A Tale of Two Surveys
Internet surveys may never replace in-person surveys as a gold standard, but they remain important tools for rapid, remote, and low-cost data collection. The West Bank and Gaza Poverty and Equity team had a unique opportunity to compare a Facebook survey with an in-person survey covering conflict exposure and potentially associated socioeconomic mental health outcomes over a similar time period. It is reasonable to expect that the estimates from internet surveys and in-person surveys would differ. In this case, the Facebook survey estimates more severe outcomes (e.g., higher exposure to conflict and worse mental health) than its in-person counterpart for most topics and populations. Multiple mechanisms may have contributed to this difference in estimates, including overrepresentation in the Facebook sample of respondents who were interested in the survey topics, reduced sensitivity bias in the context of a self-administered online questionnaire, and reporting more severe outcomes than personally experienced to encourage resource flows to perceived needs. Estimated outcomes tend to be more similar for people in Gaza, possibly because of greater homogeneity in socioeconomic experiences and exposure to violent conflict and broader interest in a survey on the effects of the May 2021 violence. The main results are robust to different ways of controlling for observable characteristics; neither alternative weights nor sample restrictions erase the systematic differences between the surveys.
| Main Authors: | , , |
|---|---|
| Format: | Working Paper biblioteca |
| Language: | English English |
| Published: |
World Bank, Washington, DC
2023-07-12
|
| Subjects: | SURVEY METHODOLOGY, INTERNET SURVEY, SELECTION BIAS, SENSITIVITY BIAS, CONFLICT, MENTAL HEALTH, CONFLICT TRAUMA, |
| Online Access: | http://documents.worldbank.org/curated/en/099749306212335845/IDU06ae9560f0dd1004b440b9c90c0ee1da42011 https://openknowledge.worldbank.org/handle/10986/39987 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|