Academic journal article Memory & Cognition

Associative Interference in Recognition Memory: A Dual-Process Account

Academic journal article Memory & Cognition

Associative Interference in Recognition Memory: A Dual-Process Account

Article excerpt

Associative interference from overlapping word pairs (A-B, A-D) reduces recall but has inconsistent effects on recognition. A dual-process account suggests that interference conditions reduce recollection but increase familiarity. This is predicted to increase recognition false alarms but have variable effects on recognition hits, depending on the relative contribution of recollection and familiarity. In three experiments that varied materials (sentences or random nouns) and test type (associative or pair recognition), interference conditions always increased recognition false alarms, but sometimes increased and sometimes decreased recognition hits. However, remember hits always decreased and know hits always increased with interference, patterns predicted of the recollection and familiarity processes, respectively. According to the dual-process view, a manipulation that affects the component processes in opposite ways can produce inconsistent patterns of recognition performance as the relative contribution of recollection and familiarity changes across tasks.

According to dual-process theories of recognition, two qualitatively different retrieval processes contribute to recognition memory (Atkinson & Juola, 1974; Mandler, 1980; Reder et al., 2000; Yonelinas, 1994). The theories vary in detail, but their descriptions of the processes are generally similar. Recollection retrieves episodic or associative details. It is typically described as a search for specific information, a process perhaps similar to that involved in the recall task (Clark, 1999; Humphreys, 1978; Mandler, 1980). Familiarity produces nonspecific information about past occurrences. A retrieval cue matches each item in memory to some degree on the basis of similarity or learned association, and the global or aggregate match is represented psychologically as a unidimensional sense of familiarity strength.

The detrimental effect on recall of studying overlapping paired associates (A-B, A-D or A-B, A-B^sub r^) is well documented in the classic literature on proactive and retroactive interference. However, the same literature notes that evidence of associative interference in recognition is often absent, or at best inconsistent (for a review, see Postman, 1976). One way to conceptualize the associative interference paradigm is as a manipulation of the similarity between items (e.g., Dyne, Humphreys, Bain, & Pike, 1990). Overlapping pairs (dog-tree, dog-lake) are more similar to one another than are nonover-lapping pairs (dish-skirt, star-ball). As a consequence, a recognition probe drawn from a study list of overlapping pairs has, on average, a greater match to items in memory than does a probe drawn from a study list of nonover-lapping pairs. As in the earlier interference literature, studies manipulating the relatedness of test probes to the study list have produced inconsistent findings. The present study describes a dual-process account that can shed light on these puzzling results.

Increasing the similarity or relatedness of nonstudied probes to the study list increases the false alarm rate. This has been observed with associatively, semantically, or orthographically related words (Anisfield & Knapp, 1968; Roediger & McDermott, 1995; Shiffrin, Huber, & Marinelli, 1995; Underwood, 1965; Vogt & tumble, 1973), word pairs (J. R. Anderson & Reder, 1999; Dyne et al., 1990), alphanumeric characters (Flagg, 1976; Reitman & Bower, 1973), geometric shapes (Medin & Schaffer, 1978), faces (Busey & Tunnicliff, 1999; Vokey & Read, 1992), pictures (Henkel & Franklin, 1998; Koutstaal & Schacter, 1997; Strack & Bless, 1994), and sentences (Bransford & Franks, 1971; Cantor & Engle, 1993; Holmes, Waters, & Rajaram, 1998; King & Anderson, 1976; Reder & J. R. Anderson, 1980). Increasing the similarity or relatedness of studied probes to the study list produces inconsistent results, sometimes slightly increasing the hit rate (Dyne et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.