Single vs. Double Source Data for TV Program Selection
比较了单一来源和双重来源两种数据方法在基于购买行为选择电视节目时的效果，发现双重来源方法在缺乏单一来源数据时是可行的替代方案。
Two methods were used to select TV programs based on product purchasing behavior—a single-source and a double-source method. The single-source method—defined as the electronic measurement of TV exposure and purchase behavior for the same household—relied on Nielsen’s Television Index (NTI). In the absence of a single-source data base, NTI was used for this purpose since it collects both TV exposure data and purchase information for a small number of product categories. Double-source data is defined as the development of a demographic profile from one data source and the application of this profile to TV exposure data from another. Double-source data was obtained by using MRI data to develop demographic profiles of heavy purchasers/owners for four product categories that are also collected by NTI. NTI was also used as the source of TV exposure in the double-source method by determining TV exposure for the demographic groups for each of the four product categories. Single-source data (solely from NTI without reference to demographics) and double-source data (from MRI and NTI utilizing demographic profiles) were then correlated across 54 programs. Correlations for the four product categories were highly significant, ranging from .54 to .92. In the absence of a single-source data base, the findings suggest that double-source methods may be a viable means of selecting TV programs based on purchase behavior.