Aging and Variability in Performance

Shammi, Prathiba; Bosman, Elizabeth A.; Stuss, Donald T.

doi:10.1076/anec.5.1.1.23

Cited by 100 publications

(93 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This article introduces the EZ-diffusion model, a simplified version of the diffusion model that is able to uniquely determine these three important unobserved variables from just three observed quantities: MRT, P c , and the variance of response times for correct decisions (VRT ). Our mathematical analysis will show that VRT is much more informative with respect to subject ability or task difficulty than is MRT, echoing recent empirical insights in the aging literature and elsewhere (e.g., Hultsch, MacDonald, & Dixon, 2002;Li, 2002;MacDonald, Hultsch, & Dixon, 2003;Shammi, Bosman, & Stuss, 1998). …”

mentioning

confidence: 83%

An EZ-diffusion model for response time and accuracy

Wagenmakers

Maas

Grasman

2007

Psychonomic Bulletin & Review

546

786

View full text Add to dashboard Cite

For a two-choice response time (RT) task, the observed variables are response speed and response accuracy. In experimental psychology, inference usually concerns the mean response time for correct decisions (i.e., MRT) and the proportion of correct decisions (i.e., P c ). The immediate problem is that MRT and P c are in a trade-off relationship: Participants can respond faster, and hence decrease MRT, at the expense of making more errors, thereby decreasing P c (see, e.g., Pachella, 1974;Schouten & Bekker, 1967;Wickelgren, 1977). This so-called speed-accuracy trade-off has for a long time bedeviled the field. Consider 2 participants in an experiment, Amy and Rich. Amy's and Rich's performance is summarized by MRT 0.422 sec, P c .881, and MRT 0.467 sec, P c .953, respectively. Amy responds faster than Rich, but she also commits more errors. Thus, it could be that Amy and Rich have the same ability, but Amy risks making more mistakes. It could also be that Amy's ability is higher than that of Rich, or vice versa. If we only consider MRT and P c , there appears to be no way to tell which of these three possibilities is in fact true. Now consider George, whose performance is characterized by MRT 0.517 sec, P c .953. George responds more slowly than Rich, whereas their error rates are identical. An explanation solely in terms of the speed-accuracy trade-off cannot account for this pattern of results, and therefore most researchers would confidently conclude that Rich performs better than George. Unfortunately, if we only consider MRT and P c , it is impossible to go beyond these conclusions in terms of ordinal relations and quantify how much better Rich does than George. Note that the same arguments would hold if the example above had been in terms of 1 participant who responds in three different experimental conditions presented in three separate blocks of trials. In this case, comparison of performance across the different conditions is complicated by the fact that task performance may be simultaneously influenced by task difficulty and response conservativeness.In sum, both MRT and P c provide valuable information about task difficulty or subject ability, but neither of these variables can be considered in isolation. When MRT and P c are considered simultaneously, however, it is not clear how to weigh their relative contributions to arrive at a single index that quantifies subject ability or task difficulty.A way out of this conundrum is to use cognitive process models to estimate the unobserved variables assumed to underlie performance in the task at hand. The field of research that uses cognitive models for measurement has been termed cognitive psychometrics (Batchelder, 1998;Batchelder & Riefer, 1999;Riefer, Knapp, Batchelder, Bamber, & Manifold, 2002), and similar approaches in other paradigms have included those of Busemeyer and Stout (2002);Stout, Busemeyer, Lin, Grant, and Bonson (2004);and Zaki and Nosofsky (2001). Here, the focus is on the diffusion model for two-choice RT tasks (see, e.g., Ratcliff, 1978...

show abstract

mentioning

confidence: 83%

An EZ-diffusion model for response time and accuracy

Wagenmakers

Maas

Grasman

2007

Psychonomic Bulletin & Review

546

786

View full text Add to dashboard Cite

show abstract

“…Higher IIV has indeed been observed in pathological subgroups showing lower cognitive functioning (Bodling, Denney, & Lynch, 2012;Burton, Strauss, Hultsch, Moll, & Hunter, 2006;de Frias, Dixon, Fisher, & Camicioli, 2007;Hultsch, MacDonald, Hunter, Levy-Bencheton, & Strauss, 2000;Murtha, Cismaru, Waechter, & Chertkow, 2002;Shammi, Bosman, & Stuss, 1998;Stuss, Pogue, Buckle, & Bondar, 1994). Numerous studies have focused on older adults, and a few on children.…”

Section: Intraindividual Variabilitymentioning

confidence: 99%

Working memory and intraindividual variability in processing speed: A lifespan developmental and individual-differences study

et al. 2014

View full text Add to dashboard Cite

Working memory (WM) and intraindividual variability (IIV) in processing speed are both hypothesized to reflect general attentional processes. In the present study, we aimed at exploring the relationship between WM capacity and IIV in reaction times (RTs) and its possible variation with development across the lifespan. Two WM tasks and six RT tasks of varying complexity were analyzed in a sample of 539 participants, consisting of five age groups: two groups of children (9-10 and 11-12 years of age), one group of young adults, and two groups of older adults (59-69 and 70-89 years of age). Two approaches were adopted. First, low-span and high-span individuals were identified, and analyses of variance were conducted comparing these two groups within each age group and for each RT task. The results consistently showed a span effect in the youngest children and oldest adults: High-span individuals were significantly faster and less variable than low-span individuals. In contrast, in young adults no difference was observed between high-and lowspan individuals, whether in terms of their means or IIV. Second, multivariate analyses were conducted on the entire set of tasks, to determine whether IIV in RTs brought different information than the mean RT. The results showed that, although very strongly correlated, the mean and IIV in speed should be kept separate in terms of how they account for individual differences in WM. Overall, our results support the assumption of a link between WM capacity and IIV in RT, more strongly so in childhood and older adulthood.Keywords Aging . Development . Individual differences . Reaction time analyses/methods . Working memory . Intraindividual variabilityThe objective of the present article is to address the issue of the relationship between working memory (WM) capacity and within-task intraindividual variability (IIV) in reaction time (RT) tasks, investigating whether this relationship might vary across the lifespan. In this introduction, we will briefly present our approach to WM-in conformity with the demand of the editor of this special issue-and then discuss why a potential relationship can be hypothesized between WM and IIV and whether individual differences in WM could be related to IIV.As a number of researchers interested in developmental and/ or individual differences in WM, we consider the performance in WM tasks to be determined by a set of underlying attentional processes. These processes, such as the activation and inhibition of relevant informational units, are considered to become more efficient with age, accounting for part of cognitive development in children, and to vary in efficacy across individuals. We consider WM to be an index of what has frequently, and rather loosely, been labeled general resources. WM capacity, while remaining severely limited, increases with age both because of an increase in neurobiological mechanisms and because the use of resources becomes more efficient with experience. There is here a major point of departure between cognitive ex...

show abstract

“…These studies investigated either iSDs (e.g., Shammi, Bosman, & Stuss, 1998), quantiles of RT distributions (e.g., Salthouse, 1993), parameters of statistical distribution functions (e.g., the exGaussian function; Spieler, Balota, & Faust, 1996;West, Murphy, Armilio, Craik, & Stuss, 2002), or parameters of theoretical process models that were fitted to the RT distributions (e.g., Ratcliff, Thapar, & McKoon, 2006a).…”

Section: Empirical Findingsmentioning

confidence: 99%

On the relation of mean reaction time and intraindividual reaction time variability.

Schmiedek

Lövdén

Lindenberger

2009

Psychology and Aging

117

104

View full text Add to dashboard Cite

Researchers often statistically control for means when examining individual or age-associated differences in variances, assuming that the relation between the 2 is linear and invariant within and across individuals and age groups. We tested this assumption in the domain of working memory by applying varianceheterogeneity multilevel models to reaction times in the n-back task. Data are from the COGITO study, which comprises 101 younger and 103 older adults assessed in over 100 daily sessions. We found that relations between means and variances vary reliably across age groups and individuals, thereby contradicting the invariant linearity assumption. We argue that statistical control approaches need to be replaced by theoretical models that simultaneously estimate central tendency and dispersion of latencies and accuracies and illustrate this claim by applying the diffusion model to the same data. Finally, we note that differences in reliability between estimates for means and variances need to be considered when comparing their unique contributions to developmental outcomes.

show abstract

Aging and Variability in Performance

Cited by 100 publications

References 0 publications

An EZ-diffusion model for response time and accuracy

An EZ-diffusion model for response time and accuracy

Working memory and intraindividual variability in processing speed: A lifespan developmental and individual-differences study

On the relation of mean reaction time and intraindividual reaction time variability.

Contact Info

Product

Resources

About