Can Larger Sample Sizes Result in Missed Opportunities?
Rousseau, B. and Ennis, D. M. (2020). IFPress, 23(3) 3-4.
The design of sensory and consumer tests relies on a combination of factors including statistical risks, methodological choices, and the size and nature of the respondent population. The research sample size is often the most discussed aspect among project team members because it requires finding a balance among representation, statistical power, timing, and costs.
Discrimination testing to investigate equivalence (or similarity) is used extensively by CPG companies, typically using the traditional concept of statistical power and a null hypothesis of no difference. However, if we rely on a nonsignificant difference to establish equivalence, then increasing the sample size may work against our objective. In this report, we illustrate how this can lead to missed opportunities and we describe a more direct and stable test of equivalence based on a pre-established consumer relevant action standard.