Recent stories in The New York Times, The Washington Post, and other media outlets have revealed new information about the government’s actions concerning prisoners at Guantánamo Bay. My thesis advisor, Kim Lane Scheppele, thought that it would be helpful for me to share the findings of my senior thesis on this topic, given this newly emerging content.

I was inspired to write this thesis when I realized that, despite the publicly available information on individual detainees, no one had systematically analyzed how various individual characteristics or factors correlate with length of detention at Guantánamo Bay Detention Camp. In my thesis, I attempt to do so, considering factors such as potential threat, age, citizenship, location of capture, and identity of initial captor.

My main findings include:

• A detainee’s citizenship is the most significant factor in determining his length of detention at Guantánamo Bay.
• My analysis suggests that a detainee’s citizenship matters because it reflects his home country’s ability (or inability) to monitor him after leaving Guantánamo; a lower capacity to monitor translates to longer detention duration.
• The identity of a detainee’s initial captor also correlates with length of detention, though less substantially than the detainee’s citizenship. Individuals captured by Pakistani officials are detained longest, while those captured by US and Coalition Forces are detained shortest.
• A detainee’s level of threat correlates only very slightly, if at all, with length of detention.

Role of Citizenship, Rather than Threat

Potential threat correlates almost not at all with length of detention, yet citizenship of the detainee correlates significantly.

To determine a detainee’s “threat level,” I use threat measurements based on a 2007 study published by the Combating Terrorism Center at West Point. I give a detainee one point each if there was any evidence that he: 1) attended a training camp; 2) was a fighter for a terrorist group or; 3) engaged in hostilities against the United States or its allies. On the graph below, each detainee is assigned a “threat score” from 0-3 based on evidence in publicly available Combat Status Review Tribunal (CSRT) documents.

As the graph shows, length of detention does not increase based on threat score. If it did, one would see a diagonal line that slopes upward as threat score increases. Rather, length of detention varies with the detainee’s country of citizenship. When one considers three of the largest groups represented at Guantánamo, Yemenis are detained nearly five years longer than Afghanis, with Saudis between these extremes. These differences remain even when the detainees from different countries have the same threat scores.

Figure 1: N=120 for Afghanistan; N=112 for Saudi Arabia; N=100 for Yemen
This graph shows the correlation between length of detention and threat level for citizens of Afghanistan, Saudi Arabia, and Yemen. Note that the correlation is only slightly positive (the best-fit lines slant slightly upward). However, for a given threat level, the clusters of Afghani detainees are detained for significantly less time than comparable Yemeni detainees. Saudi detainees are held, on average, for a length of time between the other two. For Afghanistan, R²=.004, for Saudi Arabia, R²=.010, for Yemen, R²=.011.

In my thesis, I show that the ability of a detainee’s home nation to make security commitments to the United States explains why citizenship affects length of detention. To test this explanation, I use data from a Brookings Institution study by Susan Rice, who was subsequently named US Ambassador to the United Nations. Rice’s report assessed the effectiveness of governments on a scale of 0-10. I apply these rankings to the Guantánamo context, assigning each detainee the score corresponding to his country of citizenship. I find that the higher the government effectiveness score for a detainee’s country of citizenship, the earlier detainees from that country are transferred from Guantánamo. This account is reasonable, given that a government’s effectiveness is representative of its ability to make security commitments to the United States that it can incarcerate or monitor detainees transferred into its custody. I also apply a human rights ranking system included in the Brookings study to test whether better human rights respecters get their citizens back sooner. However, I do not find a strong relationship between a detainee’s length of detention and the human rights score of his home country.

The Role of Initial Captor

I also find that the identity of a detainee’s initial captor is associated with length of detention. Detainees captured by authorities in Pakistan are detained longer than detainees captured by US and Coalition Forces, or by the Northern Alliance in Afghanistan. This result holds even when controlling for other factors such as a detainee’s citizenship, threat score, age, and location of capture. Without the use of classified information, it is difficult to assess why we see this relationship. However, given this new finding, we should explore this area further.

Conclusion

In my thesis, I sought to conduct the first mathematically rigorous, large-scale analysis of factors influencing length of detention at Guantánamo Bay. When incorporating data from 516 CSRT reports, a West Point study, and a Brookings Institution report (as well as a variety of other sources), I found that citizenship was the most important factor, followed by identity of initial captor. A detainee’s threat score correlates little, if at all, with length of detention. 

For additional information, please contact me directly at sdorison@princeton.edu. Professor Kim Lane Scheppele (kimlane@princeton.edu) is also available to discuss this study.