Last summer at a research workshop in Tanzania, my field coordinators and PIs were big proponents of using personal digital assistants (PDAs) instead of the old fashioned paper interviews. As I prepare to collect baseline data using a pretty extensive household survey (trying hard to keep it under two hours), here's some solid empirical evidence from the Journal of Development Economics in support of them.
Evidence from a randomized experiment:
We find that PAPI [Pen and paper interviewing] data contain a large number of errors, which can be avoided in CAPI [computer-assisted personal interviewing]. Error counts are not randomly distributed across the sample, but are correlated with household characteristics, potentially introducing sample bias if dubious observations need to be dropped.
Evidence from a randomized experiment:
We find that PAPI [Pen and paper interviewing] data contain a large number of errors, which can be avoided in CAPI [computer-assisted personal interviewing]. Error counts are not randomly distributed across the sample, but are correlated with household characteristics, potentially introducing sample bias if dubious observations need to be dropped.
RSS Feed