The accuracy and precision of consumer diary panel information

Date of publication: September 1, 1990

Author: John B. Frey

Abstract:

The Du Pont Company is one of the world's largest producers of Textile Fibers for Apparel End Uses. Major products include nylon, DACRON® polyester, ORLON® acrylic, and LYCRA® spandex. Its Textile Fibers are manufactured and marketed around the world and go into a large variety of apparel as well as other end uses. Among its many sources of information, Du Pont relies heavily on information from a leading soft goods diary panel company in the U. S. This information is focused on what consumers buy, where they buy, and how much they pay. The consumer diary panel data is compared with producer data and other industry data to help assess inventory fluctuations and shifts in the supply/demand balance. Consumer diary panel information is subject to a variety of errors. Because the diary is filled out by respondents in the privacy of their homes, there is no good way to assure the validity or reliability of the information. This paper reports on a study done to assess the accuracy and precision of the information. Two approaches were taken. The first of these was a "Monte Carlo" simulation of the diary process. Relying in part on drawing random numbers from specified probability distributions, the process of making purchases, recording those purchases, submitting diaries, and calculating projections was simulated on a computer. The second approach was a statistical analysis of comparative data from two competing diary panels on 47 apparel categories. Findings suggest that sampling error is only a small part of the total error and that the method used to disaggregate and project reported transaction data adds to the overall error. Furthermore, errors in period-to-period data do not "cancel out" to the extent we previously believed. Very large differences in precision exist between apparel categories. While diary information has value, it cannot be relied upon to assess overall levels of or trends in purchases with the degree of precision normally required and must be supplemented with other information for planning purposes.

  • PDF
  • This could also be of interest