|
Abstract:
|
The Consumer Expenditure Interview Survey has a three-month recall period. Over the years, researchers have noticed that the level of expenditures is highest for the most recent recall month, lower for the middle month, and lowest for the most distant month. This pattern is generally believed to be caused by respondents gradually forgetting about their expenditures over time, although other theories have been put forth as well. To improve the accuracy of the reported expenditures, research was done to identify the cause of this recall phenomenon, measure its magnitude, and possibly recommend steps to solving it. This paper presents a model for quantifying the magnitude of various sources of the observed recall effect, along with a statistical analysis of the Consumer Expenditure Survey data. Findings are presented that show the estimated forgotten amount and its magnitude relative to the model's predicted, "true" expenditure.
|
