Department of Labor Logo United States Department of Labor
Dot gov

The .gov means it's official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Bureau of Labor Statistics > Office of Survey Methods and Research > Publications > Browse Research Papers

Statistical Examination of Rounding Tendencies in the Consumer Expenditure Interview Survey

Taylor Wilson and Safia Abdirizak

Abstract

Based on the results from the 2015 Consumer Expenditure Field Staff Survey Analysis Report, 98.4% of Field Representatives stated that record use improves the accuracy of the interview. The primary purpose of this project is to conduct research regarding the use of records as it relates to the rounding effect of recall interviews. This paper tests two hypotheses, [1] the use of records reduces the rounding effect and, as a result, increases data accuracy and quality and [2] the rounded expenditure amounts and non-rounded expenditure are significantly different. The hypotheses are important in aiding large survey programs in understanding how respondent rounding will affect the underlying data quality of the survey responses. We employ a unique approach to isolate values which are most likely heaped-a coarse data property in which the respondents tend to converge their answers on 'round numbers.' This method examines the frequency of numbers represented in a domain of discrete values and examines the likelihood of observing any given value versus the rest of the values in that domain. The values which are statistically over represented by pre-defined threshold are heaped and thus more likely to be rounded. By identifying those values which have the highest probability of being a rounded value, we make an assumption that those values have actually been rounded. The findings will be of interest to survey methodologists and practitioners working in large scale survey operations with recall survey components and specific data quality goals.