A Household-Level Pairwise Comparison of Targeted Marketing Data and Self-Reported Survey Data

摘要

This research conducts a validation test of targetedmarketing data by comparing it at the household level to self-reported survey data. The pairwise comparison was limited to the following demo graphic and socioeconomic variables: age, educational attainment, ethnicity, gender, household income, marital status, number of adults, number of children in the household, and tenure. The self-reported data were collected with a mailed stated preference (SP) survey regarding transit ridership in four neighborhoods of Atlanta that consist of many hard-to-reach and hidden popu lations. A rate of accuracy was calculated using a percent of correct matches between the two datasets for each variable. Chi-squared tests where also completed using both the targeted mar keting and survey data. The findings suggest that targeted marketing data match self-reported data for neighborhoods of hard-to-reach or hidden populations at rates ranging from 17.4% to 94.5% depending on the variable. The self-reported data show that incorrect targeted marketing data ran domly occur across all populations in relation to age, gender, household income, number of adults in the household, and tenure. It does not randomly occur across ethnicity or marital status groups. Educational attainment and the number of children in the household were not testable with regards to randomness across groups. Further research should be conducted to quantify the accuracy of targeted marketing data at the household-level for population groups that are more easily surveyed or documented.