My name is Michael Zanger-Tishler, and I am a PhD candidate in sociology and social policy at Harvard. My research interests center on the criminal legal system, social inequality, and the state, drawing on quantitative methods, the sociology of race and ethnicity, and the sociology of knowledge.

 

My dissertation, “Data and Knowledge of the Criminal Legal System in a Comparative Perspective,” is a comparative study of how macro factors related to state structure, organizations, and data restructure the types of questions that social scientists ask and the type of knowledge they create. Based primarily on over 250 interviews conducted in French, Hebrew, Arabic, and English with quantitative researchers, government officials, journalists and activists producing knowledge primarily about the French, American, and Israeli states, I explain how police departments, courts, and prisons strategically release their data, how external actors navigate and develop questions based on available data, and how knowledge of the state and social inequality are conditioned by national data cultures. I contextualize these interviews by tracking dataset usage in newspapers, activist publications, and journal articles as well as reading public records requests. In this work, I argue that understanding the state as the central actor in shaping data infrastructures is critical to understanding the possibilities for quantitative knowledge creation. In other words, I analyze how contemporary statistical knowledge is affected in previously unappreciated ways by the politics of data production, publication, and accessibility, and the advantages and limitations of using these data to critique state institutions and reduce social inequality. 

​

The framework advanced in my dissertation is inspired by encountering these data dynamics in my own statistical research on disparities in the criminal legal system. Coauthoring with sociologists, applied mathematicians, economists, political scientists, and legal scholars, I conduct statistical research on the criminal legal system published in leading sociology, general science, criminology, and public policy journals. A first strand of this quantitative research applies my insights on the role of data in producing knowledge to one of the most controversial and highly debated topics in the criminal legal system: predictive AI systems. Often, critiques of these tools center around the algorithm used to make decisions. My research shows quantitatively how the type of data used to train these predictive AI systems can shape their performance and fairness in previously unappreciated ways. I demonstrate that because these tools are trained on an administrative outcome (e.g., arrest for a crime) that is a proxy for the true outcome of interest (criminal offending), a problem that is known as “label bias” can frequently arise that makes these tool less accurate and more biased than they appear.

​

I have also worked on other research related to ethnic and racial disparities in the criminal legal system. I have written articles related to crime reporting and legal cynicism, the effect of increased school policing on test scores in New York City and changing patterns of arrest and offending for cohorts of Black and White youth. I am currently continuing this research with Felix Owusu, Sandra Smith, and Pari Bishen in a large-scale evaluation of the impact of electing progressive prosecutors in North Carolina and Massachusetts on racial disparities using millions of cases from administrative court records in both states.

​

Finally, I have a longstanding interest in the sociology of antisemitism, which I study using methods from the sociology of race and ethnicity and criminology research on victimization. This line of work was inspired by the observation that there is extensive heterogeneity in how Jews experience antisemitism, but that most available surveys use broad categories that make detecting intra-Jewish differences difficult. Working around these data challenges, I analyze how antisemitism is experienced in an embodied way, showing how whether Jewish Americans can be visually identified as Jewish shapes their experiences around their identity and attitudes towards antisemitism. My first article on this topic–based on 81 semi-structured interviews and national survey data from Pew–argues that being "visually identifiable" as Jewish shapes experiences of antisemitism, attitudes towards it, and positive experiences about one's identity.​

​

If you are interested in speaking about these projects or collaborating, please send me an email!