Testing the Serial Processing Model of Item and Source Retrieval: Applying the Additive Factor Method to Source Monitoring
Tue-HS3-Talk V-01
Presented by: Hilal Tanyas
Sternberg’s (1969) additive factor method (AFM) is a longstanding technique to test the seriality of latent cognitive processing stages based on individuals’ observable response times (RTs). The AFM requires selective manipulations affecting the processing time of a single stage without changing the durations of other stages. If experimental factors influence different processing stages selectively, the combined effect of these factors on the mean RT is additive, statistically manifested by significant main effects and no interaction. In contrast, the presence of an interaction conflicts with the assumption that two stages are strictly serial, indicating temporal overlap of subprocesses to some extent. By implementing the AFM to source monitoring, our aim herein is to test whether retrieval processes for an item (e.g., what was said?) and its source (e.g., who said it?) operate serially or in parallel. Inspired by the selective effects observed on memory performance, we manipulated item encoding (i.e., generating versus reading the study items; Mulligan et al., 2006) and source similarity (i.e., dissimilar versus similar sources; Bayen et al., 1996) in a fully crossed between-subjects factorial design. In Experiment 1 (N = 128), source similarity affected source latencies but item generation unexpectedly did not result in faster item retrieval despite the expected performance benefit, preventing further application of the AFM. With a modified procedure, Experiment 2 (N = 128) yielded the expected selective effects on item and source latencies, and the additivity of these effects on item latencies suggests seriality of item and source retrieval.
Keywords: additive factor method, item memory, response time, source memory, source monitoring