Pseudocontingencies: An integrative account of an intriguing cognitive illusion.

Abstract

are gratefully acknowledged. Correspondence concerning this manuscript should be addressed via email to:kf@psychologie.uni-heidelberg.de. AbstractThe term pseudocontingency (PC) denotes the logically unwarranted inference of a Pseudocontingencies 2 contingency between two variables X and Y from information other than pairs of x i , y i observations, namely, the variables' univariate baserates as assessed in one or more ecological contexts. We summarize recent experimental evidence, showing that PCs can play a pivotal role in many areas of judgment and decision making. We argue that the exploitation of the informational value of baserates underlying PCs offers an alternative perspective on many phenomena in the realm of adaptive cognition that have been studied in isolation so far.Although PCs can lead to serious biases under some conditions, they afford an efficient strategy for inductive inference making in probabilistic environments that render baserate information, rather than genuine covariation information, readily available. Adaptive behavior hinges on our ability to extract meaningful regularities from the flood of information provided by the physical and social environment. In an uncertain world, discerning regularities often amounts to assessing statistical contingencies between two variables over a series of events, such as correlations between causes and events, signals and significant events, predictions and feedback, actions and consequences, or behaviors associated with particular personality types. It is no exaggeration to claim that the ability to assess contingencies is crucial for adaptive learning and behavior, for rational action and decision making, and -ultimately -for survival in a risky and uncertain world. Thus, contingency assessment is commonly considered a major module of inductive intelligence, as stated in several pertinent reviews (Allan, 1993;Allan, Hannah, Crump & Siegel, 2008;Alloy & Tabachnik, 1984;Arieh & Algom, 2002;Crocker, 1981;Fiedler, 2000).In this article, we present an alternative perspective on inductive learning, drawing on pseudocontingencies rather than contingencies proper. The major claim we will put forward is that when organisms figure out correlations or contingencies (in case of nominally scaled variables), they often engage in a cognitive inference process that is sensitive to something different from a genuine correlation. To start with a definition, the term "pseudocontingency" denotes the (logically unwarranted) inference of a contingency between two variables X and Y from information other than pairs of x i , y i observations. In empirical reality, there are many situations in which joint observations of two or more variables are not available, or in which environmental or mnemonic constraints preclude the use of such genuine contingency information. In these situations, pseudocontingencies (PCs) can be inferred from other information, particularly, from the covariation of unequal baserates of high and low levels of X and Y over an ecological context fact...

On the adaptive flexibility of evaluative priming

“…A study by Freytag, Bluemke, and Fiedler (in press) revealed that EP is also sensitive to base ratedriven strategies in a paradigm called "pseudocontingencies" (cf. Fiedler, Freytag, & Meiser 2009). Although this growing evidence for alternative strategies and interventions is quite in the spirit of our flexible cue-utilization approach, the focus of the present research was on the interplay of two generic cues that are potentially relevant to any EP setting.…”

Pseudocontingencies can override genuine contingencies between multiple cues

“…A comprehensive answer to this crucial question is beyond the scope of this short research note (for a longer discussion, see Fiedler et al, 2009). Suffice it to point out that PCs are more than a worst-case remedy in the absence of more appropriate inferences.…”

“…They often function as useful inference schemes in their original learning context and lead to erroneous inferences only when carried over to new environments. This insight pertains to the pseudocontingency (PC) illusion in particular (Fiedler & Freytag, 2004;Fiedler, Freytag, & Meiser, 2009). The affix pseudo refers to the fact that a contingency between two dichotomous variables, or a correlation between any pair of variables, is inferred from information that does not fit the logic of a correlation.…”

“…Previous research has been confined to the elementary case of PC inferences between only two focal variables, although PC effects have been obtained in both simple and complex task settings, using both dichotomous and continuous stimulus variables (see Fiedler et al, 2009). PCs have been observed for the assessment of correlations in personality tests (Fiedler & Freytag, 2004), groups and stereotypically associated behaviors (Freytag, 2003;Meiser & Hewstone, 2004), evaluation of student achievement (Fiedler, Freytag, & Unkelbach, 2007), and operant feedback learning (Kutzner, Freytag, Vogel, & Fied ler, 2008).…”

Contingency is used to prepare for outcomes: Implications for a functional analysis of learning

“…Both of them also appear to be independent of contingency and different from causal judgments. Moreover, related research is also showing that people tend to use base rates, rather than contingencies, to infer correlations and to make predictions (Fiedler, Freytag, & Meiser, 2009). …”