Digital filters in behavioral research.

Logue (1983) wrote a how-to guide for cleaning messy response-rate graphs.

The paper shows step-by-step math tricks called digital filters.

You run the filter on raw counts to strip out random jumps while keeping real trends.

Filtered curves look smooth but still show the real behavior change.

Noise that hides schedule effects or therapy gains drops out.

You can spot small, steady shifts you would miss with naked-eye graph reading.

PLISKOFF (1963) first showed us raw IRT scatter; Logue (1983) gives the mop to clean it up.

Nevin et al. (2005) and Gaucher et al. (2020) both plot choppy rate data; their findings would pop clearer after this filter.

Cao et al. (2026) split persistence into five lines; smoothing each line first would make the splits sharper and easier to sell to caregivers.

Next time your graph looks like a heart-rate trace, run a digital filter before you judge the intervention. Clean pictures speed up team decisions and keep parents nodding instead of squinting.