Issues in Proxy Reporting in the Consumer Expenditure Survey

Proxy reporting about expenditures Although proxy reporting is ubiquitous in surveys – for example in completing household rosters, reporting labor force participation, and detailing the health status of spouses and children – the use of proxy reporting for detailed reporting of expenditures in the Consumer Expenditure (CE) Survey entails challenges that differ from those for other topics and other surveys. Moore’s (1988) review of studies about proxy reporting concluded that the studies available at that time did not point to a clear advantage for self reporters, but that the available evidence was not definitive. However, even though differences between proxy reports and self reports about participation in the labor market or major illnesses may have been modest (see Moore 1988), these findings cannot be extrapolated to reporting about expenditures by a consumer unit. More recent research suggests that proxy reports are most likely to be comparable to those given by the person if the proxy has an opportunity to have direct knowledge (either by participation or conversation) of the event, and the opportunities for such knowledge can be expected to be greater if the event reported about is engaged in regularly over a long period of time and is salient or important for other reasons, such as emotional or financial reasons (see, e.g., Sudman et al. 1994 ; Schwarz and Wellens 1997; Kojetin and Miller 1993; Tucker and Miller 1993). Clearly, labor force participation and major illnesses may share some of these features in a way that many expenditures do not. The somewhat limited literature that addresses issues relevant to assessing proxy reporting about expenditures has recently been reviewed by Mathiowetz (2010) and earlier by Kojetin and Jerstad (1997). There is evidence that proxy reporting about expenditures might be of poor quality. Kojetin and Jerstad (1997) compared answers of self and proxy reporters about expenditures in a task simpler than that in the CE Interview Survey. They concluded that proxy reports were incomplete, even when considering reports at the category level, and more detailed reports were of even poorer quality. The number of