Ranking Visualizations of Correlation (reVISit Replication)
Replicating correlation perception studies across encodings with reVISit sequencing and staircase designs.
I co-designed a replication of correlation just-noticeable-difference studies to understand how people read evidence across scatterplots, parallel coordinates, hexbins, and heatmaps. Using reVISit dynamic sequencing and staircase designs, we gathered large-scale crowdsourced data backed by preregistered analyses and provenance-aware logging. I built JND stimuli and adaptive staircases so we could calibrate correlation thresholds per participant while keeping the task replayable.
- Goal: Update classic correlation perception results with richer interaction traces and modern encodings.
- Method: OSF preregistration; crowdsourced experiment with adaptive sequencing and replayable logs.
- Status: Published; contributing to a first-authored short paper.