![LockUp](/sites/g/files/toruqf261/files/styles/8x3_1440w_540h/public/2020-09/Lock-up-920x425_0.jpeg?itok=KXHJ9BvN)
The Demobilizing Effects of Pretrial Incarceration
Problem: This study will examine the effect of pretrial incarceration on political participation. To date, scholars have reported conflicting conclusions about the effects of the carceral state on voter turnout, and no studies have focused on pretrial incarceration. Pretrial incarceration is central fixture of the “shadow carceral state,” which refers to practices and institutions that exist outside of formal sentencing but rely on the coercive power of the state to impose punishment (Beckett and Murakawa 2012). The omission of pretrial incarceration and the “shadow carceral state” more generally obscures key aspects of the American government and may lead scholars to underestimate its full effects on political participation.
Sources: We are assembling a dataset of court records from several jurisdictions across the U.S. that span multiple election cycles. We are collecting these records from court systems directly and from court websites through webscraping as needed. We will then use probabilistic techniques to merge these court records with voter files.
Methods: Our project will leverage sources of random variation in the likelihood that a defendant is detained pretrial, including variation that arises from sharp changes in pretrial practices and assignment to more or less lenient judges. We will use instrumental variables, difference-in-differences designs, and matching methods.
Challenges: One challenge is ensuring voting data is reliable and accurately merged with court records. Voter files are frequently updated to remove individuals for a variety of reasons (moving, death, etc.). These purges may be correlated with pretrial conditions, which could bias our results. Missing data and typographical inconsistencies across records also complicate the merge process. Another challenge is ensuring variation in pretrial outcomes is sufficiently random, which is crucial for our research design. This requires both institutional knowledge and empirical validation.
Findings: We expect that pretrial incarceration may affect voting behavior for several reasons. For many, it is highly punitive (5 months on average in some jurisdictions) and triggers economic and social losses, which are antecedents of political participation. The experience of pretrial incarceration, which is imposed disproportionately on people of color and those who are too poor to post bail, may also convey striking messages about government and political efficacy. Preliminary findings from Philadelphia reveal that pretrial incarceration triggered by high bail amounts has substantial demobilizing effects on subsequent voter turnout. This project will assess the generalizability of these preliminary findings and extend them by collecting more data and leveraging new research designs.