Abstract:
|
In complex surveys spanning diverse populations, the interviewed sample can generate data with a wide range of quality. Some respondents forget information, provide information in incorrect places, misunderstand questions, or even refuse to provide data at all, while others will supply answers that almost perfectly represent their situations. In the context of a highly-detailed household financial survey, we investigate respondent-level determinants of data quality. Do specific demographic groups provide higher quality data than others? Are any respondent traits, including financial literacy, particularly indicative of higher quality data? Results suggest that greater respondent engagement, interest, education, or confidence in financial topics could generate increased data quality for financial questions.
|