Specific and nonspecific treatment factors in the experimental analysis of behavioral treatment efficacy.

Eugenia Gras et al. (2003) wrote a how-to guide for researchers. They show ways to separate real treatment parts from fake ones.

Fake parts are things like the client liking you, expecting to get better, or just having a set schedule. The paper gives blueprints for trials that keep these parts the same across groups.

The authors say most behavior studies skip these controls. When we skip them, we can’t tell if gains come from our ABA moves or from plain attention.

They list easy fixes: add a sham treatment arm, keep visit length equal, and script your praise.

Carr et al. (1985) ran a six-month trial that did exactly what M et al. warn about. They stacked relaxation, diet, and weight loss on top of clinic check-ins. The extras gave zero added blood-pressure drop, matching the target paper’s point that nonspecific care often looks like it works.

Critchfield et al. (2003) showed the flip side. They removed rewards and saw more work, not less. Their discovery hinged on tight single-case controls, the kind M et al. want scaled up to group trials.

Normand et al. (2023) add the ethics layer. If you run these tight trials with your own clients, you must guard dual roles and get outside review, keeping the science clean and the client safe.

Next time you test a new prompt-fading package or token board, build in a placebo play group that gets equal time, equal praise, and equal structure. Only the target skill should differ. This one move tells you if your procedure is the hero or just the spotlight.