GED holders in prison read better than those in the household population: why?

Harlow, Caroline Wolf ; Jenkins, H. David ; Steurer, Stephen 等

In the last few decades the U.S. Department of Education (DOE) has conducted two nationally representative surveys assessing adult literacy in the United States. In 1992 the U.S. Education Department's statistical agency, the National Center for Education Statistics, sponsored the first survey, the National Adult Literacy Survey (NALS). Respondents included nationally representative samples of persons in the general population and in prisons, both federal and state. Findings focusing on prisoners were published in 1994 in Literacy Behind Prison Walls. Comparing household and prison respondents, a key finding was that inmates scored substantially lower than the household population on three separate literacy skills-reading, understanding documents and general quantitative facility.

A second survey, the National Assessment of Adult Literacy (NAAL) conducted in 2003, had similar results although prisoners as a whole scored better in 2003 than they had in 1992 in their reading and quantitative skills. However, in 2003, prisoners with a GED as their final education experience scored better in reading skills than persons in the general population with the equivalent education.

To discover why, the authors decided to explore three avenues--a review of the literature, a questionnaire of leaders in correctional education and a more specific analysis of the NAAL and GED databases on prisoners.

Literature Review

The U.S. prison population traditionally differs significantly from the general or household population in a variety of measures, including age, race, gender and educational attainment. These differences--more youthful, disproportionately minority, overwhelmingly male, higher levels of learning disabilities, and fewer years of education completed--are generally associated with lower educational attainment and run counter to the research question at hand--how did the prison population with a GED score significantly higher on their reading skills than the household population with a GED?

The literature on prison education has largely been descriptive and more recently focused on "what works" or the impact of prison education programs on the post release success of participants. The studies stand in contrast to Martinson's widely known and quoted finding of 1974 that "nothing works." Recidivism has been a primary focus of this research and, as a result, there have been a number of studies of individual and multi-state programs. Most recently, meta-analyses have brought together a number of small studies to produce findings of wider applicability. These analyses produce large sample sizes from a number of individual research studies with acceptable research methodologies that include control or comparison groups. Examples of studies with large samples and multistate designs or meta-analyses include: Aos, Miller and Drake; Harer; Saylor and Gaes; Wilson, Gallagher and MacKenzie; Vaher and Visher; and Steurer, Smith and Tracy.

Early reports of inmates' academic performance suggested that offenders may be as or more competent than their community counterparts. MacCormack's book The Education of Adult Prisoners (193l) reported that inmates participating in a University of California extension program at San Quentin performed at a high level of academic performance, with 57% of the student papers submitted receiving a grade of "A" and less than 1% receiving a grade of "F". MacCormack reported that university authorities stated that San Quentin student work was of higher quality and fewer prison students dropped out than was true in the community programs.

MacCormack also reported that inmate students had higher completion rates and achieved higher grades than their community counterparts in a University of Wisconsin extension program at Wisconsin State Prison. He postulated that maturity and fewer distractions may distinguish inmate students from their community counterparts. However, the degree to which the community and prison programs were equivalent (instruction, curriculum, materials and grading) is unknown. Differences in any of these factors may have been responsible for the reported differences.

Writing in the Journal of Correctional Education, Edwards-Wiley and Chivers (2005) reported that Ball State University faculty evaluated inmate students as having academic ability equal to campus students. However, inmate college students were judged to have put greater effort in their courses than campus students. Again, motivation of prison students was seen as a factor in explaining the different outcomes for students of similar abilities. This finding may have resulted from highly limited post secondary opportunities for incarcerated students who would have been highly motivated to participate in this limited opportunity. The reported higher levels of motivation did not appear to be based on objective measures (attendance, for example) between the two groups of students, but rather on instructor reports.

Mack and Duttlinger, who wrote in the 1991 Yearbook of Correctional Education, reported higher grades were achieved by inmate college students than their campus counterparts taking the course with the same curriculum and texts. Generally, the inmate population over-represented those characteristics associated with lower levels of achievement (male, minority, learning disabled, and school drop outs). Admittedly, these factors are associated with the general inmate population, not the small percentage involved in post secondary education. In a population where approximately 50% are high school dropouts and academic achievement levels at intake are widely reported to be in the 6th to 8th grade range in math and reading, inmates who qualify for college level programs are exceptional.

The researchers concluded that campus students should be performing at higher levels than their prison counterparts as a result of having greater resources for support and remediation available on the college campus than at the prison. The survey findings suggested factors such as motivation, maturity and peer support which may account for the difference between the academic performance of the campus and inmate college students and, by extension, may suggest why inmates unexpectedly scored higher on portions of the NAAL than their community counterparts.

Both faculty and campus and prison students were asked to explain the findings. For faculty respondents (27 respondents-78% return rate), responses related to student motivation and commitment, having fewer distractions and more time to study accounted for 54% of the responses. Other factors included: peer support 15%, maturity (age) 12% and qualifications (12%).

The student responses closely tracked those from the faculty survey with 53% of students reporting. They indicated that differences in performance were due to high levels of motivation and commitment and fewer distractions (33%). Peer support was mentioned by 11% of the students. Maturity was mentioned nearly twice as often by the faculty compared to the students despite the similar average age (33 yrs. for inmates and 30 years for campus students).

Finally, when asked why they enrolled in college courses, inmates noted career goals (92%), self satisfaction (84%), intellectual stimulation (74%), and post release employment (65%) as the top four reasons for participation, reasons which suggest that the prison students were highly motivated.

The literature which suggested that inmate students can be highly motivated and achieve at levels higher than their community counterparts is not the only conclusion reached by researchers studying how inmates and community students compare. Writing in the Journal of Correctional Education, Brahmasrene (2001) examined the academic achievement of inmates and regional campus students who had participated in a post secondary course in economics. Unlike Mack and Duttlinger, Brahmasrene employed a standardized test published by the National Council on Economic Education (NCEE) to compare the two student groups rather than a more subjective comparison based on faculty reports. Brahmasrene reported there were no significant differences between the inmate students and the campus based students. His findings, based on objective criteria, may question the findings of Mack and Duttlinger or they may simply be the result of two different groups of students in different programs with different levels of performance.

Another study of inmate students with a comparison group of community based students was conducted at a federal correctional institution and a comparison group of students at a nearby technical institution. In that study, Wood (1978) reported no differences in reading comprehension on standardized measures. Again, the finding is based on a post secondary education program and a group of inmate students who were atypical of the general inmate population. Although the inmates did not score higher than their community counterparts, they did achieve at the same level and thus reinforce the conclusion that high levels of academic performance by inmates is not unusual or unexpected.

Findings From The Naal

The NAAL offers a wealth of information to shed light on the finding that prison inmates with a GED read better than their counterparts in the general population. The NAAL included a nationally representative sample of adults 16 years of age and older residing in the general population who were personally interviewed and tested to obtain estimates of their literacy competence. In addition, a nationally representative sample of prison inmates was included for the survey. Approximately 18,000 adults in the general population and 1,200 prison inmates became NAAL participants.

Sampled respondents in the general population were asked about their general language and education background, their social, political, labor force, and welfare participation, their literacy practices, and demographic characteristics. Prison inmates were administered a similar questionnaire, and additional questions were added to ascertain their criminal history, current offenses, and time served on their sentence. Questions not relevant to incarcerated individuals but of interest for the general population were omitted from the prisoner questionnaire.

After the general interview, sampled respondents were administered a series of tests to measure their prose, document and quantitative literacy levels. Prose literacy estimates the knowledge and skills needed to search, comprehend, and use information from such materials as editorials, news stories, brochures, and instructional materials. Document literacy consists of the knowledge and skills to understand documents, such as job applications, payroll forms, transportation schedules, maps, tables and drug or food labels. Quantitative literacy includes identifying and performing computations, either alone or sequentially, as balancing a checkbook, figuring out a tip, completing an order form, or determining the amount of interest on a loan from an advertisement. All the test items measured skills needed in the adult world from test materials taken from adult experiences.

In this paper we are looking at the differences between the prison and general populations on their prose literacy and in particular at the differences for those in the prison and general populations whose highest level of educational attainment is the GED or another high school equivalency. This group constituted 28% of prison inmates and 5% of those in the general population (table 1). In this group, prisoners scored 10 points higher (270) than those in the general population (260) on the prose literacy scale. This difference is a statistically significant difference.

Table 1. Percentage of the adult prison and household populations
and their mean prose literacy scores by educational attainment, 2003

Educational Population Prose
attainment literacy
 score

 Prison Household Prison

Total 100 % 100 % 257

Less than high 9 % 9 % 199
school

Some high 28 10 * 235
school

GED/high 28 5 * 270
school
equivalency

High School 13 26 * 264
graduate

Post 22 51 * 282
secondary

Number 1,358,298 214,440,093 1,358,298

Educational
attainment

 Household

Total 275 *

Less than high 160 *
school

Some high 228
school

GED/high 260 *
school
equivalency

High School 262
graduate

Post 302 *
secondary

Number 214,440,093

* Significantly different from prison population.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

(Footnote: Standard t tests were used to determine statistical significance. Statistical significance is reported at p<.05. Differences between averages or percentages that are statitically significant are discussed by using terms such as higher or lower. Differences that are not statistically significant are referred to as not significant or not different.) Among both prisoners and householders, those with a GED scored about the same as those with a high school diploma on reading and significantly better than those with only some or no high school.

Since prisoners as a whole did not do as well as the general population on the NAAL prose scale, one would expect that prisoners with a GED would not do as well as the same group in the general population. In the total sample prisoners scored on average 257 and the household population scored 275, a significantly higher score. It may be that the GED group in the general population was composed primarily of groups that did not do as well as others on the literacy tests (for example, minorities or persons who speak English as a second language) and these groups were not well represented among prisoners with a GED.

Personal characteristics

Demographically, prison and household groups with a GED look very different. Prison inmates are overwhelming male (96%) compared to persons in the general population (53%), generally minority (62% vs. 34%), younger (with two-thirds under the age of 40 compared to half) and more likely to be born in the United States (96% vs. 90%) (table 2).

Table 2. Demographic characteristics of prison and household
populations whose highest educational attainment is the GED
or high school equivalency, and their mean prose literacy
score, 2003

Demographic Population Mean
characteristics prose
 literacy
 score

 Prison Household Prison

Gender

Female 4.5 47.4 * 270 **

Male 95.5 52.6 * 270

Race

White 38.6 % 65.6 * % 275

Black 44.6 14.5 270

Other 16.9 19.9 259

Age

16-24 15.6 % 18.6 * 286

25-39 52.3 31.1 * 272

40 or older 32.1 50.3 259

Country of birth

US 96.4 90.4 * 272

Other 3.6 9.6 * ***

Demographic
characteristics

 Household

Gender

Female 263 *

Male 257

Race

White 270

Black 233 *

Other 244

Age

16-24 258 *

25-39 268

40 or older 256

Country of birth

US 264 *

Other 210

* Significantly different from the prison population.

** Estimate based on fewer than 45 unweighted cases.

*** Too few cases to produce an estimate.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

Males in the prison GED population scored higher than GED males in the general population on the prose literacy scale (270 to 257). In the total prison and household populations the achievement is reversed with males in prison scoring below those in the general population (257 to 273). So it would be expected that the male GED group in the general population should score higher than that of prisoners. However, the reverse is true. Male prisoners in the GED group scored about the same as all males in the general population, including those with much more educational training.

Blacks in prison with a GED are better readers than those in the household population and read about the same as whites in prison and whites in the general population. Blacks in prison with a GED scored significantly higher than blacks in the general population ((270 to 233). Blacks with a GED in prison scored 270, whites in prison 275 and whites in the general population 270. The high literacy rates of blacks in prison, who comprise 45% of the prison GED population compared to 15% for the household population, is one of the factors that make literacy rates higher for the prison GED population than those in the household population.

Young inmates, age 16 to 24 with a GED, read better than their counterparts in the general population (286 to 258), another group that pushes up the literacy score for prisoners compared to those in the general population.

Disabilities

Generally, it is thought that disabilities can contribute to an inability to perform cognitive tasks such as reading. Then it could be expected that the presence of a disability would affect a person's ability to read. A higher percentage of people with disabilities, and in particular learning disabilities, in the GED group in the general population could help explain their lower prose scores. However, roughly the same percentage of those in the prison and household populations with a GED reported a disability (35% and 37%) (table 3). These disabilities included difficulty seeing or hearing, a learning disability or a disabling physical condition.

Table 3. Physical disabilities of adult prison and household
populations whose highest educational attainment is the GED
or high school equivalency, 2003

 Population

 Prison Household Prison

Physical 100.0 % 100.0 %
disabilities

Any reported
disability

Yes 35.4 % 37.4 % 261

No 64.6 62.6 275

Difficulty seeing

Yes 12.9 14.6 255

No 87.1 85.4 272

Difficulty hearing

Yes 7.6 12.5 * 250

No 92.4 87.5 272

Learning disability

Yes 15.2 6.4 * 270

No 84.8 93.4 270

Other
disabling
physical
condition

Yes 14.3 20.9 * 258

No 85.7 79.1 272

 Prose
 literacy
 score

 Household
Physical
disabilities

Any reported
disability

Yes 254

No 263

Difficulty seeing

Yes ** 246

No 262 *

Difficulty hearing

Yes ** 256 *

No 260

Learning disability

Yes 248 *

No 261

Other
disabling
physical
condition

Yes 257

No 261

* Significantly different from the prison population.

** Estimate based upon fewer than 45 un-weighted cases.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

All those who reported a disability scored lower than their counterparts-261 compared to 275 for prisoners and 254 and 263 for householders. Prisoners without disability scored higher than those in the general populations (275 and 263) but prisoners and householders who acknowledged some disability did not score significantly different (261 and 254).

Of all disabilities, a learning disability is expected to affect the ability to read more than any other. A higher percentage of those in prison said they had a learning disability compared to those in the general population (15% and 6%). Although it might be expected that prisoners with a learning disability would score lower than others, prisoners with and without a learning disability had the same prose scores (270), indicating that the learning disability either made no difference or had been overcome. For those in the general population, a learning disability did make a difference, with those with the disability scoring lower than those without (248 and 261). Having a disability did not make a difference in prisoners prose score (both were 270) but it did for those in the general population (248 vs. 261).

Both prisoners who reported a learning disability and those who did not scored higher on the prose scale than those in the general population (270 vs. 248 for those with a disability and 270 vs. 261 for those without.). Hence, another difference between the GED graduates is the strong showing of those with a learning disability in prison compared to those in the general population, even though a much higher percentage of prisoners had a learning disability.

Experiences with Languages

The United States is a country with immigrants from many lands who do not speak English or learned it after childhood. It may be that individuals' language experiences, particularly when young, have impeded their ability to read English. Compared to prisoners, the household population reported a significantly higher percentage of people who spoke another language before school (16% vs. 8%), who first learned another language before English (7% and 3%) and who usually speak another language now (5% vs. 1%) (table 4). These groups, although relatively small, would be expected to lower the prose scores for the general population more than for those in prison.

Table 4. Languages of prison and household populations whose
highest educational attainment is the GED or high school
equivalency, and their mean prose literacy score, 2003

 Population Mean
 prose
 literacy
 score

Languages Prison Household Prison

Spoken
before
school

English 91.8 % 84.3 * % 272 ***
only

Other 8.2 15.7 * 250

Spoken at
home while
growing
up

English 84.4 % 81.1 % 273
only

Other 15.6 18.9 257

Language
first
learned to
read and
write

English 97.1 % 93.3 * % 271

Other 2.9 6.7 * **

Languare
usually
spoken
now

English 98.8 % 94.6 * 271

Other 1.2 5.4 * **

Languages Household

Spoken
before
school

English 264 *
only

Other 232

Spoken at
home while
growing
up

English 265 *
only

Other 234 *

Language
first
learned to
read and
write

English 263 *

Other 195

Languare
usually
spoken
now

English 263 *

Other 187

* Significantly different from the prison population.

** Too few cases in cell to calculate.

*** Estimate based on fewer than 45 unweighted cases.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

Householders whose experience was with a language other than English scored lower on the prose literacy scales than other householders. For householders, those who spoke a language other than English before school scored 232 on the prose scale while those whose only experience was English scored 264. Similar patterns were demonstrated with the other measures of language experience. In particular, those who learned to read and write in another language scored significantly lower than those who only learned English (195 vs. 263). So the presence of higher percentages of householders who learned another language while young can contribute to the lower scores on the prose literacy scale.

However, those with only English experience in the general population also continued to score lower than prisoners. Those in the general population who spoke only English before school scored 264 while prisoners scored 272. The relationship was the same for those who spoke only English at home (265 vs. 273), learned to read and write English first (263 vs. 271) and usually speak English now (263 vs. 271).

Library Use

Prisoners have a different life style from those who are free to go places as they please, and prisoners generally have more free time to engage in activities that can increase their literacy skills. Almost half of prison inmates with a GED report that they went to their library at least weekly compared to one/tenth of those in the general population (table 5). Relatively few did not frequent a library. About 2 in 10 of the prison population and 4 in 10 of the household population never went to the library. In the populations who never went to the library, prisoners scored significantly better than householders in their reading abilities (268 and 254).

Table 5. Library use by prison and household populations whose
highest education attainment is the GED or high school equivalency
and their mean prose literacy scores, 2003

 Population Mean
 prose
 literacy
 score

Library use Prison Household Prison

At least weekly 46.7 % 9.1 * % 272

At least once or 29.8 50.5 * 270
twice a year

Never 23.5 40.4 * 268

Borrowed from
the library in
the past month

Yes 46.5 % 12.9 * % 278

No 53.5 87.1 * 264

Library use Household

At least weekly 265

At least once or 264
twice a year

Never 254 *

Borrowed from
the library in
the past month

Yes 280

No 257

* Significantly different from the prison population.

** Too few cases in cell to calculate.

*** Estimate based on fewer than 45 unweighted cases.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

For both prisoners and householders, those who borrowed a book in the previous month scored significantly better on the prose scale than those who did not. However, borrowing reading material does not differentiate prisoners from householders. Both groups scored about the same on the literacy score (278 and 280). A much larger percentage of prisoners compared to householders borrowed a book (27% vs. 13%). This difference in the size of the groups would result in prisoners obtaining higher overall prose scores, since a higher proportion fall into the borrowing group.

Reading frequency

Borrowing books from the library itself does not increase ease of reading. Most education specialists say that the act of reading is what increases one's facility at reading more than any other activity. It could be that prisoners, who can have extensive periods of free time, also read more than those in the general population and that gives them the edge in taking the prose literacy test.

A significantly higher percentage of GED prisoners report reading newspapers, magazines, and books daily than those in the general population (47% of prisoners vs. 39% for householders for newspapers and magazines and 50% of prisoners vs. 22% in the household population for books) and a significantly higher percentage of householders report reading letters and notes daily (31% vs. 44%)(table 6).

Table 6. Reading practices of prison and household populations
whose highest educational attainment is the GED or high
school equivalency and their mean prose literacy score, 2003

 Population Mean
 prose
 literacy
 score

Reading Prison Household Prison Household
practices

Total 100.0 % 100.0 * %

Frequency %
of reading
newspapers
or
magazines

Every day 47.2 38.7 * % 270 262

One or more 35.6 44.9 * 273 261
times per
week

Less than 17.2 16.4 264 252
once a
week/never

Frequency
of reading
books

Every day 50.3 % 22.0 * % 270 266

One or more 34.7 28.7 277 260
times per
week

Less than 15.0 49.2 * 253 257
once a
week/never

Frequency
of reading
letters or
notes

Every day 31.4 % 44.1 * % 276 266

One or more 50.4 34.1 * 271 257
times per
week

Less than 18.3 21.5 258 252
once a
week/never

Reading
practices

Total

Frequency
of reading
newspapers
or
magazines

Every day

One or more *
times per
week

Less than
once a
week/never

Frequency
of reading
books

Every day

One or more *
times per
week

Less than
once a
week/never

Frequency
of reading
letters or
notes

Every day *

One or more *
times per
week

Less than
once a
week/never

* Significantly different from the prison population.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

Generally prisoners and householders who reported reading newspapers, magazines or books every day had about the same prose scores (270 vs. 262 and 270 vs. 266) although prisoners who read letters or notes daily scored higher than householders who did (276 vs. 266). So prisoners and householders who read daily or almost daily had similar reading skills but since smaller percentages of householders read frequently, reading practices of householders contributed to their lower scores.

TV watching

TV watching is sometimes viewed as an important factor contributing to the lowering of reading ability in the country by supplanting time spent reading. Since prisoners live in a much more structured environment than householders, they may be allowed to watch TV to keep them quiet. Surprisingly, prisoners watched less TV than householders. Almost a third of prisoners indicated that they watch TV an hour or less a day compared to a fifth of householders (table 7). Among those who watched two or more hours of TV per day prisoners scored higher on prose literacy than householders (275 vs. 259) for those who watched 2-3 hours and (267 vs. 258) for those watching 4 or more hours.

Table 7. TV watching by prison and household populations whose
highest educational attainment is the GED or high school
equivalency and their mean prose literacy score, 2003

 Population Mean
 prose
 literacy
 score

 Prison Household Prison

Hours watch TV,
videotapes, or DVDs
each day

1 hour or less ** 31.6 % 19.1 * % 268

2-3 hours 34.9 43.5 * 275

4 or more hours 33.5 37.4 267

 Household

Hours watch TV,
videotapes, or DVDs
each day

1 hour or less ** 265

2-3 hours 259 *

4 or more hours 258 *

* Significantly different from the prison population.

** Includes those who never watch TV.

Source: U.S. Department of Education, Institute of Education
Sciences, National National Center for Education Statistics,
2003 National Assessment of Adult Literacy.

For these groups TV watching did not appear to have an effect of their reading. None of the differences in prisoners scores based on their hours watching TV tested as being statistically different. The same was true for the household population.

Computer usage

Prisoners were significantly less likely to be using a computer compared to those in households. Since prisoners live in a restricted environment, they are less likely to have access to computers as a result of Internet security concerns, fewer computers, less instruction in their use. and less technology for instruction. Over three-quarters of prisoners said they had never used a computer compared to a third in the general population. Inmates were even less likely to have used specific computer applications common to the work and school worlds of those in the general population. Less than one in ten had used a word processor, a spreadsheet, or a CD-ROM to look up information, while about a third of the householders had used a word processor or a CD-ROM and almost a quarter had used a spreadsheet. (Table 8)

Table 8. Computer use of adult prison and household populations
whose highest educational attainment is the GED or high school
equivalency and their mean prose literacy score, 2003

 Population Mean
 prose
 literacy
 score

 Prison Household Prison

Total 100.0 % 100.0 %

Ever use
computer

Yes 21.8 % 63.8 * % 269

No 78.2 36.2 * 271

Write using
a word
processor

Yes 9.5 % 37.0 * % 267 **

No 90.5 63.0 * 271

Use a
spreadsheet

Yes 4.6 % 22.5 * % 274 **

No 95.4 77.5 * 270

Look up
information
on a CD-ROM

Yes 8.5 % 32.8 * & 268 **

No 91.5 67.4 * 270

 Household

Total

Ever use
computer

Yes 268

No 245 *

Write using
a word
processor

Yes 276

No 251 *

Use a
spreadsheet

Yes 274

No 256 *

Look up
information
on a CD-ROM

Yes 277

No 252 *

* Significantly different from the prison population.

** Estimate based on fewer than 45 un-weighted cases.

Source: U.S. Department of Education, Institute of Education
Sciences, National Center for Education Statistics, 2003
National Assessment of Adult Literacy.

For prisoners using a computer was not related to literacy rates. None of their mean literacy scores were statistically different irrespective of their computer usage. For householders, computer usage separated out the more literate from those with lower scores. Those using a word processor, a spreadsheet, or a CD-ROM scored significantly higher on the prose scale than those who had not used those computer applications.

Mandatory Education and Incentives

Mandatory education is a policy adopted by the Federal Bureau of Prisons in 1981 and implemented by a number of states in subsequent years. This policy requires educational participation of offenders who enter the correctional system testing below various established education thresholds. Mandatory education had been adopted by 22 states by 2003, as determined by a survey conducted by one of the authors, H.D. Jenkins, and Jerry McGlone 2002). In its various forms, mandatory education is linked with numerous incentives (parole consideration, time off sentences, stipend/wages) for participation as well as a variety of penalties for nonparticipation.

Mandatory education may be a factor in explaining the unexpected and higher levels of academic achievement on the NAAL by requiring educational participation on the part of inmates who, absent mandatory education, may not have voluntarily participated in educational programming while in prison. Correctional education is "delivered" to the inmate students without cost and other barriers (transportation, child care) which face some adults in the community.

One of the authors, C.W. Harlow (2003), reported that state inmates participated in education programs (ABE, GED, VOC and post secondary) at an overall rate of 57% in 1991 and 52% in 1997. That is a rate significantly higher than participation levels in the community for adults. Additionally, correctional education is generally delivered several days a week and for several hours a day, suggesting an intensity of the educational experience not provided most community based adult education students.

Although states participating in the NAAL are not identified, and thus cannot be correlated with the states with mandatory education requirements, the Federal Bureau of Prisons and the states participating in NAAL are likely to have included some mandatory education prisons in the study. At these locations, inmates who may not have participated in educational programming in the community were involved in programs that could have resulted in their outperforming their community counterparts on the NAAL.

Polling conducted by the authors reconstructed state and Federal systems of incentives for participation in correctional education in 2003 when the NAAL was conducted. It is instructive to understand what incentives were generally in place during that time period to determine how inmates may have been encouraged to participate in correctional education and thus may have been more likely to be involved in an educational program than their community counterparts.

In an attempt to understand the difference between the incarcerated students and students in the community, incentives in the correctional system may serve to involve inmates who would not otherwise have participated in education, thus better preparing them academically to outscore their community counterparts in the NAAL. Time off sentences and improved opportunities for parole are among the most powerful incentives available for an incarcerated population.

For this study the authors surveyed the Federal Bureau of Prisons and the departments of corrections of all states, asking them about the kinds of incentives each provides for educational participation. In an email to each department we asked if they provided any incentives for educational participation. Incentives reported include the following (table 9):

Table 9. Number of Correctional Systems Reporting the Use Of
Incentives for Educational Participation

Incentives for Program Participation Systems
 Reporting

Total Department of Corrections reporting 21

Total reporting use of incentives 19

Time off sentence for all programs including education 7

Time off sentence for educational participation only 6

Pay for school participation/assignment 11

Special consideration for parole with educational 4
participation

Qualifications for higher level jobs or assignments within 9
system

No incentive 2

Note: Individual incentives add to more than total because systems
may use more than one type of incentive.

These incentives are uniquely provided in a prison environment and cannot be applied to the household population in the same manner as can be achieved within the controlled environment of a prison.

Twenty state Departments of Corrections and the Federal Bureau of Prisons responded to our emails and follow up telephone calls. Of these 21 departments, 19 indicated that they used incentives to motivate prisoners to work to gain an education. Table 9 reports the survey findings. Of the 21 Departments reporting, 11 departments said they gave pay for school participation (generally as inmates' work program), 9 used educational attainment as a qualification for higher level work assignments, 7 gave time off sentence for program participation, including educational participation, 6 gave time off their sentence for participation in education programs, and 4 said they gave special consideration to education participants in considering parole. Other states which did not respond to our survey may also have used these or other incentives to motivate prisoners to participate in education programs.

Ged Pass Rates In 2003

The GED Testing Service has supplied us with information on persons both in the general population and in correctional settings who took and passed the GED test. These data are not generally published and the authors are grateful to the GED Testing Service staff for the data (Table 10).

Table 10. Number of candidates who were tested, completed the test
and passed the GED in 2003.

Testing Tested Completed Completion Passed Pass
Centers rate Rate

Total 668,618 561,932 84.0 % 393,688 70.1 %

In 29,887 23,306 78.0 17,132 73.5
correctional
facilities

In other 638,731 538,626 84.3 376,556 69.9
facilities

Within the territorial United States approximately 670,000 persons took the GED in 2003 (Table 10). Of these, approximately 4% were tested in correctional facilities. Although a lower percentage of prison inmates completed the test, of those that did, a higher percentage of prisoners than the general population who completed the test passed (74% vs. 70%). Correctional education program administrators say that they generally do not allow prisoners to take the GED until they are ready to pass, since they must pay a fee for each prisoner who takes the test.

The GED tests include sections on writing, social studies, science, reading and math. On the whole, those who took the test in correctional settings scored on average somewhat lower than those in the general population in each section and on their total score (Table 11). The total average score for prisoners was 2,606 and those in the general population 2,669.

Table 11. Standard mean scores for GED test passers, 2003

Testing Number Total Writing Social Science Reading
Centers of Studies
 passers

Total * 373,007 2,667 511 537 553 569

Correct. 17,049 2,606 493 529 542 560
Facilities

Other 355,958 2,669 512 537 554 569
places

Testing Math
Centers

Total * 496

Correct. 483
Facilities

Other 497
places

* Due to incomplete data, total numbers are different from the
number of passers in the above table.

In recent years it has been widely reported verbally to some of the authors by state directors of correctional education that the GED passing rate and overall scores for prisoners are higher than for other GED takers. Official GED Testing Service data do not support such contentions on a national level. Individual states, however, may still experience higher pass rates for their prisoner population than their other GED takers.

Analysis

In the foregoing literature review and in-depth analysis of NAAL and GED Testing Service data a number of possible explanations for the higher performance of inmates on prose literacy were discussed.

Literature review suggested factors which account for the difference between the academic performance of the on campus and inmate college students and, by extension, may suggest why inmates with a GED scored higher on reading on the NAAL than their community counterparts. These included student motivation and commitment, fewer distractions and more time to study. Other factors included peer support, maturity (age) and qualifications. Inmates noted career goals, self satisfaction, intellectual stimulation and post release employment. All of these suggest that inmate students can be highly motivated and achieve at a level higher than their community counterparts. Although the inmates did not generally score higher in college than their community counterparts, they did achieve at the same level and thus reinforce the conclusion that high levels of academic performance by inmates is not unusual or unexpected.

Generally, factors associated with difficulty reading in the general population did not work to lower NAAL scores among prisoners. For example, literacy rates of blacks in prison were the same as white prisoners while blacks in the general population scored lower than whites, one of the factors that make the literacy rates higher for the prison GED population than those in the household population. Young inmates, age 16 to 24 with a GED, read better than their counterparts in the general population and constitutes another group that pushes up the literacy score for prisoners compared to those in the general household population. Prisoners with a learning disability scored as well as other prisoners, a phenomenon not seen in the household population. And prisoners who watched TV read as well as prisoners who didn't, which was not true of the household population

The NAAL data also show that prisoners and householders who read daily or almost daily had similar reading abilities, but smaller percentages of householders read frequently. Reading practices of householders may have contributed to their lower scores. Reading scores, like many other life activities, improve with practice.

For both prisoners and householders TV watching did not have an effect on their reading. None of the differences in prisoners NAAL scores based on their hours watching TV tested as being statistically different. The same was true for the household population.

For householders computer usage separated the more literate from those with lower scores. Those using a word processor, a spreadsheet, or a CD-ROM scored significantly higher on the prose scale than those who had not used those computer applications. For prisoners, using a computer was not related to literacy rates. None of the literacy scores of prisoners was differentiated by whether or not they used a computer and its applications.

Since the 1980s many states and the Federal Bureau of Prison have instituted mandatory education policies which include a wide variety of incentives for educational participation. Many prisoners who would otherwise not enroll in education programs participate in education to earn time off their sentence, better assignments and other rewards during incarceration. Whether or not the incentives produce a higher level of achievement among incarcerated students has not been proven statistically.

The 2003 data shared with the authors by the GED Testing Service indicated that students who took the GED test in correctional settings scored on average somewhat lower than those in the general population in each of the five test sections and on their total score. Since there are many samples of prose literacy in the GED reading test, one could have expected that prisoners would perform better than their householder counterparts. However, the population of prisoners who took the GED test in 2003 was only a small part of the prisoners represented in the NAAL and, therefore, these groups may be somewhat dissimilar. In any case, 2003 GED testing data do not help explain the higher scores for prose literacy on the 2003 NAAL survey.

Conclusions

A major finding of this study is that prisoners do read more and watch less TV than their householder counterparts. This is encouraging because it underlines the value of encouraging people to read in all settings. Prisoners have more time and opportunity to read than householders and do so. The result is a higher prose score.

There are many criticisms of the correctional systems in the United States, but the fact that the setting encourages education and more reading is a very positive finding. While the data from this study are not extensive it does serve to inform policy and procedure. The creation of more resources and opportunities for reading is important for everyone, especially for prisoners. In fact, it also improves public safety.

It is has been demonstrated in a number of studies of the relationship of education and recidivism that inmates who participate in education are significantly less likely to commit crimes again. And, all this results in significant future savings to the taxpayer and lower crime rates (Three State Study by Steurer, Smith and Tracy and Bazos and Houseman). The NAAL study seems to reinforce the general perception that "education reduces crime".

References

Aos, S., Miller, M. and Drake, E., Evidence Based Adult Correctional Programs: What works and What Does Not, Washington State Public Policy Institute, 2000.

Bazos, A. and Housman, J., Correctional Education as a Crime Control Program, USLA School of Public Policy and Social Research, Department of Policy Studies, Correctional Education Association, 2004.

Brahmasrene, T., Comparing the Academic Achievement of Inmates with Regional Campus Students in Economics Courses, Journal of Correctional Education, 54 (4), 152-155, 2001.

Coley, R. and Barton, P., Locked Up and Locked Out: An Educational Perspective on the U.S. Prison Population, Educational Testing Service, 2006.

Edwards-Wiley, E. & Chivers, T. & N., Perceptions of Inmate Student's Abilities to Succeed. Journal of Correctional Education, 56 (1), 65-86, (2005).

Greenberg, E., Dunleavy, E., Kutner, M. and White, S., Literacy Behind Bars: Results from the 2003 National Assessment of Adult Literacy Prison Survey, National Center for Education Statistics, US Department of Education, 2006.

Haigler, K.O., Harlow, C.W., O'Connor, P. and Campbell, A., Literacy Behind Prison Walls: A Profile of the Prison Population from the National Adult Literacy Survey, National Center for Education Statistics, US. Department of Education, 1994.

Harer, M., Recidivism and Federal Prisoners Released in 1987, Office of Research and Evaluation, Federal Bureau of Prisons, 1994.

Harlow, C.W., Education and Correctional Populations, NCJ 195670, Bureau of Justice Statistics, U.S. Department of Justice, 2003.

Jenkins, H.D. and McGlone, J., Status of Mandatory Education in State Correctional Institutions, Office of Vocational and Adult Education, U.S. Department of Education, 2002.

Kirsch, I.S., Jungeblut, A., Jenkins, L. and Kolstad, A., Adult Literacy in America: A First Look at the Results of the National Adult Literacy Survey, National Center for Education Statistics, US. Department of Education, 1993.

MacCormick, A., The Education of Adult Prisoners, National Society of Penal Information, 1931.

Mack, M. and Dittlinger, L., Inmates and Regional Campus Students: A Comparison of Academic Achievement, Yearbook of Correctional Education, 213-228, 1991.

Martinson, R., What works?--questions and answers about prison reform, Public Interest, 35, 22-54, 1974.

Saylor, W. & Gaes, G., PREP: A Study of Rehabilitating Inmates Through Industrial Work Participation and Vocational and Apprenticeship Training, Office of Research and Evaluation, Federal Bureau of Prisons, 1996.

Sherman, L., et al., Preventing Crime: What Works, What Doesn't, What is Promising, The National Institute of Justice, U.S. Department of Justice, 1996.

Steurer, S., Jenkins, H.D., and Harlow, C.W., Unpublished survey of state and Federal Directors of Correctional Education on incentives used to encourage prisoner participation in educational programs, 2008.

Steurer, S., Smith, L., and Tracey, A., Education Reduces Crime: Three-State Recidivism Study, Centerville, UT: Management and Training Corporation, 2003.

Wilson, D., Gallagher, C., and MacKenzie, D., A Meta Analysis of Correctional Based Education, Vocational, and Work Programs for Adult Offenders, Journal of Research in Crime and Delinquency, 37 (4) 347-368, 2000.

Wood, D., Learning Assistance for Inmates in a Federal Correctional Institution, Paper presented at the Annual Meeting of the Western College Reading Association. March, 1978.

Yahner, J. and Visher, C., Illinois Prisoner's Reentry: Success Three Years After Release, Urban Institute, 2008.

Caroline Wolf Harlow, Ph.D.

H. David Jenkins Ph.D.

Stephen Steurer, Ph.D.