Presents guidelines for users of the simple Yes/No signal detection experiment who want to design their experiments to ensure a given precision for estimators of d. Additionally, simulations of small-scale Yes/No and rating-method experiments are employed to investigate when the use of asymptotic results is justified. The way in which the pay-off matrix for the Yes/No experiment may be manipulated to optimize precision is demonstrated. Large experiments may be needed to provide precise unbiased estimation of d, with sampling distributions that are not appreciably skewed.
|Number of pages||7|
|Journal||British Journal of Mathematical and Statistical Psychology|
|Publication status||Published - May 1990|
- experimental design
- signal detection