Background: The U.N. Convention against Torture has been in force for more than 20 years, but torture continues in over 150 countries (Amnesty International, 2009). In the U.S., torture's repercussions--chronic pain, anxiety, depression, PTSD—are felt chiefly by refugees (Quiroga & Jaranson, 2005). Experts estimate the prevalence of torture survivorship among U.S. refugees at 60%, suggesting a U.S. population of over 500,000, but there is no reliable national data (Mollica, 2004). Lack of data is a problem for agencies and professionals who serve refugees or seek funding for torture-treatment services. This study pilots the use of human rights data to predict the resettlement of African refugees in the U.S. The central hypothesis is that government use of torture predicts an outflow of refugees who eventually settle in the U.S., and that use of torture continues to affect the exodus of refugees up to two decades later.
Methods: This study used the Cingranelli-Richards (CIRI) dataset, a longitudinal repository for international human rights information (Cingranelli & Richards, 2010), and the sample comprised 47 African countries for which CIRI data was available (1990 to 2006). CIRI categorizes countries annually according to frequency of torture: never, occasionally, or frequently, coded in this study as 1, 2, and 3. Additionally, CIRI codes when torture occurs during political chaos (coded here as 4). This study used U.S. Office of Refugee Resettlement data for information on the 153,350 African refugees admitted in 2000-2009. To highlight changes over time, torture and refugee variables were summed into time-defined, continuous variables: Torture 1990-94, Torture 1995-99, and Torture 2000-06, Refugees 2000-04, Refugees 2005-09, and Refugees Total. Bivariate correlations and linear regressions were used to explore the effects of allegations of torture on the resettlement of refugees in the U.S.
Findings: Bivariate correlations between torture and Refugees Total were significant across all time periods (r= .475, .436, .436). When Refugees Total was regressed on the three torture variables, the model was significant (F=5.729 p<.01) and torture explained almost 29% of the observed variance in refugees over the 17-year study period. Thus, every unit increase in torture in a given country, 1990-1994, predicted 1561 refugees seeking safety in the U.S. in 2000-09. These findings were echoed in successive regression models as torture was shown to predict U.S. refugee resettlement in 2000-04 (r2=.268) and 2005-09 (r2=.280).
Implications: Analysis indicates a significant predictive relationship—29% overall—between allegations of torture and African refugees seeking resettlement in the U.S. Furthermore, this analysis shows that past torture (1990-94) remains a significant predictor (r2=.155) of refugee resettlement more than a decade later in 2005-09. This time gap between the torture and the refugees' arrival may be an important finding for social workers who need to plan and budget for refugee services five, 10 and even 20 years in the future.
