Separating sensitivity from response bias: Implications of comparisons of yes-no and forced-choice tests for models and measures of recognition memory.

Abstract

A fundamental challenge to psychological research is the measurement of cognitive processes uncontaminated by response strategies resulting from different testing procedures. Test-free estimates of ability are vital when comparing the performance of different groups or different conditions. The current study applied several sets of measurement models to both forced-choice and yes-no recognition memory tests and concluded that the traditional signal-detection model resulted in distorted estimates of accuracy. Two-factor models were necessary to separate memory sensitivity from response bias. These models indicated that (a) memory accuracy did not differ across the tests and (b) the tests relied on the same underlying memory processes. The results illustrate the pitfalls of using a single-component model to measure accuracy in tasks that reflect 2 or more underlying processes.

Subliminal Smells can Guide Social Preferences

“…Given the uneven distribution of response probabilities (odor was present on 75% of trials and absent on 25% of trials), d 0 could have been critically distorted by individual response bias (b; Kroll, Yonelinas, Dobbins, & Frederick, 2002). We therefore adjusted d 0 for variability in b, obtaining standardized scores.…”

d' is not appropriate for contrasting yes-no and forced-choice recognition

“…On the other hand, some research suggests that the YN/2AFC relationship for discrimination is variable and depends on the stimuli and encoding conditions (Cook, Marsh, & Hicks, 2005). Further still, some research suggests that d'YN is an unreliable measure of discrimination and can lead to inflated measures of YN discrimination (Kroll et al, 2002). These latter two findings neither support the idea of equivalent discrimination between tasks nor the idea of superior 2AFC discrimination.…”

“…Table 1). Other published studies have shown this trend (e.g., Kroll et al, 2002;Yonelinas et al, 1992).…”

“…3 Unfortunately, these demonstrations of SDT's failure to predict the relationship between YN and 2AFC discrimination have been largely overlooked in subsequent research. The many studies that have contrasted YN and 2AFC discrimination explicitly state that the V2 adjustment was made for 2AFC performance (e.g., Bastin & Van der Linden, 2004;Cook, Marsh, & Hicks, 2005;Khoe et al, 2000;Kroll et al, 2002;Smith & Duncam, 2004) or indicate that the adjustment was made by referring to the 2AFC d' equation in Macmillan and Creelman (1991) or Green and Swets (1966) (e.g., Holdstock et al, 2002;Westerberg, Palier, Holdstock, Mayes, & Reber, 2006 Khoe et al, 2000, Yonelinas et al, 1992, a sentiment that provides converging support for the predictions of SDT as they apply to old/new discrimination. A number of other studies suggest that discrimination is superior in the 2AFC task even after the V2 adjustment (e.g., Aggleton & Shaw, 1996;Creelman & Macmillan, 1979;Deffenbacher et al, 1981), a sentiment that converges on the idea that the YN/2AFC relationship for discrimination differs from the relationship found in the detection literature.…”

“…However, very little research has explicitly examined the status of the V2 assumption. The V2 assumption is a very important assumption for the numerous studies that have used d' to contrast performance in yes-no (YN) and twoalternative forced-choice (2AFC) discrimination tasks (e.g., Bastin & Van der Linden, 2003;Cook, Marsh, & Hicks, 2005;Deffenbacher, Leu, & Brown, 1981;Green & Moses, 1966;Holdstock et al, 2002;Khoe, Kroll, Yonelinas, Dobbins, & Knight, 2000;Kroll, Yonelinas, Dobbins, & Frederick, 2002;Mayes et al, 2002;Nolde, Johnston, & Raye, 1998;Westerberg, Palier, Holdstock, Mayes, & Reber, 2006;Yonelinas, Hockley, & Murdock, 1992). If the V2 assumption is violated, then d' will not allow for balanced comparison to be made between YN and 2AFC discrimination as SDT claims that it will.…”