Jm. Ennis et al., THURSTONIAN MODELS FOR VARIANTS OF THE METHOD OF TETRADS, British journal of mathematical & statistical psychology, 51, 1998, pp. 205-215
Citations number
24
Categorie Soggetti
Psychology, Experimental","Psychologym Experimental","Statistic & Probability","Mathematics, Miscellaneous","Mathematics, Miscellaneous","Statistic & Probability
Variants of the method of tetrads are discussed and their psychometric
functions are both derived and presented in table form. For the first
variant, the Unspecified method, subject response is correct provided
that the subject, when presented with two stimuli from one set and tw
o stimuli from a I;second set, correctly segregates the four stimuli i
nto their two appropriate groups. in the second variant, the Specified
method, subject response is correct provided that the subject., when
presented with two stimuli from one set and two stimuli from a second
set, correctly discerns the two strongest stimuli. For the Unspecified
method, P-c=1 - 2 integral(-infinity)(infinity) phi(x)[2 Phi(x)Phi(x-
delta)-[Phi(x-delta)](2)]dx, where P-c is the probability of correct r
esponse, and phi(x) and Phi(x) are the density and distribution functi
ons respectively, for the standard normal distribution. The parameter
delta is the perceptual difference between stimulus means and correspo
nds to the signal detection measure of sensitivity d'. For the Specifi
ed method, P-c=1-2 integral(-infinity)(infinity)phi(x)Phi(x)[2 Phi(x-d
elta)-[Phi(x-delta)](2)]dx For each of the two variants, tables connec
t the probability of correct response to delta.