Cl. Leonard et D. Orchard, THE PROBLEM OF GENERALIZING TO A LANGUAGE POPULATION - A RANDOM CONTROVERSY, Journal of speech and hearing research, 39(2), 1996, pp. 406-413
The rationale behind the use of item statistics as well as objections
to their use are reviewed. It is argued that investigators conducting
language research with disordered populations should follow the exampl
e of those conducting language research with normal populations and co
nsider calculating item statistics if they wish their results to gener
alize beyond the specific items tested.