Examining the effects of extra legal factors on immigration decision making in asylum cases
MetadataShow full item record
Given the ever-present debate over immigration law and policies since the inception of our nation, combined with recent controversies over the Arizona immigration laws and the executive order authorizing Deferred Action for Childhood Arrivals (DACA), it is without a doubt that immigration law is an important area to study. The entire immigration legal system, embodied within the Executive Office for Immigration Review, is completely under the control of the Department of Justice and hence, the Attorney General of the United States. The Immigration Court and the Board of Immigration Appeals (BIA) are the two systems of interest within this dissertation. The two administrative agencies are given much deference in their decision-making process. This dissertation seeks to study whether Immigration Judges and BIA members base the decision-making process on strictly legal principles, and not extra-legal factors. For this dissertation, I only study asylum case outcomes. At the Immigration Judge level, I coded important factors of each Immigration Judge and estimated an OLS regression of which of these factors influence decision-making. At the BIA level, I conducted a two-level factor test when coding my variables. At the first level, I coded for BIA member factors. At the second level, I coded for respondent factors. The quantitative study for the BIA estimates a Multinomial Logit Regression Model to determine which factors were important in predicting BIA judge votes in these asylum cases. The results show that both Immigration Judges and BIA member are affected by extralegal factors. This means that because the process is mostly unchecked, these "judges" enjoy a wide latitude of discretion in the decision-making process, especially in asylum cases, where discretion plays a large role. The findings confirm that these decision-makers do not strictly adhere to a legal reading of the law, but that they are affected by their own characteristics, as well as characteristics of the respondent before them.