Working Paper: NBER ID: w19543
Authors: Thomas F. Crossley; Joachim K. Winter
Abstract: When designing household surveys, including surveys that measure consumption expenditure, numerous choices need to be made. Which survey mode should be used? Do recall questions or diaries provide more reliable expenditure data? How should the concept of a household be defined? How should the length of the recall period, the level of aggregation of expenditure items, and the response format be chosen? How are responses affected by incentives? Can computer-assisted surveys be used to reduce or correct response error in real time? In this paper, we provide a selective review of the literature on these questions. We also suggest some promising directions for future research.
Keywords: household expenditure surveys; data quality; survey design
JEL Codes: C81; D12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| survey mode (C83) | response behavior (C92) |
| response behavior (C92) | quality of expenditure data (C82) |
| survey mode (C83) | comprehension of questions (C83) |
| survey mode (C83) | ease of recall (G41) |
| survey mode (C83) | confidentiality concerns (I13) |
| comprehension of questions (C83) | quality of expenditure data (C82) |
| ease of recall (G41) | quality of expenditure data (C82) |
| confidentiality concerns (I13) | quality of expenditure data (C82) |
| recall methods (D18) | underreporting due to forgetting and telescoping errors (C83) |
| diary methods (C90) | compliance issues and respondent fatigue (C83) |
| disaggregated expenditure categories (H76) | higher estimates of total expenditure (H59) |
| aggregated questions (E10) | systematic underreporting (P37) |
| different response formats (C83) | systematic biases in reporting (C83) |