Determining an Appropriate Sample Size to Estimate IDP Shelter Area

Authors: Scott Basford*, Oak Ridge National Laboratory
Topics: Geographic Information Science and Systems, Remote Sensing, Middle East
Keywords: IDP settlements, internally-displaced persons, Syria, sample size, informal settlements, remote sensing
Session Type: Paper
Day: 4/14/2018
Start / End Time: 8:00 AM / 9:40 AM
Room: Grand Chenier, Sheraton, 5th Floor
Presentation File: No File Uploaded


The Population Density Tables (PDT) project at Oak Ridge National Laboratory provides building occupancy estimates for a variety of facility types at both national and sub-national scales. The project relies primarily on open source data collection to determine population density. Commonly, area is estimated by applying digital area measurement tools to satellite imagery, but this work is often tedious and redundant. This is particularly prevalent among refugee and internally-displaced persons (IDP) settlements, which generally feature a large number of relatively uniform structures. In such environments it is useful to determine a minimum threshold beyond which further measurements are unnecessary to ascertain average structure area. To accomplish this, we deploy a smart sampling tool in which two tests—that the mean changes no more than 5% with each additional sample and that the margin of error is no more than 10% of the mean—must be passed a minimum of 10 times before the sample measurements are considered to accurately represent the true average structure area. In this presentation, I describe the results of an investigation in which I assess whether a lower minimum-to-pass value (2 or 5, specifically) produces a comparable average to the current standard of 10. I do so through a case study of 54 Syrian IDP settlements situated in 9 clusters near the Turkish border. The results of the examination reveal that a lower minimum-to-pass value is sufficient for determining an accurate building area within Syrian IDP settlements.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login