Should we run the Inequity Game online? Using valuable “virtual” resources influences children’s choices

• In Event: Created (Un)Equal?: Impact of Resource Type and Value on Children's Fairness Concerns

Abstract

Appealing resources are important for incentivized measures of prosocial behavior. If denying oneself resources in service of trustworthiness or fairness is to be costly in a psychological study, much like in “the real world,” the resources should be rewarding. But might variations in the kind and value of a rewarding resource affect children’s decisions? Do virtual “tokens,” conducive for use in online studies, serve as effective stand-ins for resources like candy, which is often used in-person? To address these questions, we ran incentivized tasks with separate samples of 6—11-year-olds tested in-person (n = 60) and online (n = 87). Our online methods, conducted over Zoom, closely recreated our in-person methods, except for the key difference of resource type: we used candy in-person and virtual tokens (ultimately “exchanged for real prizes”) online. Would our online sample generate similar results as our in-person sample?

Participants completed the Trustworthiness Task (Amir et al., 2021) and Inequity Game (Blake & McAuliffe, 2011) in counter-balanced order. In Trustworthiness, participants chose how many resources to share with an absent child whose prior actions had benefited them. In the Inequity Game, a measure of inequity aversion, participants chose whether to accept or reject several allocations between themselves and an absent child. Allocation “acceptances” distributed resources to each child; “rejections” meant no one received them. Advantageous Inequity allocations consisted of 4 resources for the participant and 1 for the absent child; rejections caused neither child to receive tokens for that trial.

Trustworthiness: A linear model run on the online data, with number shared as the dependent measure and age (months) as the key predictor, found that sharing increased with age (β = .02, SE = .007, p = .003), replicating similar in-person findings (Figure 1). A model including the online and in-person samples, with sample type as a predictor, found no effect for sample (p = .44). Inequity Game: A generalized linear mixed model run on the online data, with Advantageous Inequity rejections as the dependent measure, found no significant effect of age (p = .35); in-person, however, rejections increased with age (p = .002), as in prior studies (Blake et al., 2015). A model including both datasets found a marginal effect of sample (p = .06) and a significant age by sample interaction (β = .06, SE = .03, p = .02), driven by older children rejecting fewer trials online than in-person. Thus, key results differed across samples (Figure 2).

Trustworthiness results were similar across samples, but Inequity Game results diverged. We attribute the reduced Advantageous Inequity rejections online to older children’s greater reluctance to “waste” tokens than to “waste” candy (used in-person). (In Trustworthiness, all tokens were divided between the participant and the absent child and thus were never “wasted.”) Our results indicate mixed prospects for tasks involving tokens, which are conducive to online administration. Even in situations when children experience inequity aversion, efficiency concerns (Choshen-Hillel, Lin, & Shaw, 2020) may reduce children’s Inequity Game rejections if high-value resources, like fungible tokens, are used.

Authors

• Richard E Ahl, Boston College
  Presenting Author

• Kelsey Hannan

• Aaron Baker

• Dorsa Amir, Boston College

• Katherine McAuliffe, Boston College