Using Eye Tracking to Investigate the Evaluation-Performance Relationship in Visual Attention

摘要

Recently, two accounts have attempted to account for the effects of evaluation on performance, the mere effort account (Harkins, 2006; McFall, Jamieson, & Harkins, 2009) and the focus of attention account (Muller & Butera, 2007; Normand, Bouquet, & Croizet, 2014). The focus of attention account suggests that undergoing evaluation causes a reduction in processing capacity, which, in turn, leads individuals to prioritize the most relevant information and disregard less focal information. In support of this notion, Normand, Bouquet, and Croizet (2014) report an experiment in which participants were asked to look at a fixation cross followed by a brief abrupt onset cue that flashed in one of four locations forming a square followed by a display of four letters (three Qs and one O) in the same square. The participants' task was to identify the location of the O as quickly and as accurately as possible. The brief onset cue appeared either in the location where the O would appear (valid trials), in a location where the Q would appear (invalid trials), or onsets flashed in all four locations (neutral trials). Normand et al., found that participants subject to evaluation exhibited smaller cueing effects (i.e., the difference between invalid and valid trial reaction times) than their non-evaluated counterparts, which they interpreted as evidence for the cue having less impact on reaction times.;Alternatively, the mere effort account argues that undergoing evaluation leads participants to be highly motivated to perform well, which potentiates (i.e., makes more likely) the prepotent (i.e., dominant) response. Task performance is then dependent on whether this response is correct, as well as whether or not participants have the knowledge, opportunity and motivation to correct it if it is incorrect. Normand et al. proposed that the spatial cue presented prior to the target array represents a dominant response, which should have produced larger, not smaller, cueing effects. However, this prediction for mere effort does not take into account the motivation to correct this response when it is incorrect, as seen in previous work (e.g., McFall, Jamieson, & Harkins, 2009). Additionally, Normand et al. draw conclusions about this process by simply interpreting terminal reaction time data even though a more definitive measure is available. The current work aimed to replicate and expand upon this research with the addition of eye-tracking, which allows a direct test of the two accounts.;In Experiment 1, using the stimulus parameters reported by Normand et al. (2014, Exp. 1), we found that the potential for evaluation did affect terminal reaction times, but not because participants looked away from or toward the cue. In fact, our best estimate is that participants looked at the cue only 30% of the time, perhaps because the letters in Normand et al.'s display were so large that the participants could see the target without even moving their eyes.;In Experiment 2, by changing stimulus parameters, it appeared that we were successful at producing a prepotent response since participants looked at the cue on 74% of the trials. However, once again, the significant evaluation effect in terminal reaction times was not the result of evaluated participants looking away from or toward the cue more than non-evaluated participants. Instead, through exploratory analyses, we found that what appeared to be a strong tendency to look at the cue was the result of a bias to look toward the top-left of the target display, which was accentuated for evaluated participants.;In Experiment 3, we attempted to better understand this location bias and concluded that this behavior was the actual prepotent response in these designs. In both Experiments 2 and 3, undergoing experimenter evaluation led to potentiation of the location bias, which produced slower reaction times for valid trials in Experiment 2, and, interestingly, no effect on overall reaction times in Experiment 3. Each of these experiments support the mere effort account and not the focus of attention account.;Experiments 4 and 5 represent an attempt to produce a set of stimulus parameters that simultaneously eliminate the location bias seen in Experiments 2 and 3 and make eye movements to the abrupt onset the prepotent response. In Experiment 4, we found that the stimulus parameters actually produced a ceiling effect for eye movements to the abrupt onset, which precluded testing the potentiation hypothesis of mere effort. In Experiment 5 in which we reduced the potency of the cue, we found that evaluated participants were faster to respond on invalid trials, and trended towards being faster on valid trials, but the eye-tracking data revealed no evidence for potentiation of the prepotent response (i.e., orienting to the location of the abrupt onset). Additional analyses revealed that evaluated participants were fastest on trials in which they did not orient to the abrupt onset or to the target location, suggesting that being evaluated still led to motivated responding, but without potentiation of the prepotent response.;Taken together, these findings support a motivational account of performance under evaluation despite the limitations revealed in the mere effort account. In addition, the research demonstrates the pitfalls of the common practice of using single behavioral measures (e.g., reaction time) to infer mediating process.