There’s a mean (not really) but funny (maybe only to me) trick I like to play on my students that starts with a lesson on an experiment. It works for almost any kind of cognitive psychology study, and I start it by walking my students through the design of an experiment—the research question, important variables, conditions, measures, etc.—before asking, “Has anyone ever participated in a study like this before?” When I’m lucky enough to have at least one student raise their hand, I’ll turn to them, wearing the best facsimile of excitement as I can, and ask, “How was it?” In the best outcomes of this gag, you can almost see the gears of diplomacy turning in their heads as they try to figure out the politest way to answer. “You can say it was boring,” I’ll tell them, and once I see their smile of relief, I immediately hit them with, “I’m going to tattle on you, though.”

Of course, I always follow up with reassurances that I’m 1) kidding (“Tattling? At this age? To whom??”) and 2) fully aware that many psychology studies are just kind of boring. And then, in a way that is both serendipitous and pedagogically valuable, what starts out as a gag for my own amusement often turns into a discussion about how a participant’s overall mood and level of motivation could potentially affect the quality of their data. More broadly, is tedium the necessary price for scientific control? Or are there ways to collect high-quality data while also making experiments an enjoyable, dare I say even fun, experience?

This brings us to a topic nearly as fun as teaching undergrads: Video games.

Drew J. McLaughlin and Arthur G. Samuel (pictured below), authors of the featured article in this post, say that we can make experiments that are “as equally useful for researchers as [they are] enjoyable for the participant”. They join a growing movement of experimental gamification, where studies can be made to feel more like a video game (and therefore become more fun and interesting) by adding features like nicer graphics, competition, and narrative explanations for experimental tasks. In their recent Behavior Research Methods paper, which has the excellent title of “Lettuce entertain you: Assessing Sandwich Builder as a measure of auditory short‑term memory”, they introduce and evaluate a game that, as one might guess, involves building sandwiches for the purpose of assessing short-term memory.

Specifically, players of Sandwich Builder take on the role of a drive-thru restaurant worker (or should we say “sandwich artist”? Probably not, since the game presumably doesn’t reflect any real chains). After listening to a customer order a sandwich with a list of ingredients, players click on the various ingredients in the order they were said by the “customer,” who is actually an algorithmically generated audio file built from recordings of real voices. Once they think the sandwich is ready, players click a button reading, “BUILD IT”. I imagine that this part is particularly satisfying.

The classic experiment being gamified by Sandwich Builder is what’s called a forward recall task, where participants are presented with a sequence, typically of words or numbers, and then must recall the sequence in the correct order, usually by saying it or typing it on a computer. Instead of numbers or unrelated words, Sandwich Builder presents a sequence of ingredients. Instead of reciting the sequence back, players click on ingredient icons, and the whole endeavor is (deli-)wrapped in fun graphics and a pretend setting.

But of course, the authors needed to confirm that Sandwich Builder could assess short-term memory at least as well as traditional methods. To do this, they had online participants play Sandwich Builder and also complete some traditional variations of forward recall tasks. These tasks included video and audio versions of a digit span task, where participants were presented with sequences of numbers that they then had to type into a response box in the correct order. Participants also completed video and audio versions of a free recall task, where longer sequences of words were presented very quickly, and the participants then had to recall as many words as they could.

For Sandwich Builder to work as a feasible assessment of short-term memory, players’ game scores need to reflect their short-term memory abilities in a similar way to the other tests. And indeed, as shown in the figure below, Sandwich Builder scores were significantly correlated with the outcomes of all four traditional methods, which strongly suggests that it can provide estimates of short-term memory that are just as sensitive as these other tests.

To see if Sandwich Builder also provided the emotional advantages of gamification, the authors had their player participants complete questionnaires about mood, tiredness, and motivation after each task. As shown in the figure below, regardless of what order they completed the tasks or played the game and played the game, participants reported that they were more motivated, less tired, and in a more positive mood after playing Sandwich Builder than after any of the more traditional paradigms.

On top of its consistency with traditional measures as well as its advantages for participant experiences, the authors also wanted to see if players’ Sandwich Builder scores remained consistent across play sessions and whether they could predict later performance on other kinds of memory tests. Evidence of this kind of prediction validity helps to demonstrate Sandwich Builder’s ability to predict players’ skills in other areas while also reconfirming the game’s use as a measure of short-term memory.

To do this, the authors asked their player participants to return for another session, where they once again played Sandwich Builder and also completed two additional tasks where they transcribed two types of audio recordings, one being a speech recording distorted by artificial noise and the other being a recording of accented speech from a second-language speaker. As shown in the figure below, not only were player scores fairly reliable across play sessions, but they also corresponded to the accuracy of participants’ skills for both transcription tasks.

As final confirmation of their gamified experiment’s utility, the authors also explored the possibility that Sandwich Builder scores might correlate with something unrelated to short-term memory. For this test, they hoped to NOT find such a correlation; after all, a good test should only reflect one (or very few) processes. If it reflects many unrelated skills, then it becomes extremely difficult, maybe even impossible, to know whether game scores are driven by short-term memory or some other unrelated skill. Fortunately, when the authors compared player scores to an unrelated personality measure of extraversion, there was no correlation between players’ tendencies for extraversion and their elite gamer skills (by which I mean their Sandwich Builder scores).

The outcome of McLaughlin and Samuel’s evaluation is a new kind of short-term memory assessment that affords many, if not all, of the strengths of traditional measures while also granting the advantages of just being more enjoyable for participants, which in turn could lead to higher-quality data. The best part for anyone interested in using Sandwich Builder for their own studies is that you won’t even have to wait in a proverbial deli line; the game is already available for free via the Gorilla experiment platform.

Featured Psychonomic Society paper

McLaughlin, D.J., & Samuel, A.G. (2025). Lettuce entertain you: Assessing Sandwich Builder as a measure of auditory short-term memory. Behavior Research Methods, 57, 191. https://doi.org/10.3758/s13428-025-02707-1