The effects of low alcohol beverages on alcohol consumption and impairment.

Researchers worked with four adults who drank at one bar. They served the same mixed drinks at half the usual alcohol strength and half the price. The bar switched back and forth between full-strength nights and low-alcohol nights. Staff measured how many drinks each person bought and took breath samples to check blood alcohol levels.

On almost every low-alcohol night, the four patrons drank fewer total ounces of alcohol. Their breath alcohol readings were also lower. Cutting the drink strength in half reliably cut their intake and impairment.

Whitehouse et al. (2014) and Silva et al. (2020) used the same flip-flop design in classrooms. They compared token earn versus token loss, and adding versus removing game tokens. All three studies show that quick A-B-A-B comparisons work in real-world settings.

Pliskoff et al. (1972) and Siegel et al. (1970) ran early token economies with adults and preschoolers. Like the bar study, they tracked simple counts—tokens spent or handwriting strokes—to judge if the program worked. The tactic of watching natural behavior counts is decades old.

Meuret et al. (2001) tracked who dropped out of a hospital token system. Their mixed results remind us that even simple behavioral plans need steady data. The bar study had only four patrons, so watching for early drop-outs matters there too.

You can shrink problem behavior by shrinking the reinforcer. Less alcohol in the glass meant less drinking, just like smaller token pay-offs can cut disruptive acts. If you run a token economy, try cutting the size or value of the backup reinforcer first. You might keep the same behavior change with lower cost and fewer side effects.