reception had a rate of 34% and those with neither a GED nor a high school diploma had a rate of 31.5%. The differences are statistically significant (p<.000).

Two findings from this research are especially significant: (1) Oklahoma’s low recidivism rate found in other studies has been replicated, and (2) offenders obtaining a GED while incarcerated were less likely to recidivate over the time frame of this study than those entering ODOC with a high school diploma and those who were released with neither a high school diploma nor a GED. Limitations of the actual number of offenders re-arrested or arrested in a jurisdiction other than the ODOC are not available, and may cause the findings here to underreport recidivism.

Noteworthy characteristics of GED completers include younger age at ODOC reception, males, white, single, those with shorter sentences, from the smaller, less urbanized counties, minimum security at release, and those less likely to have been previously incarcerated. Those without these characteristics--older age at reception, females, minority, married or divorced, those with longer sentences, those from the 3 largest counties, non-minimum security levels, and repeat offenders would seem to be appropriate targets for increased participation in GED programming.

Although there are some limitations with the kind of data used in this analysis, the findings are highly instructive. Perhaps the most sobering finding is that those entering ODOC with a high school diploma have the highest recidivism rates. Clearly, having a diploma did not insulate these persons from arrest, conviction, and incarceration. Whether additional education or vocational training gained while in prison might help after release remains to be determined.

Intervening variables in the relationship between education and recidivism might well include the positive effects of the "normalization" of the prison environment, educational intervention in the earlier stages of one’s criminal career, improvement in self-esteem, greater motivation, and greater employability, etc. While the mechanisms which might operate here have yet to be fully elucidated, it is clear that there are advantages to GED completion during incarceration.

With the current frenzy for "Tough on Criminals," and politicians and news media portraying offenders in a negative light, this study indicates the need for sound educational programs within the correctional field. While not wanting to "coddle" inmates, it is important to provide offenders with the necessary tools to return to society with a least some chance of becoming productive citizens.
 
 

TABLE 1

		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	High School Diploma	13051	52.4	54.7	54.7
	GED Completion	1742	7.0	7.3	61.9
	No High School Diploma, No GED	9086	36.5	38.1	100.0
	Total	23879	95.8	100.0
Missing	System Missing	1037	4.2
	Total	1037	4.2
Total		24916	100.0

 

TABLE 2

		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	1990	4323	18.1	18.1	18.1
	1991	4273	17.9	17.9	36.0
	1992	4824	20.2	20.2	56.2
	1993	4825	20.2	20.2	76.4
	1994	5625	23.6	23.6	100.0
	Total	23870	100.0	100.0
Missing	System Missing	9	.0
	Total	9	.0
Total		23879	100.0

 

TABLE 3

Selected Offenses of ODOC Releasees, 1990-1994
 

		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	Other Offenses	6943	29.1	29.1	29.1
	Homicide/Assault	1139	4.8	4.8	33.9
	Robbery	1202	5.0	5.0	38.9
	Sex Offenses	1075	4.5	4.5	43.4
	Burglary	3308	13.9	13.9	57.3
	Larceny	2284	9.6	9.6	66.9
	DUI	3012	12.6	12.6	79.5
	Illegal Poss/Dist CDS	4896	20.5	20.5	100.0
	Total	23859	99.9	100.0
Missing	System Missing	20	.1
	Total	20	.1
Total		23879	100.0

 

TABLE 4

Recidivism		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	No	16037	67.2	67.2	67.2
	Yes	7842	32.8	32.8	100.0
	Total	23879	100.0	100.0
Total		23879	100.0

 

TABLE 5

			Comparison Group			Total
			High School Diploma	GED Completion	No High School Diploma, No GED
Sex	Male	Count	11299	1562	7903	20764
		% within Comparison Group	86.6%	89.7%	87.1%	87.0%
	Female	Count	1752	180	1174	3106
		% within Comparison Group	13.4%	10.3%	12.9%	13.0%
Total		Count	13051	1742	9077	23870
		% within Comparison Group	100.0%	100.0%	100.0%	100.0%

Pearson Chi-square: 13.055, DF=2, p<.001.
 
 

TABLE 6

			Comparison Group			Total
			High School Diploma	GED Completion	No High School Diploma, No GED
	13 to 19	Count	891	383	950	2224
Age at Reception		% within Comparison Group	6.9%	22.1%	10.6%	9.4%
	20 to 29	Count	5586	880	3954	10420
		% within Comparison Group	43.3%	50.8%	44.0%	44.1%
	30 to 39	Count	4288	346	2630	7264
		% within Comparison Group	33.2%	20.0%	29.3%	30.8%
	40 to 49	Count	1546	105	986	2637
		% within Comparison Group	12.0%	6.1%	11.0%	11.2%
	50 or over	Count	590	18	459	1067
		% within Comparison Group	4.6%	1.0%	5.1%	4.5%
Total		Count	12901	1732	8979	23612
		% within Comparison Group	100.0%	100.0%	100.0%	100.0%

Pearson Chi-square: 614.559, DF=8, p<.000.
 
 

TABLE 7

			Comparison Group			Total
			High School Diploma	GED Completion	No High School Diploma, No GED
Race	White	Count	7367	1280	5435	14082
		% within Comparison Group	56.6%	73.9%	60.0%	59.1%
	Indian	Count	850	125	707	1682
		% within Comparison Group	6.5%	7.2%	7.8%	7.1%
	Hispanic	Count	309	50	315	674
		% within Comparison Group	2.4%	2.9%	3.5%	2.8%
	Black	Count	4490	278	2603	7371
		% within Comparison Group	34.5%	16.0%	28.7%	31.0%
Total		Count	13016	1733	9060	23809
		% within Comparison Group	100.0%	100.0%	100.0%	100.0%

Pearson Chi-square: 305.776, DF=6, p<.000.
 
 
 
 

TABLE 8

			Comparison Group			Total
			High School Diploma	GED Completion	No High School Diploma, No GED
	Less than 1 year	Count	1004	319	739	2062
Length Of Sentence		% within Comparison Group	7.7%	18.3%	8.1%	8.6%
	1-2.99 years	Count	4808	414	3154	8376
		% within Comparison Group	36.8%	23.8%	34.7%	35.1%
	3-5.99 years	Count	5123	623	3408	9154
		% within Comparison Group	39.3%	35.8%	37.5%	38.3%
	6-9.99 years	Count	850	153	545	1548
		% within Comparison Group	6.5%	8.8%	6.0%	6.5%
	10 or more years	Count	1266	233	1231	2730
		% within Comparison Group	9.7%	13.4%	13.6%	11.4%
Total		Count	13051	1742	9077	23870
		% within Comparison Group	100.0%	100.0%	100.0%	100.0%

Pearson Chi-square: 380.930, DF=8, p<.000.
 
 

TABLE 9

			Comparison Group			Total
			High School Diploma	GED Completion	No High School Diploma, No GED
Recidivism	No	Count	8578	1235	6224	16037
		% within Comparison Group	65.7%	70.9%	68.5%	67.2%
	Yes	Count	4473	507	2862	7842
		% within Comparison Group	34.3%	29.1%	31.5%	32.8%
Total		Count	13051	1742	9086	23879
		% within Comparison Group	100.0%	100.0%	100.0%	100.0%

Pearson Chi-square: 30.585, DF=2, p<.000.
 
 

7,035 inmates participate in education programs. (1997, August). Inside Corrections, p. 6.

Babbie, E. (1995). The practice of social research. (7th. ed). Belmont, CA: Wadsworth.

Batiuk, M.E., Moke, P., & Rountree, P. W. (1997). Crime and rehabilitation: Correctional education as an agent of change--A research note. Justice Quarterly, 14(1), 167-180.

Bureau of Justice Statistics. (1987, May). Examining recidivism. (BJS Publication No. NCJ-96501). Washington DC: National Criminal Justice Reference Service.

Bureau of Justice Statistics. (1989, April). Recidivism of prisoners released in 1983. (BJS Publication No. NCJ-116261). Washington DC: National Criminal Justice Reference Service.

Bureau of Justice Statistics. (1987, May). Recidivism of young parolees. (BJS Publication No. NCJ-104916). Washington DC: National Criminal Justice Reference Service.

Bureau of Justice Statistics. (1993, March). Survey of state prison inmates, 1991. (BJS Publication No. NCJ-136949). Washington DC: National Criminal Justice Reference Service.

Chown, B & Davis, S. (1986). Recidivism among offenders incarcerated by the Oklahoma Department of Corrections: A survival data analysis of offenders released January 1982 through July 1986 and case-control study of offenders received in 1985 and 1986. Oklahoma City, OK: Oklahoma Department of Corrections Planning and Research.

Clear, T. R. & Cole, G.F. (1994). American corrections. (3rd ed.). Belmont, CA: Wadsworth.

Egan, S. (1993, February). Correctional education: An important tool in corrections. Corrections Today, 55(1), 10, 30.

Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34(4), 575-607.

Gerber, J. & Fritsch, E.J. (1995). Adult academic and vocational correctional education programs: A review of recent research. Journal of Offender Rehabilitation, 22(1/2), 119-142.

Harer, M.D. (1994, January 25). Recidivism among federal prison releasees in 1987: A preliminary report. (Federal Bureau of Prisons, Office of Research and Evaluation).

Hoffman, P. B. (1994). Twenty years of operational use of a risk prediction instrument: The United States Parole commission’s Salient Factor Score. Journal of Criminal Justice, 22(6), 477-494.

Jones, W. (1993, February). Building inmates’ skills through training, industry and education. Corrections Today, 55(1), 74, 76.

Kernodle, J.R., Jr., Joyce, C.C., & Farmer, R.J. (1995). Changing the behavior of DWI First Offenders. Journal of Offender Rehabilitation, 22(3/4), 113-128.

Lawrence, D. (1994, August). Inmate students: Where do they fit in? Journal of the Oklahoma Criminal Justice Research Consortium, 1, 43-51.

Marquart, J.W., Cuvelier, S.J., Burton, V.S., Jr., Adams, K., Gerber, J., Longmire, D., Flanagan, T.J., Bennett, K., & Fritsch, E. (1994). A limited capacity to treat: Examining the effects of prison population control strategies on prison education programs. Crime & Delinquency, 40(4), 516-531.

Maxfield, M.G. & Babbie, E. (1995). Research methods for criminal justice and criminology. Belmont: Wadsworth.

Menninger, K. (1968). The crime of punishment. New York: Viking.

Morgan, K.D. (1995). Variables associated with successful probation completion. Journal of Offender Rehabilitation, 22(3/4), 141-153.

Office of Correctional Education. (1997, January). Using correctional education data: Issues & strategies. Washington, DC: U.S. Department of Education.

Office of Correctional Education. (1996, January 30). Survey of state correctional educational systems: Analysis of data from 1992 field test. Washington, DC: U.S. Department of Education.

Oklahoma Department of Corrections. (n.d./1997). Highest grade completed, [WWW document]. URL http://www.doc.state.ok.us/docs/Lstgrade.htm.

Oklahoma Department of Corrections. (n.d../1997). Oklahoma Department of Corrections FY 97 Budget. [WWW document]. URL http://www.doc.state.ok.us/Spreadsh/WPSUM.htm.

Oklahoma Department of Corrections. (n.d./1997). Reading level. [WWW document]. URL http://www.doc.state.ok.us/docs/educate.htm.

Oklahoma Department of Corrections. (n.d./1997). Recidivism. [WWW document]. URL http://www.doc.state.ok.us/ docs/recidivi.htm.

Oklahoma Department of Corrections. (1995). Program description guide. Oklahoma City, OK: author.

Platt, J.S., Bohac, P.D., & Barnes W.G.W. (1993, February). The year 2000: Changing to meet challenges in correctional education. Corrections Today, 55(1), 64, 66-69, 86.

Sims, B. & Jones, M. (1997). Predicting success or failure on probation: Factors associated with felony probation outcomes. Crime & Delinquency, 43(3), 314-327.

Snarr, R.W. (1996). Introduction to corrections. (3rd ed.). Dubuque, IA: Brown & Benchmark.

Tewksbury, R., & Taylor, J.M. (1996, September). The consequences of eliminating Pell Grant eligibility for students in post-secondary correctional education programs. Federal Probation, 60(3), 60-63.

Welch, M. (1996). Corrections: A critical approach. New York: McGraw-Hill.

Yeonopolus, J.M. (1994, August). Innovative Programs for Inmates. Journal of the Oklahoma Criminal Justice Research Consortium, 1, 53-61.