News: News Archives
http://www.aaas.org//news/releases/2003/0623hr-methods.shtml
Human Rights Data Analysis System
Includes Ways of Ensuring Accuracy
Since there are so many sources from which data can be collected (legal records, radio shows, newspapers), the HRDC came up with a "controlled vocabulary"a set of parameters that can turn qualitative information (such as a narrative) into a calculable data set, according to Romesh Silva. (See description of violation category, "Rape.")
By processing qualitative information using a controlled vocabulary, a violation like rape, or other violations such as sexual abuse or genital abuse (which also appear in the controlled vocabulary list) are transformed into well-defined statistical data.
When such information-processing techniques are applied to large amounts of qualitative information, the result is the construction of large-scale datasets (which detail the complex events and relationships between different violations and their associated victims and perpetrators). However, the reliability of this process must be continuously monitored and measured in order to ensure high quality data.
Once the data is pooledbringing together information from multiple organizations-researchers can estimate the total number of violations, including those never documented by any group. The term for this type of analysis is "multiple systems estimation."
Silva describes the method as akin to a set of Venn diagrams: imagine six circles (representing the six NGOs' sets of violations data). If the data is the same, many of the circles overlap and the number of actual violations will be estimated as smaller. However, if there is only slight overlapping, then the violations will be estimated to be higher.
The multiple systems estimation is also a de facto check and balance for the NGOs. For example, if one group was to fraudulently create data, there would be no overlap whatsoever with the five other groups, and thus send a signal that something was awry.
Carol Hoy
23 June 2003
Return to main article, "Turning Tales of Trauma into Hard Data to Draw Accurate Portrait of Conflict."
