Practitioner Development

Applied behavior analysis and statistical process control?

Hopkins (1995) · Journal of applied behavior analysis 1995
★ The Verdict

Skip automated SPC rules—eyeball your graphs and stay close to every data point.

✓ Read this if BCBAs who review or share visual analysis in team meetings.
✗ Skip if RBTs who only run programs and never plot data.

01Research in Context

01

What this study did

Martin (1995) wrote a warning paper, not an experiment.

He looked at how factories use Statistical Process Control (SPC).

Then he asked, "What if ABA starts using the same rules?"

02

What they found

The author says SPC could hurt ABA.

Automated control limits might replace your eyes on the graph.

That distance could lead to poor treatment choices.

03

How this fits with other research

Castañe et al. (1993) counted child studies and found most skipped treatment integrity checks.

Martin (1995) says, "See? We already drift from data; SPC would make it worse."

Falakfarsa et al. (2022) show the problem still lives—less than half of recent papers report integrity data.

Cox (2024) extends the same worry to value-based care analytics, proving the 1995 warning keeps echoing.

04

Why it matters

Your graph is your client’s voice.

Letting a red line tell you "significant" steals the chance to ask why.

Keep marking phase changes by hand, note anomalies, and talk with your team.

Stay close to each data point and your treatment stays honest.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Before you draw a phase line, write one sentence on the graph that says why you picked that day.

02At a glance

Intervention
not applicable
Design
theoretical
Finding
not reported

03Original abstract

This paper examines Pfadt and Wheeler's (1995) suggestions that the methods of statistical process control (SPC) be incorporated into applied behavior analysis. The research strategies of SPC are examined and compared to those of applied behavior analysis. I argue that the statistical methods that are a part of SPC would likely reduce applied behavior analysts' intimate contacts with the problems with which they deal and would, therefore, likely yield poor treatment and research decisions. Examples of these kinds of results and decisions are drawn from the cases and data Pfadt and Wheeler present. This paper also describes and clarifies many common misconceptions about SPC, including W. Edwards Deming's involvement in its development, its relationship to total quality management, and its confusion with various other methods designed to detect sources of unwanted variability.

Journal of applied behavior analysis, 1995 · doi:10.1901/jaba.1995.28-379