Tourism Surveys: Sample Size,Accuracy,Reliability,and Acceptable Error

Abstract

Misunderstanding of terminology for expressing acceptable error in values estimated (e.g., total expenditure) using tourism surveys can cause confusion and inappropriate use of estimates. There is confusion about estimates being accurate, biased, being within a certain percentage, for example, 5%, and being reproducible with an acceptable amount of error. This article clarifies such matters. The need for standards to assess accuracy is introduced. Statistical concepts that relate to understanding acceptable error are introduced using examples and graphics based on simulation. Variability in responses is documented using results from two large national surveys. Sample size to achieve particular levels of acceptable error for means, totals, and change/difference are determined graphically. Sample sizes are given for some values of acceptable error that estimate users might want. Implications of large sample sizes being needed to have an acceptable amount of error, appropriate use of small samples, and options for achieving acceptable error are introduced.