|
Abstract:
|
This paper describes the sample design used at the Guatemalan National Police Archive (GNPA). The Archive contains millions of documents, which were initially found mixed together and in poor physical condition. Given the Archive size and lack of a traditional sample frame, we opted for a multi-stage random PPS sample using the Archive's topography for stages 1 and 2. For stages 3 and 4, frames were created on location. The sampling faced several challenges, including movement of the documents as they were being restored and organized, and uncertain resource availability. To manage these difficulties we drew iterative sample waves. After rounds of evaluation, we modified the sampling to reduce one stage, making the sampling more efficient. Over 2 years of sampling, we have selected 20,000 documents. Next, we may use adaptive sampling to search for documents of interest.
|
