Bootstrap Investigation of Respondent Sample Size for Environmental Preference

Although environmental researchers use many different sizes of respondent samples, few environmental studies provide statistical rationales for selecting those sizes. This article utilizes bootstrap analysis to calculate split-block correlations of preferences of randomly selected respondents to a variety of environmental stimuli. Two analytic methods were used, raw score ratings and comparative choice. Analysis showed split-block correlations of .90 and higher could be achieved with 25 to 30 respondents with rating protocols and 10 to 15 respondents with comparative-choice protocols.