Frequency effects in recognition and recall

Abstract

Stimulus frequency, which is often evaluated using normative word frequency, is among the variables that have the most diverse and puzzling effects on memory. Word frequency can either facilitate or impair memory performance depending on the study and testing conditions. Understanding why and under what conditions frequency has positive or negative effects on performance is crucial for understanding basic properties about the human memory system. As a result, the study of word frequency has led to the development of multiple memory models. This chapter summarizes the current knowledge concerning word frequency effects on item recognition, associative recognition, free recall, cued recall, serial recall, and source memory. We also discuss how word frequency interacts with manipulations concerning presentation rate, list-composition, age of the participants, memory load, midazolam injections, response deadlines and remember-know judgements. This review of frequency effects in memory identified four major classes of empirical findings, which can be further subdivided into a total of 21 key phenomena that any theory should account for. Based on these phenomena, we identify three high-level principles that characterize the diverse effects of frequency on memory – the probe dependency principle, the dual process principle, and the resource demands principle.

Memory resources recover gradually over time: The effects of word frequency, presentation rate, and list composition on binding errors and mnemonic precision in source memory.

“…in fewer contexts (for a recent review, see Popov & Reder, 2020a;Osth & Dennis, 2015;Reder et al, 2000Reder et al, , 2007Dennis & Humphreys, 2001;Criss et al, 2011). As Criss et al (2011) argued, a null or a positive effect of high frequency cues is inconsistent with most current memory models.…”

“…The fact that the LF cue retrieval advantage can be masked depending on list composition and presentation rate is theoretically important. The LF retrieval advantage in item recognition has been attributed to less contextual competition for LF words because they have been experienced in fewer contexts (for a recent review, see Criss et al, 2011;Dennis & Humphreys, 2001;Osth & Dennis, 2015;Popov & Reder, 2020a;Reder et al, 2000Reder et al, , 2007. As Criss et al (2011) argued, a null or a positive effect of high frequency cues is inconsistent with most current memory models.…”

“…These two factors create a trade-off (Popov & Reder, 2020a;Reder et al, 2007), and, depending on the task conditions (e.g., presentation rate, list-composition, test type), can produce either a positive or a negative effect of frequency 2 (see Figure 1).…”

“…This inconsistent pattern of results is unsatisfactory because it makes it difficult to evaluate memory models. Numerous models have been proposed to explain word frequency effects in item recognition and free recall (for a review, see Popov & Reder, 2020a), and here we will focus on the source of activation confusion (SAC) model (Popov & Reder, 2020b;Reder et al, 2000Reder et al, , 2007. SAC depends on the following assumptions to explain a wide array of word frequency effects: 1) memory formation depletes a limited resource that recovers gradually over time; 2) HF words have stronger preexisting representations in memory compared to LF words; 3) the amount of resources required to process an item is an inverse function of its existing strength; 4) as a result, processing LF words deplete more resources than do HF words, leaving fewer resources to create episodic associations between words and an experiential context; 5) cues associated with more episodic contexts are less effective in retrieving any specific episode; and 6) LF words are associated with fewer contexts and, as such, spread more activation to episodes they are connected to.…”

“…Normative word frequency has played a key role in empirical research and theoretical development on human memory (Clark, 1992; Glanzer & Bowles, 1976; Hulme et al, 2003; Mandler et al, 1982; Reder et al, 2000; Popov & Reder, 2020b). Its prominence is due to the fact that it can either facilitate or impair memory performance depending on the nature of the study task, the test, and the details of the study sequence (for a review, see Popov & Reder, 2020a). For example, high-frequency (HF) words are recognized less well in item recognition tests, leading to more false alarms and fewer hits than low-frequency (LF) words (e.g., Glanzer & Adams, 1990; Reder et al, 2000).…”