Cast a critical eye.

Until the 1990 US census there are no real yardsticks for evaluating the reliability of small area estimates for projections yet the reliability of these numbers is vital to those who use such data in making important business investment decisions. The International Council of Shopping Centers (ICSC) analyzed the small area estimates and projections offered by 6 private data companies for 3 metropolitan areas--Baltimore Detroit and Phoenix. The ICSI requested data on 1) population and numbers of households 2) current average household size 3) current number and percentage of households in broad income groups and 4) average household income. The coefficient of variation is used to evaluate the estimates and projections of population and households; Cramers statistic is used to evaluate household income distributions. This analysis demonstrates that there is relatively little difference in the population and household totals provided by the data companies at the metropolitan and county levels. Also within stable market areas the different among the providers are not great for standard units of geography even as small as census tracts. Buyers can be confident that regardless of the data company they use the numbers would not be much different. In rapidly changing areas on the other hand a buyer would be wise to verify small area data through field investigation and interviews with local sources such as public planning agencies. Because of the difficulties in defining zip code geography and allocating census tract information to zip codes buyers should be cautious with zip code data when accuracy is critical. Likewise businessess targeting specific income groups should be aware that there are wide variations in updated come data at the local level.