Tuesday, October 2, 2012

Data Collection Methods

STICERD Fieldwork in Development Economics, managed by Markus Goldstein, provides resources useful for doing fieldworks.

Martinelli and Parker (2007) "Deception and Misreporting in a Social Program" provide caution against using self-reporting information on the ownership of status goods (goods that many other people own).

De Mel, McKenzie, Woodruff (2008) provide guidance for measuring microenterprise profits.

Svensson (2003) provides the way to solicit information on bribe payments. See section III.

Delavande, Gine, and McKenzie (2010) discuss how to solicit subjective expectations (such as expected return to schooling) from people in developing countries.

Zwane et al (2011) show that being asked about their behavior does change their behavior subsequently. Researchers should consider the impact of surveys on the outcomes of interest when they design the questionnaire.

Mckenzie (2012) argues that for outcomes serially uncorrelated (such as profits), skipping the baseline survey and conducting multiple follow-up surveys can help increase the statistical power.

Development Impact, the blog by four World Bank economists working on randomized control trials, often reports the studies in which the data collection methods are randomly assigned.

  • April 26, 2011: when interviewing a household member privately is good and when it is bad
  • April 27, 2011: what method of consumption expenditure survey is the best
  • October 11, 2011: how labor supply should be measured for farmers
  • October 18, 2011: how access to financial services (bank accounts, insurance, etc.) should be measured.
Demombynes (2012) talks about the need to make your data publicly available.

No comments: