Last spring a small group of my colleagues and I corresponded about the idea that the number of children with reading problems predicts the need for prison beds. In “Internets, Help Me Find A Citation! Does Any State Include Education Data in Estimating Future Needs for Prison Capacity,” Liz Ditz asked for help with that same popular idea. I sent her some notes from a search that I’d conducted last spring on the topic. Although you can read my notes there, the short of them is that there’s a lot of talk about the idea, but not much data. So, why not get some data and examine the question?
Can one predict imprisonment from levels of reading competence? If there is a predictive relationship between imprisonment and reading, are there other factors (e.g., poverty) that might account for it? I decided to make a quick-and-dirty run through the data that I could snag easily (please read the caveats). I pulled data from these sources
- The US Department of Justice on prison populations for each of the states in the US; these data show the percentage of each state’s population who are in jail or prison for stays of longer than one year;
- The US. Department of Education National Assessment of Educational Progress for each of the states in the US; the data are for the percentage of school populations that test below the basic level of reading competence at the fourth and eighth grades;
- The US Census Bureau on the percentage of people living in poverty for each of the states.
I correlated these data. The table at the right shows the results.
As is clear from the correlations, there appears to be a relationship between the percentage of prisoners in the states with the percentage of students who read below the basic level (r = .56 and .53 for 4th and 8th grades respectively). The inference that many will want to make from those data is that lower levels of reading cause higher levels of imprisonment.
But, not so fast, cowkid. Check the correlation between the percentage of people living in poverty and the percentage imprisoned. I’ve not had a chance to run the statistical comparison of them to verify this, but that correlation (r = .624) looks like it’s larger than the ones for reading competence (r = .56 and .53 for 4th and 8th grades respectively). Furthermore, the correlations of percentages of people living in poverty and percentages of students reading below basic at 4th and 8th grades (r = .678 and .614, respectively). Looks like there’s a need to employ a more complex model than this simple examination of bivariate correlations.
When I get the time to work with these data, I’ll let folks know the results. I’d welcome help with both the thinking and the analyses.
Caveats: For several reasons, these results should not be taken as definitive. There are likely other factors that also affect these variables separately and in combination. There are also potential statistical features of the data (e.g., restricted ranges in the measures) that have not been examined in this analysis. Remember that this is just a blog entry, not a peer-reviewed research paper.
Sphere: Related Content
Poverty correlates with most of our social ills including illiteracy and imprisonment. To suggest that illiteracy causes criminality (or imprisonment–maybe illiterate criminals get caught more than literate ones) goes a little bit beyond the data as John admits in his caveats but it opens up an interesting topic to blog about. While we are speculating way beyond what data will allow, I want to hazard a guess that the connection between poverty and prison is mediated by school failure which in turn is partly predicted by low reading levels although serious, repeated disciplinary problems are likely a stronger predictor of both school failure and eventual imprisonment. I am remembering Hill Walker telling us in a seminar that he believed imprisonment could be predicted very accurately by weighing the size of a student’s cumulative file at the end of 5th grade–the file being mostly full of behavior incident reports. One last intriguing fact—studies of adolescents in juvenile incarceration suggest that about 30% have learning disabilities.
George, these are wonderfully comments. Your insight and contextualization (may I make up that word?) help to remind us that these simple statements (% reading problems ==> number of needed prison beds) are simplifications.
Also, I agree that there is likely a relationship between weight of cumm files and imprisonments. Hill likely had that right (among the many things he had right!). What should we do about these relationships?
I suggest that we explain them clearly when we talk with the press. Your more nuanced comments point the right direction.
It is true. I have actually interviewed someone who is involved with building prisons and he confirmed that they look at student test scores for projections. I am an education writer currently working on a story about this. Sometimes data won’t tell you what people will, if you bother to ask.
Sherry, thanks for the news. It is valuable to talk with people and get their observations and opinions. Did your informant have any evidence about the frequency or likelihood of construction planners checking reading results? Did your informants also disclose that they look at poverty rates?
Now, that would be an interesting survey to conduct: Find the folks who bid on contracts to build and actually do build prisons. Interview them intensively….
Sherry, I would love to hear more about your research into this.
The connections between lack of education (in particular reading/literacy and vocabulary), poverty and crime seem obvious… yet the fact that the correlation is accepted enough to become part of prison planning, but not to mobilize our entire society to make sure our kids learn to read by 3rd grade is just mind-boggling to me.
Now there’s another drastic 10% cut to the education budget in California, along with the state taking an expensive loan to build more prison beds. This insanity does not make any sense, not practically, morally, or economically.
The nonprofit I manage, eSpindle Learning, will provide free vocabulary tutoring licenses to all 3rd graders in the SF Bay Area for the 2008/09 school year. We’re a small organization, and this will be a big push for us, but I seriously losing sleep about these stats along with what I see just in my immediate surroundings and talking with teachers and administrators.