Inefficient Eye Movements: Gamification Improves Task Execution, But Not Fixation Strategy

Abstract

Decisions about where to fixate are highly variable and often inefficient. In the current study, we investigated whether such decisions would improve with increased motivation. Participants had to detect a discrimination target, which would appear in one of two boxes, but only after they chose a location to fixate. The distance between boxes determines which location to fixate to maximise the probability of being able to see the target: participants should fixate between the two boxes when they are close together, and on one of the two boxes when they are far apart. We “gamified” this task, giving participants easy-to-track rewards that were contingent on discrimination accuracy. Their decisions and performance were compared to previous results that were gathered in the absence of this additional motivation. We used a Bayesian beta regression model to estimate the size of the effect and associated variance. The results demonstrate that discrimination accuracy does indeed improve in the presence of performance-related rewards. However, there was no difference in eye movement strategy between the two groups, suggesting this improvement in accuracy was not due to the participants making more optimal eye movement decisions. Instead, the motivation encouraged participants to expend more effort on other aspects of the task, such as paying more attention to the boxes and making fewer response errors.

Six of one, half dozen of the other: Suboptimal prioritizing for equal and unequal alternatives

“…However, using hoops of unequal size did not help participants to solve the focus-or-divide task optimally, as there is no systematic tendency to shift from centre to side positions as Δ increases. This replicates previous versions of this experiment (Clarke & Hunt, 2016;James et al, 2017;James et al, 2019). Accuracy.…”

“…We confirmed and extended this conclusion in an experiment presented in the supplementary material, using a different manipulation of the symmetry between the two possible targets and a different task context. Briefly, we used an eye movement task and presented a small, brief target inside either a left or right square (see also Clarke and Hunt, 2016;James et al, 2019;Morvan and Maloney, 2012). Participants had to choose a place to fixate in anticipation of the target appearing.…”

“…Goodnow, 1995;Phillips and Edwards, 1966), although there are limits (for a review, see Camerer & Hogarthm, 1999). In a gamified version of the focus-or-divide dilemma (James et al, 2019), in which a penguin character could earn fish rewards for accurate performance, participants improved their performance on the task relative to participants who were not given this additional motivation.…”

“…This focus-or-divide dilemma is an example of a resource-allocation problem, and is a simplified version of a problem we face routinely in daily life: for example, in deciding which projects to try and accomplish in a given timeframe, or deciding where to wait for a person when our rendezvous point was vaguely defined. In experiments that present focus-or-divide dilemmas, people consistently demonstrate an inability to apply the simple solution described above (Clarke & Hunt, 2016;James, Clarke, & Hunt, 2017;James, Reuther, Angus, Clarke & Hunt, 2019;. In fact, the vast majority of people, across a range of task contexts, do not alter their decisions about whether to pursue one goal or two with the difficulty of the goals.…”

“…Participants have been shown to have all the information they need to implement the correct solution (James et al, 2017). Moreover, participants who explicitly understand the optimal solution can implement it easily (Hunt et al, 2019). In other words, this decision problem has an obvious-in-retrospect solution, but something about the way the problem has been presented to participants prevents them from discovering or implementing it.…”