Academic journal article Memory & Cognition

The List Strength Effect: A Contextual Competition Account

Academic journal article Memory & Cognition

The List Strength Effect: A Contextual Competition Account

Article excerpt

Research on the list strength effect (LSE) has shown that learning some words on a list more strongly than others impairs memory for the weakly learned words when tested with a recall task. Norman (2002) demonstrated that the LSE also occurs within the recollection process of a recognition test. In this study, a mechanistic dual-process account of the LSE was tested that simultaneously makes predictions concerning additional sources of context in interference effects. In two experiments, we attempted to replicate Norman's (2002) findings and provide the basis for our modeling efforts. We found evidence for a recollection LSE in raw measures of responding, with memory performance also benefiting from reinstatement of perceptual characteristics at test. However, large differences in the hits between the lists were accompanied by small differences in false alarms, such that when d' is calculated, differences between the lists are not significant. We propose an account of the LSE whereby the actual effect of competition between items on the list is small, although present, and difficult to distinguish from large effects of bias due to the strength manipulations. We argue that our findings provide support for a mechanistic explanation of LSE that is based on competition of source activation and changes in the thresholds for responses.

The study of interference is central to understanding how memory works. Indeed, recent attention to the mechanisms underlying interference has allowed better understanding of memory processes in general (e.g., Anderson, 1981; Chandler, 1991; Kim & Glanzer, 1995; Murnane & Shiffrin, 1991a; Willis & Underwood, 1983; Yonelinas & Jacoby, 1994). In particular, the list length and list strength paradigms have been popular methods for investigating interference in recall and recognition tasks. List length manipulations test the sensitivity of memory to the number of items on a list, whereas list strength manipulations test memory sensitivity to the relative strength of items on a list. These types of experiments can help determine how information is related in memory and how information that is stored in the same context influences memory for other information.

List length experiments manipulate the number of items on a list and compare memory for lists with different numbers of items. List length manipulations have been found to affect both recognition and recall (e.g., Gillund & Shiffrin, 1984; Gronlund & Elam, 1994; Ohrt & Gronlund, 1999). Some researchers have questioned whether previously demonstrated list length effects (LLEs) in recognition were due to methodological confounds, such as retention interval, rather than to interference (Dennis & Humphreys, 2001; Murdock & Kahana, 1993). However, the results have held under carefully controlled conditions (Cary & Reder, 2003; Ohrt & Gronlund, 1999). Thus, at least under some conditions, increasing the number of items on a list makes it more difficult to remember any one of the items.

Tulving and Hastie (1972) were the first to investigate the effects of stronger learning of some items on memory for more weakly learned items. They found a list strength effect (LSE) for free recall, such that when some words on a list were presented twice, performance for the singly presented items was impaired. Ratcliff, Clark, and Shiffrin ( 1990) found some evidence for an LSE in the cued recall paradigm, but they were unable to find a consistent LSE for recognition. Many other researchers have also found little or no evidence for impairment in recognition tasks based on list strength (e.g., Hirshman, 1995; Murnane & Shiffrin, 1991a; Ratcliff, Sheu, & Gronlund, 1992; Yonelinas, Hockley, & Murdock, 1992). Murnane and Shiffrin (199Ib) were the first to show consistent evidence of an LSE for recognition, but only when each repetition of a word was presented within a unique sentence. They proposed that each presentation in a unique context is stored as a separate memory trace, creating a list length manipulation, rather than a list strength manipulation in which an individual trace would be strengthened. …

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.