https://jingkaihong.netlify.app/category/jusgement-and-decision-making/Jusgement and Decision MakingWowchemy (https://wowchemy.com)en-usMon, 24 Nov 2025 00:00:00 +0000https://jingkaihong.netlify.app/media/icon_hucbde1a20235b5cb75c554ae0daae2535_146828_512x512_fill_lanczos_center_3.pnghttps://jingkaihong.netlify.app/category/jusgement-and-decision-making/https://jingkaihong.netlify.app/publication/hong_2025a/Mon, 24 Nov 2025 00:00:00 +0000https://jingkaihong.netlify.app/publication/hong_2025a/<!-- ### abstract Work with process tracing methods has identified reliable patterns in choice for accuracy, reaction times and attention. The most successful family of models at capturing patterns across multiple process measures has been evidence accumulation or Drift Diffusion Models (DDMs). Since DDM models most often incorporate only the value difference between options, most prior research has focused on the effect of varying value difference (easy versus hard choices). However recent work has demonstrated the effect of value magnitude on choice, particularly Ting and Gluth (2025). We use the data from 6 eye tracking experiments to attempt to replicate their conclusions and critically extend their approach. In addition, we present simulations of a DDM with an attentional bias (the aDDM) to demonstrate the emergent phenomena and model predictions, to compare against empirical data. Our findings largely replicate Ting and Gluth’s conclusions for reaction times and attention bias at last fixation, but do not find the reported effect upon choice accuracy. We also report novel analyses that show the timecourse of attentional bias for high versus low value magnitude trials. This shows that although the bias in the final fixation is robustly larger for low value magnitude trials, the onset is earlier and larger for high value magnitude trials. When considered alongside the need for motor response and action planning, this suggests the attention patterns may not be unequivocally incompatible with the aDDM. However, the aDDM and other models still fail to capture the overall pattern of results. --> <h3 id="talk-presented-at">Talk presented at</h3> <ul> <li> Subjective Probability Utility and Decision Making (SPUDM) (Presentation), Vienna, 2023 </li> <li> Society for Judgement and Decision Making (SJDM) (Poster), San Francisco, 2023 </li> <li> European Group of Process Tracing Studies (EGPROC 2025) (Presentation), Netherlands, 2025 </li> </ul> <div class="alert alert-note"> <div> Manuscript under review at <em>Cognition</em>. </div> </div>https://jingkaihong.netlify.app/publication/hong_2025b/Sun, 23 Nov 2025 00:00:00 +0000https://jingkaihong.netlify.app/publication/hong_2025b/<!-- ### abstract Across three datasets (N = 150) involving choices with dynamic visual stimuli, we examined how early and late information contributes to decision outcomes. We present analysis techniques that measure and visualize the relative weight given to incoming information throughout deliberation. There is consistent over-weighting of early information (primacy). Recency effects were present in all tasks employing free-response conditions, but over-weighting peaked several hundred milliseconds before a response was made, then rapidly reduced, with later incoming information having no detectable influence upon choice. This pattern is consistent with non-decision time after a choice is formed and in preparation of a motor response. We demonstrate this apparent recency bias can be produced even by standard Drift Diffusion Models. However, eye-tracking data shows the timecourse is incompatible with visual attention accounts and the gaze cascade phenomena. Recency effects vanished under the complete-observation condition, a pattern incompatible with memory and evidence decay accounts. --> <h3 id="talk-presented-at">Talk presented at</h3> <ul> <li> Subjective Probability Utility and Decision Making (SPUDM) (Presentation), Lucca, 2025 </li> </ul> <div class="alert alert-note"> <div> Manuscript under review at <em>Psychological Science</em>. </div> </div>https://jingkaihong.netlify.app/publication/hong_2025x/Fri, 21 Nov 2025 00:00:00 +0000https://jingkaihong.netlify.app/publication/hong_2025x/<!-- abstract Decision-making often occurs under conditions of limited cognitive resources and stringent time constraints. However, it remains unclear whether and when these con-straints affect preferences, and if so, why. The current research used three online studies to compare how time pressure and cognitive load separately and jointly influence gain/loss framing sensitivity in risky choices, a phenomenon that has demonstrated sensitivity to cognitive resource constraints in past work. Participants completed a gamble selection task, choosing between a sure option framed as a gain or loss and a probabilistic gam-ble, while experiencing manipulations of cognitive load (low vs. high) and time pressure (varying amounts of time to make the choice). Consistent with prior literature, we ob-served robust gain/loss framing effects: participants were more risk averse under gain frames and more risk seeking under loss frames. Severe time pressure (e.g., a 1-second decision window) intensified these framing effects, suggesting that constraining decision time increases reliance on intuitive, frame-dependent heuristics. This effect declined and vanished entirely under more moderate time pressure constraints. Cognitive load, how-ever, did not significantly alter framing sensitivity, even though evidence suggested that it reduced available cognitive resources and sped up response times in no-time-pressure conditions. These findings indicate that while both time pressure and cognitive load limit cognitive resources, and perhaps through similar mechanisms (i.e., reducing the time spent on choice), these limits must generally be quite severe to influence behaviour. The results highlight important distinctions in how common resource constraints influ-ence the construction and expression of preferences, informing both theoretical models of decision-making under constraints and practical interventions aimed at improving deci-sion quality. --> <div class="alert alert-note"> <div> Preparing for submission to <em>Journal of Experimental Psychology General</em>. </div> </div>