Intervention effects and relative variation as dimensions in experts' use of visual inference.

The team asked 21 ABA experts to judge 120 single-case graphs.

Each graph showed a treatment line and data points.

The experts rated how sure they were that the treatment worked.

The researchers then looked at which graph features matched high ratings.

Experts changed their ratings when they saw clear up or down shifts.

They did not change ratings when the data points were more spread out.

In short, they cared about effect patterns, not variability.

Wolfe et al. (2023) later tested this with new graphs.

They found that steep trend and big effect size still drive agreement.

Variability mattered a little, but less than trend, matching the 1982 view.

Diller et al. (2016) seemed to clash: they said variability swayed BCBA ratings.

The gap is about design type: 1982 used simple AB graphs, 2016 used multielement.

In multielement graphs, overlapping lines make variability stand out more.

So both papers can be right; the graph style changes what pops out.

When you eyeball a graph, first look for a clean level shift or trend change.

Do not let wide data scatter scare you off if the pattern is steady.

If you are using multielement designs, expect variability to grab your eye—pause and check the trend anyway.

Train new staff to spot effect patterns first; save variability checks for later.