This analysis confirmed a statistically reliable difference in responses per scheduled reinforcer between the F group (M = 34, SD = 45) and both the SIVR (M = 87, SD = 54) and N groups (M = 82, SD = 55).
The greatest number of participants fall into the highest-rate category for both the SIVR and N groups, while the greatest number of F participants fall into the lowest-rate category.
As it was, the performances of participants in the N group often resembled the high-rate performances of the SIVR group.
As in Experiment 1, a SIVR group provided a comparison condition.
The SIVR condition employed the same instruction (see Table 3) and schedule as was used for this group in Experiment 1, except that the value of the parameters was halved for both the training and test schedules.
At the 75th percentile, test-phase response rates are uniformly high for participants in each of the variety groups, with the notable exception of DRL, and resemble the performance of the participant from the SIVR group.
The means for each type of density correspondence were then compared to the mean for the SIVR group using t tests.
017 (Bonferroni-corrected for three tests), chi-square analyses revealed a statistically reliable difference between the DRL group and both the SIVR group, [chi square](1, N = 63) = 8.
Comparisons of the variety trained groups to SIVR training further attested to the influence of discriminative schedule properties on variety-trained FI performance.