Background

The U.S. Government Accountability Office (GAO) issued report GAO-08-940, entitled U.S. Asylum System: Significant Variation Existed in Asylum Outcomes across Immigration Courts and Judges. In the report, the GAO recommended that the Executive Office for Immigration Review (EOIR) use the GAO’s methods and findings to examine cost effective options for obtaining statistical information on immigration judges’ asylum decisions to identify judges in need of training and supervision. The GAO noted that current EOIR asylum outcome studies did not take into account available data on the characteristics of asylum seekers or immigration judges when compiling asylum decision data. The report concluded that the statistical methods applied by GAO in its analysis provided a more complete, accurate, and useful picture of the asylum rulings. The GAO recommended that EOIR duplicate its multivariate models for EOIR’s internal use, which would provide a resource to ensure that immigration judges are effectively supervised.

Project Summary

Fentress was tasked to analyze recent data on grants and denials of asylum seekers’ applications, and to compare the results obtained by analyzing recent data identified in the GAO report. Fentress was provided with over 3.5 million asylum cases where decisions were made between October 1994 and April 2007 that had to be meticulously queried to identify the specific records used for analysis in the original GAO report. Once the correct records were selected for analysis, Fentress performed the necessary logistic regression to successfully replicate the results found within the original GAO report. Fentress was then provided with a new set of data from asylum cases where decisions were made between May 2007 and September 2011. The data mining efforts were replicated on the new data prior to performing logistic regression. Fentress created a report which compared the two sets of results regarding the grant and denial rates among immigration courts and immigration judges throughout the country. EOIR was provided with step-by-step instructions on how to replicate the statistical analysis in the future.

Results

This project provided EOIR with insight into whether or not immigration courts and immigration judges throughout the country were displaying bias in their asylum decisions. The step-by-step instructions created by Fentress could be used by EOIR staff as part of a periodic review of asylum grant and denial rates. As necessary, extra supervision and/or training could be provided to those judges whose asylum decisions appear to reflect statistical bias.