The LSMS program has developed a Computer Assisted Personal Interviewing experiment in order to: (i) improve the measurement of core indicators in LSMS surveys; and (ii) improve the quality of the data that is generated by improving its accuracy, relevance, and timeliness. This last point encompasses taking advantage of new technologies.
Computer Assisted Personal Interviewing systems (CAPI) that allow direct data entry (no paper questionnaire) provide a tool for checking data as it is entered during the interview so that errors can be rectified immediately. Research in the U.S has shown that, compared to standard, post-field work data entry systems, CAPI systems lead to a substantial reduction in data errors.
CAPI in Tanzania
A randomized survey experiment specifically designed to compare pen-and-paper (PAPI) to computer-assisted personal interviewing (CAPI) has been done. An existing full CAPI survey of 1,200 households on the island of Pemba in Zanzibar was augmented with 320 CAPI interviews with disabled consistency checks and 320 PAPI interviews. The analyses found an average of 10 inconsistencies per survey in PAPI, and virtually non in CAPI; a reduction that is primarily due to CAPI's automated routing system. The results also suggest that CAPI significantly reduces the number of outliers in food consumption data, likely due to the availability of checks, clarifying pictures and pre-defined unambiguous item specific unit lists. the study found further evidence that interview time, excluding time taken to run consistency checks, is cut by12% and there is no impact on respondents' perceptions of the interview.