A look at the method behind the Post Afghan poll

A total of 209 professional interviewers were assigned to this public opinion survey project, conducted for The Washington Post and its partners by the Afghan Center for Socio-Economic and Opinion Research. The Post's Director of Polling, Jon Cohen, reports.
By Jon Cohen
Washington Post Staff Writer
Monday, December 6, 2010; 6:06 AM

Conducting a public opinion survey in Afghanistan is a big challenge. There is no current census of the population, some places are inaccessible for security reasons, and most adults have no formal education. But it is possible to collect quality data.

Here's how.

Random selection is the key component in any poll, and interviews for this survey were conducted in person among adults selected at random based on where they live.

The first step in the sampling process used here by the Afghan Center for Socio-Economic and Opinion Research in Kabul was to select random locations in each of the country's 34 provinces. The center randomly selected districts within provinces and villages or neighborhoods within districts.

Residences were then selected by the "random route, random interval" method, in which interviewers chose households and adults within households based on a predetermined formula. Male respondents were interviewed only by male interviewers, female respondents only by female interviewers. The final sample included 227 sampling points that were adjusted by population of province and sex by region, according to estimates from the Afghan Central Statistics Office.

Of the country's 398 districts, 59 were inaccessible for security reasons during the field period of Oct. 29 to Nov. 13. Female interviewers could not work in an additional 75, including all of Paktika and Uruzgan provinces. Despite these limitations, the poll represents 85 percent of the national population.

The survey had a contact rate of 86 percent and a cooperation rate of 94 percent, for a response rate of 81 percent.

© 2010 The Washington Post Company