iowa court upholds mass restoration of felon voting rights
from the des moines register and talkleft.com: a district court in muscatine has upheld iowa governor tom vilsack's executive order restoring voting rights to all former felons in that state. as i wrote in august, the order raises some interesting legal questions. shelly schaefer and i are presenting a paper framed around some of them in a department workshop november 29:Chris Uggen and Shelly Schaefer “Voting and the Civic Reintegration of Former Prisoners”
When Iowa governor Tom Vilsack restored voting rights to all former felons in that state this July Fourth, he noted that “research shows that ex-offenders who vote are less likely to re-offend.” The National Review countered that “the problem with Vilsack’s claim is that there is absolutely no research to support it. Not one longitudinal study exists showing the effects of the restoration of voting rights on crime rates or recidivism.” We undertook such a study this summer, by matching criminal records with voting records. We conceptualize voting as a form of “civic reintegration,” analogous to the work and family ties that are well-established in life course criminology. For our 1990 Minnesota release cohort, we find that approximately 20 percent of the former felons registered to vote. Our event history analysis shows that felons who voted in the previous biennial election have a far lower risk of recidivism than non-voting felons, and that this effect holds net of age, race, gender, and criminal history. The talk will discuss the strengths and limitations of our data and covariate adjustment approach for making causal inferences, the implications of felon enfranchisement for public safety, and the viability of weaving former felons back into the citizenry as stakeholders.11/10 UPDATE: 11/29 talk cancelled (bumped for job candidates - new date to be announced).
posted by christopher uggen @ 8:49 PM
11 comments
11 Comments:
I tend to be a causal pessimist in these matters, although a good-natured one and certainly respectful of anyone putting themselves out there with the work of research. Now that there are some instances of repeals, there is at least the possibility of looking at before and after time-series comparisons of crime rates, although of course those are low-powered. I'm presuming that either (a) not many noncitizen-felons get to remain in the US or (b) there is no systematic data on the proportion of felonies perpetrated by citizens vs. noncitizens, so that you can't evaluate hypotheses about changes in the citizen/noncitizen recidivism rate.
Hey Chris, thanks for the heads up on Iowa. I would have missed that. On a related note, I presented some of your work (with Behrens and Manza) on the history of FD laws to a class, and students quickly asked about voting and recidivism. The discussion then turned to how we might evaluate this, and we debated whether it would be ethical to randomly assign some ex-prisoners as “voting eligible” and disenfranchise the control group. The students were divided on the ethics, legality, and feasibility of such a study. Any thoughts (anyone)?
well jeremy, now you can say that you concur with the national review on at least one issue -- they are causal pessimists too (though not so good-natured, i'd wager). i've made some simple graphical attempts at the interrupted time-series thing, but recidivism measures change too much over time (and, esp. space) to get much traction. i'll try to put a little section on this in the paper, perhaps looking at a few consistently measured and enforced offenses. i'd love to get your thoughts on the paper once we get it presentable.
ryan, i guess the supreme court might frown on random assignment of voting rights. my personal gendankenexperiment would test a "practical citizenship" intervention in the prisons. ethically, prisoners would probably have to volunteer for consideration, but then they'd be randomly assigned to treatment and control. those assigned to treatment status would do some sort of inside/outside voting, do service for community-based organizations that interest them (e.g., stuffing envelopes), or make calls for politicians.
i'd bet that prisoners assigned to the treatment group would be more likely to vote upon release (my intermediate outcome) and there's a fightin' chance they'd be less likely to recidivate. [yes, i recall how my last bet with you turned out :( ] i know that this wouldn't identify the independent treatment effect of voting, but that's ok with me. my goals are simple: (1)less crime; and, (2) more reintegration. i'm tempted to write a grant to actually conduct this experiment.
I would like to know the independent treatment effect of voting, in fact that is what most interesting to me about your new paper. Does voting make in independent contribution to an understanding of criminal behavior above and beyond other types of reintegration (family-- work) known to be important?
Some of your other work suggests the effect of voting washes out when other measures of reintegration are used, but your more recent piece with Shelly seems to suggest significant effects.
Its really an interesting methodological and sociological question.
et tu punk? another causal pessimist. seriously, though, you raise a good point. we should be able to establish that the effect holds net of family and work (of course, we mostly "know these to be important" based on observational studies of the sort we're estimating). would you buy a propensity score model of voting rather than a covariate adjustment approach?
in my last comment, i just meant that i don't care whether i'm getting the treatment effect of the physical act of voting or the treatment effect of a general "civic reintegration" construct measured in an experiment designed to promote civic reintegration. i suppose i could do the experiment, and then further adjust for work and family effects (presuming that the experiment also affected these processes) or incomplete randomization.
finally, though i agree with your read of the paper, i wouldn't interpret the other work quite so negatively. that research was based on a general community sample rather than a sample of convicted felons. as i recall, voting remained significant for some outcomes at .05 (and you've convinced me i should calculate my eigenvalues to see whether i had sufficient power for that analysis!).
thanks for the comments -- i think we're both facing tough challenges on the causality front.
Sorry man, I didn't mean to come off as negative... I remember thinking "most" of the voting-crime effects went away, although some held. I thought it was really neat paper.
As for propensity models, as you well know, they are just another part of the tool kit. If the findings from earlier work held, I think thats an important contribution.
The issue of voting, however, seems well suited for propensity models and "treatment effects"
To me your distinction (or lack of) between the "voting" effect and the more general "reintegration effect" is really interesting and a learning tool about research methods, maybe even ethics, from someone with the highest standards on both counts.
What does it mean if voting has no effect on crime when, for instance, marriage is included?
a) Does that mean the crime-voting relationship is spurious... and stated simply, that voting doesn't matter for crime and shouldn't be discussed as though it does?
b) Does that mean that voting is part of a larger constellation of behaviors associated with civic reintegration? If so……..
Where, and how, do you draw the line around the constellation of important, yet not significant, behaviors included in the construct of reintegrating?
I know the usual suspects, theory, previous work, ect. help some, but it strikes me as a potentially difficult issue.
Chris: Before seeing your response to my comment and your colloquy with PSP, I wrote a long post about this on my blog (or, more accurately, where I used this as an example to talk about my causal pessimism).
thanks, jeremy. i'm so glad you're thinking about these issues -- the paper will be much better if we're writing with a friendly skeptic like you in mind.
Chris!
You're breaking my heart, man.
We're not noticed, we're scorned, and when we are noticed for something other than a "Why do they vote for Bush" type rant, people think we're the state that produces all the potatoes.
And this picture?!?
That's the best picture you could think of to represent Iowa?
You couldn't have a picture of Kurt Warner, two-time NFL MVP and one-time Super Bowl MVP?
(Also a proud grad of UNI, like yours truly)
You couldn't have put up our world-leading meth production numbers?
You couldn't even give us a Tom Arnold or Ashton Kutcher nod?
For shame, city boy. For shame.
woz, i guess i could have used a slipknot picture. but why are you hating on max, the maker of fine-quality state-themed all-wood miniatures and shadow boxes? i even included a link on the iowa picture for you -- if you click it you can order one your own self for a mere twenty-six dollars. i'd suggest getting one for home and one for the office. plus, i tossed out a tommy bolin reference yesterday too. what? you don't recognize the scion of sioux city, the greatest of all iowa musicians? what do they teach in iowa history classes anyway? for shame iowegian boy, for shame...
Somehow I knew this tension would arise. As someone who considers herself "bi-cultural" Minnesotan-Iowan (having spent nine years in Iowa City), might I suggest Grant Wood?
Post a Comment
<< Home