Not All People Are Cut Out for Strategic Criterion Shifting

Abstract

Decisions are often based on memory, but memories are often vague and ambiguous. Therefore, one must establish some standard for how strong a memory must feel before it is accepted as part of one’s past. This standard of evidence, called a decision criterion, should be adapted to the demands of a given situation: Sometimes it is critical to accept only strong memories as legitimate, whereas at other times it is more appropriate to act on the basis of relatively weak memory evidence. The ability to shift between these two standards as circumstances warrant is a critical element of successful decision making. We review recent work examining criterion shifting in recognition memory and argue that the true nature of criterion shifting may be seen only at the individual level: Whereas some people shift their criterion quite adaptively, others do not shift at all.

Measuring memory is harder than you think: How to avoid problematic measurement practices in memory research

“…However, insofar as K does not well-describe the shape of ROCs in this domain, these conclusions could indicate that it is response bias that has large and reliable individual differences and is correlated with intelligence, rather than (or in addition to) visual working memory "capacity" per se. In fact, it is known that response bias, and propensity to adapt this bias to a task, tends to be quite stable within individuals (Aminoff et al 2012;Kantner & Lindsay, 2012;Miller & Kantner, 2020). It is unlikely this completely explains all such correlations; for example, there is some work where researchers extract a single latent factor using multiple VWM tasks and test for an association between this factor and a measure of general intelligence (Unsworth, Fukuda, Awh, & Vogel, 2014; although such covariance techniques can still confound variations in memory strength with response bias; for a discussion of alternative linking modeling approaches see Turner, Forstmann, & Steyvers, 2019).…”

Who gives a criterion shift? A uniquely individualistic cognitive trait.

“…Alternatively, strategic criterion shifting tendencies and tendencies to report meta awareness in test items via confidence ratings may represent completely different decision process. Miller and Kantner (2020) conducted post hoc analyses on previously reported data in which participants both shifted criteria and rated confidence, but found no relationship between the extent of criterion shifting and metacognitive bias. However, no studies have systematically assessed this relationship a priori.…”

“…In recognition memory, criterion placement is the threshold of familiarity strength that must be exceeded to recognize items. Criterion placement, like criterion shifting, is quite variable across individuals (Aminoff et al, 2012, 2015; Frithsen et al, 2018; Kantner et al, 2015; Kantner & Lindsay, 2012, 2014; Layher et al, 2018; Miller & Kantner, 2020). Despite large between-subjects variability, Kantner and Lindsay (2012, 2014) propose that the within-subject consistency of criterion placement over time makes it a stable cognitive trait.…”

You can’t “count” how many items people remember in working memory: The importance of signal detection-based measures for understanding change detection performance

“…Our simulations and empirical results reveal that these types of memory demands are precisely the kind that can lead to changes in response bias, and that variations in K estimates lead to spurious conclusions as to the source of these purported correlations with intelligence. Given that much prior works suggests that there are substantial individual differences in response bias (Aminoff et al, 2012;Kantner & Lindsay, 2012;Miller & Kantner, 2020), it follows that a substantial part of the shared variance between intelligence and VWM capacity in such studies could instead reflect an association between intelligence and response bias. An analogous criticism has been repeatedly made in the study of the relationship between intelligence and cognitive control, where it remains unclear whether associations between intelligence and performance on cognitive control (e.g., Eriksen Flanker tasks) reveal shared variance between executive function and intelligence, or shared variance between individual differences in third variables, such as response policies (e.g., speed/accuracy tradeoffs in cognitive control tasks) and intelligence (e.g., Burgoyne & Engle, 2020;Frischkorn & Schubert, 2018).…”