Figure 4-1
Responses for subsequent targets. A, Hits Trend toward more hits for expected compared with unexpected targets. B, CR. Main effect of foil similarity. We examined the effect of contextual expectation on corrected-recognition decisions (hits – FA), for the three foils. Importantly, as there are multiple foils tied to each target, each with a different distribution of responses based on the level of similarity, standardized signal detection theory (SDT) measures, such as d′, cannot be computed. Instead, a similar approach has been used (Yassa and Stark, 2011) to calculate a lure discrimination index using the probability of new responses to similar foils and targets. Corrected recognition rates were subjected to a two (expected vs unexpected) by three (F1, F2, F3) repeated-measures ANOVA, performed using JASP v0.8.6 (https://jasp-stats.org/). Although this approach takes into account the tendency to respond old, it is also important to note it provides less power compared with the mixed-effects modeling. Nevertheless, similar to the results for the mixed-effects model, using a 2 (expectation) by 3 (foil similarity) repeated-measures ANOVA we found a main effect of foil similarity (F(2,46) = 23.7, p < 0.001, ɳ2p = 0.508 and a main effect of expectation (F(1,23) = 5.11, p = 0.034, ɳ2p = 0.182). The interaction between foil similarity and expectation approached, but failed to reach statistical significance (F(2,46) = 1.51, p = 0.23, ɳ2p = 0.06). Again, the beneficial effect of contextual expectation violation was manifested in F1 (t(23) = 2.15, p = 0.042, Cohen's d = 0.44), there was a trend in the same direction for F2 (t(23) = 1.59, p = 0.12 Cohen's d = 0.33), and no difference between expectation condition was observed in F3 (t(23) = 0.14, p = 0.89, Cohen's d = 0.03). Download Figure 4-1, TIF file