Db. Allison et al., REFLECTIONS ON VISUAL INSPECTION, RESPONSE GUIDED EXPERIMENTATION, AND TYPE-I ERROR RATE IN SINGLE-CASE DESIGNS, The Journal of experimental education, 61(1), 1992, pp. 45-51
The use of single-case research designs for the evaluation of experime
ntal outcomes is a distinguishing characteristic of applied behavior a
nalysis. Behavior analysts traditionally rely on visual inspection of
graphed data in lieu of statistical analyses. Proponents of visual ins
pection claim lower Type I error rates relative to statistical analysi
s. However, behavior analysts also employ response guided experimentat
ion (i.e., allowing accumulating data to guide the further course of t
he experiment). Based on data published by Matyas and Greenwood (1990)
, we argue that when visual inspection is combined with response guide
d experimentation, Type I error rates could easily be as high as 25%.
Tables of error rates as a function of alphas, autocorrelation, and nu
mber of looks are presented. We suggest caution in interpreting conclu
sions based on single case designs employing visual analysis and respo
nse guided experimentation.