ABSTRACT

This chapter reviews some of the results from the panel survey portion of the Springville Project. In social science research, panel surveys typically comprise cross-sectional samples of informants who are re-interviewed for the same variables at different points in time. Using data from the Springville panel survey, the chapter demonstrates the impact of both the Small N Effect and the Gap Effect on results from panel surveys in sociolinguistics. The chapter suggests that researchers should use extreme caution in reaching conclusions from data with small Ns. The Springville panel study data provides further confirmation that low-token interviews are unreliable measures of an individual's use of a linguistic feature, which is refered to as the Small N Effect. Linguists who are actively planning a panel survey should consider the quantity of data and the frequency of its collection very carefully in the study design.