I recently came across an interesting article from the Journal of Refugee Studies, "A Rare Examination of Typically Unobservable Factors in US Asylum Decisions," which analyzes data from 81 asylum cases to determine the "unobservable factors" that influenced the decisions in those cases.  The authors had access to cases litigated by the Human Rights Initiative of North Texas ("HRI") between 1998 and 2005.  As the authors explain, the small number of cases and the selection bias in the samples (HRI only takes cases deemed to have a good prospect of success) makes their findings suggestive only.  Nevertheless, the study points to some interesting possibilities concerning how decisions are made. 

The authors refer to prior studies, which correlate asylum grants with factors other than a "well-founded fear" of persecution, for example, "the judge's prior work experience (particularly prior INS experience), and legal representation strongly influence asylum outcomes."  Also-

While some legal scholars and human rights activists might expect that human rights conditions and evidence of credible fear of persecution would be the most important factors in the determination of whether to prevent a particular asylum seeker from being returned to a situation that would threaten their life or physical integrity, these empirical studies suggest that outcomes are more likely to be based on economic and security concerns of the state than the merit of the claim.

Unlike most prior studies, the authors here, Linda Camp Keith and Jennifer S. Holmes, had access to the asylum seekers' files, and could look at many factors-gender, marital status, education level, religion-that are not normally available.  Some of the findings are quite surprising, and are examined below: 


The authors write: "we expect that a female will be less likely to receive asylum due to possible cultural biases of the decision makers concerning women as viable threats to government and due to possible cultural differences that affect women's credibility."  In fact, the authors found that within the cases studied, being female decreased the chances of an asylum grant by a statistically significant amount.  Indeed, of all the factors studied (except possibly religion), gender was the most significant "unobservable factor" that determined the outcome in an asylum seeker's case. 


The authors expected that asylum seekers with higher levels of education might be viewed as economic migrants.  They found "applicants with a primary or middle school education had greater success than higher educated applicants."  However, they noted that "the sample size is small in the primary and middle categories" (although the results do seem to have some statistical significance).  They concluded, "Overall, there does not seem to be much variation in the grant rate due to educational attainment."  People who speak English, however, are statistically more likely to succeed in their cases. 

In my own experience, I have always believed that more educated applicants are more likely to win asylum.  More educated people are better able to understand the system, they are more likely to articulate their stories consistently (which decision makers rely on to judge credibility), it is easier for well-educated decision makers to relate to them, and they are considered less likely to become a burden on our society.  I imagine that these positive factors outweigh any negative perception that well-educated asylum seekers are economic migrants.

What are the odds that this single Christian female would qualify for asylum?



The authors distinguished between asylum seekers with Judeo-Christian religions and those with non-Judeo-Christian religions.  They expected that asylum seekers with non-Judeo-Christian religions would be more likely to win asylum.  In fact, the authors found that having a non-Judeo-Christian religion was the most influential "unobservable factor" in gaining asylum.  Their results in this regard were considered statistically significant.

This is the one observation that seems to me flawed.  The authors' cases come from an NGO in Texas, and so many of their Judeo-Christian asylum seekers likely come from Mexico, Central America, and South America.  People from these countries are very unlikely to gain asylum in the U.S.  Asylum seekers with non-Judeo-Christian religions, on the other hand, likely come from other regions of the world-regions where it is more likely that they will be granted asylum.  For example, many of my cllients are asylum seekers from Ethiopia and most are Christian.  They have a very high likelihood of success in their cases (usually based on political persecution).  Thus, had the authors used data from an NGO in my area (where we have many Christian asylum seekers), their results concerning religion would likely have been different.

Marital Status

A 2000 study suggested that single people were less likely to gain asylum in the United States, presumably because decision makers view them as likely economic migrants.  However, the authors of the current study found that being married significantly decreases the odds of an asylum grant.  I've never noticed any difference in the grant rates for my married vs. single clients.  However, whenever an asylum applicant lists numerous young children on their application, it makes me worry that an adjudicator will be more hesitant to grant, knowing that the grantee's entire family will be "following to join" him in the United States.

Some Thoughts

As the authors point out, their sample size is small, and the results are only suggestive.  Nevertheless, it seems safe to say that "unobservable factors"-or at least factors that are not related to the legal requirements for asylum-do influence decisions in asylum cases.  I imagine the same is true in criminal cases and civil cases.  Not that this makes the situation any better, but the fact is, such "improper" influences are difficult to eliminate in any type of case.  More study is clearly needed.  If Immigration Judges and Asylum Officers can be made aware of the biases that influence their decisions, perhaps that will be a first step towards reducing those biases.