Probability learning in an uncertain world: How children adjust to changing contingencies

Abstract

We regularly make predictions about future events, even in a world where events occur probabilistically rather than deterministically. Our environment may even be non-stationary such that the probability of an event may change suddenly or from one context to another. 4–6 year olds and adults viewed 3 boxes and guessed the location of a hidden toy. After 80 trials with one set of probabilities assigned to the 3 boxes, the spatial distribution of these probabilities was altered. Adults easily responded to this change, with participants who maximized in the first half (by choosing the most common location at a higher rate than it was presented) being the fastest at making this shift. Only the older children successfully switched to the new location, with younger children either partially switching, perseverating on their original strategy, or failing to learn the first distribution, suggesting a fundamental development in children’s response to changing probabilities.

Statistical learning abilities and their relation to language

“…Thus, mapping individual differences in SL cannot focus only on how individuals differ in their ability to learn a novel set of regularities from scratch, but should also consider variance in individuals' capacity to learn novel information given existing assimilated regularities. This can be done by in‐task manipulations that simulate the learning of two consecutive structures, examining how regularities acquired during exposure to the first stream impacts learning of those in consecutive input (see Starling, Reeder, & Aslin, for preliminary findings), or alternatively, by quantifying real‐world statistics using corpora and assess how they impact learning of an artificial stream that requires updating of this statistical knowledge (and see Fine & Florian Jaeger, for a related work in the context of adaptation to novel syntactic structures). Relatedly, learning in the real world commonly involves multiple regularities present simultaneously (in contrast to common SL tasks in which participants are exposed to a single isolated input stream).…”

Testimony bias lingers across development under uncertainty.

“…In Experiment 1, weight to the testimony-following model decreased with age when the testimony only sometimes indicated the correct reward location. Additionally, adults were better at detecting a one-time pattern change in a probability-learning task than 4–six-year-old children (Starling et al, 2018). Further, individuals use minimal information to make trait-level attributions about social agents (Uleman et al, 2008), and children who are younger than 5 years of age are less accurate when making predictions about a social agent’s future behavior because they are not able to readily incorporate consistency cues (Boseovski & Lee, 2006).…”

“…Second, without a nonsocial comparison, we cannot assert that these patterns are caused by the social nature of the cue. Because age-related differences in flexible updating have been found in nonsocial contexts (Starling et al, 2018), it is possible that age-related differences are primarily driven by executive function, memory, or other cognitive skills that track with age. Nevertheless, Experiment 3 provides evidence that children can and do update their choices and strategies in the face of a one-time change in testimony.…”

The cognitive costs and advantages of children’s exposure to parental relationship instability: Testing an evolutionary-developmental hypothesis.

“…In expanding on this work, we posited that adapting to the minimal, fleeting occurrence of reward in unpredictable rearing environments shape the development of implicit, specialized cognitive abilities to successfully capitalize on immediate payoffs by quickly detecting and acquiring random, fluctuating resources. Accordingly, our paper was designed to test the hypothesis that parental relationship instability predicts children’s enhanced implicit cognitive skills in a reward probability learning task designed to assess the ability to detect fluctuating reward based on immediate, trial-by-trial feedback (Ashby & Maddox, 2005; Starling et al, 2018).…”