The Impact of Decentralized Data Entry on the Quality of Household Survey Data in Developing Countries: Evidence from a Randomized Experiment in Vietnam |
| |
Authors: | Glewwe, Paul Dang, Hai-Anh Hoang |
| |
Affiliation: | Paul Glewwe (corresponding author) is a professor in the Department of Applied Economics at the University of Minnesota Hai-Anh Hoang Dang is a consultant in the Policy Research Group at the World Bank; his email address is hdang{at}worldbank.org |
| |
Abstract: | Computers were provided to randomly selected districts participatingin a household survey in Vietnam to assess the impact on dataquality of entering data within a day or two of completing theinterview rather than several weeks later in the provincialcapital. Provision of computers had no significant effect onthe observed distribution of household expenditures and thusno effect on measured poverty. Provision of computers reducedthe mean number of errors per household by 5–23 percent,depending on the type of error. Given the already low rate oferrors in the survey, however, the goal of increasing the precisionof the estimated mean of a typical variable can be achievedat a much lower cost by slightly increasing the sample size.Provision of additional computers did substantially reduce thetime interviewers spent adding up and checking the data in thefield, with the value of the time saved close to the cost ofpurchasing desktop computers. |
| |
Keywords: | |
本文献已被 Oxford 等数据库收录! |
|