Deliberative Choice Strategies in Youths: Relevance to Transdiagnostic Anxiety Symptoms

Abstract

Aberrant decision-making characterizes various pediatric psychopathologies; however, deliberative choice strategies have not been investigated. A transdiagnostic sample of 95 youths completed a child-friendly sequential sampling paradigm. Participants searched for the best offer by sampling a finite list of offers. Participants’ willingness to explore was measured as the number of offers sampled, and ideal task performance was modeled using a Markov decision-process model. As in previous findings in adults, youths explored more offers when lists were long compared with short, yet participants generally sampled fewer offers relative to model-estimated ideal performance. Searching deeper into the list was associated with choosing better price options. Analyses examining the main and interactive effects of transdiagnostic anxiety and irritability symptoms indicated a negative correlation between anxiety and task performance ( p = .01, η p2 = .08). Findings suggest the need for more research on exploratory decision impairments in youths with anxiety symptoms.

Methodological Remarks Regarding Optimal Stopping Tasks and the Implications for Sampling Biases

“…In other words, the only change compared to Furl et al (2019) was that the images of faces were replaced with numerical smartphone prices. Because of the use of numbers instead of image-based stimuli, we hypothesised that this adaptation to the paradigm would be sufficient to induce an undersampling bias, in line with the results of previous studies that employed numerical stimuli (e.g., Baumann et al, 2020;Bearden & Connolly, 2007;Cardinale et al, 2021;Costa & Averbeck, 2015;Furl & Averbeck, 2011;Sonnemans, 2000).…”

“…Bayesian ideal observer model Human behaviour on our optimal stopping tasks was compared to a Bayesian ideal observer model, for which performance is Bayesian optimal. This computational Markov decision process (MDP) model has been used in previous literature, including Costa and Averbeck (2015), Furl et al (2019) and Cardinale et al (2021) (for a mathematical description of the model, see Supplementary Materials, Text B). Just like the historically used Gilbert and Mosteller model (Gilbert & Mosteller, 1966), the ideal observer model's expectations about future option values are based on a standard normal distribution, from which future options are assumed to be generated.…”

“…Previous studies exploring decision-making on economic optimal stopping tasks have reported that decision makers primarily stop searching too early compared to models of optimality (undersampling) (Bearden et al, 2006;Cardinale et al, 2021;Costa & Averbeck, 2015;Guan et al, 2014;Seale & Rapoport, 1997;Sonnemans, 2000). However, there are also examples of specific optimal stopping tasks on which people sample too much (oversampling), such as when choosing a date (Furl et al, 2019).…”

“…The current paper addresses this question by investigating various methodological task features that may affect sampling biases in three separate studies. This is important in light of recent research suggesting that optimal stopping tasks might have a wider real-world application, for example as part of cognitive behavioural therapy in anxiety disorders (Cardinale et al, 2021), or even as a general measure of problem solving ability and psychometric intelligence (Lee et al, 2005). For these kinds of real-world applications to be realised, more uniform and standardised procedures for studying human behaviour on optimal stopping tasks are warranted.…”

“…Presently, numerous versions of optimal stopping tasks prevail, which complicate direct comparisons between studies. For instance, countless different stimuli are used across the literature to indicate the value of an option (e.g., Baumann et al, 2020;Cardinale et al, 2021;Costa & Averbeck, 2015;Furl et al, 2019;Goldstein et al, 2020;Guan & Stokes, 2020), but there is reason to suggest that the type of stimulus (e.g., numbers or images) might affect human sampling behaviour. Specifically, a study by Costa and Averbeck (2015) found that participants undersampled on an economic optimal stopping task compared to a Bayesian ideal observer model.…”