Teaching Mind-Wandering with the SART: A 45-Minute Class Plan

The SART (Sustained Attention to Response Task) is one of the easiest experimental tasks to run in an intro psychology, cognitive psychology, or educational psychology class. It takes four minutes, needs no sign-in, and produces an intuitive measure: the moments when a stray thought leaked into a response. With a group link, every student's result pools into an anonymous distribution you can put on screen right away. This guide covers what the task measures, how to structure a 45-minute session, and which numbers from which papers to cite.

What it measures

Digits from 1 to 9 flash one at a time, and students press a button for every digit except 3. Because almost every trial is a 'press', the response becomes automatic, and at the moment attention drifts away from the task the hand presses for the 3 as well. Those presses that should have been withheld (commission errors) are the core measure. Alongside them you look at mean reaction time and how uneven the reaction times are (variability).

That is exactly why Robertson and colleagues (1997) built the task. In both people with traumatic brain injury and healthy adults, the number of presses on the 3 tracked self-reported everyday attentional slips (where did I put the keys, what was I about to say). The starting point, then, is that four minutes in the lab reflect, to a useful degree, attentional failures in daily life.

The 45-minute session, step by step

1. Before class: create a group from the test page's 'Create a group link' (test = SART, retest = 4 weeks), and post the link in the class chat or on a slide. Students just tap it, no sign-in.
2. Opening, 5 min: ask 'How many times did your mind wander in the last 10 minutes?' and take a show of hands. The scatter of answers is today's question: how common is mind-wandering, and can we measure it?
3. Test, 4 min: everyone runs it quietly on their own device. Have them silence notifications and leave the result screen open when done (individual scores stay with each person).
4. Distribution, 5 min: open the group hub on the projector — anonymous histogram, median, and head count. Set the rule first: nobody announces where they landed.
5. Lecture, 15 min: automatic responding and 'decoupling' — the moment attention shifts from the external task to internal thought (Smallwood & Schooler, 2006). Use the numbers in 'Key studies' below as they are.
6. Discussion, 10 min: pick two of the questions below. It works well to separate the spread across people (why so different?) from the variation within a person (does the same person differ by time of day?).
7. Wrap-up, 5 min: announce the post-test in four weeks and ask each student to log one variable in the meantime (sleep hours, caffeine, time of day). On retest day, 'what moved with the score?' becomes the second lesson.

Key studies — numbers you can use as they are

• Robertson, Manly, Andrade, Baddeley & Yiend (1997) — the original SART paper, titled 'Oops!'. Presses on the 3 correlated with a questionnaire of everyday attentional failures and were more frequent after traumatic brain injury. The source for 'a lab measure reflects daily life'.
• Smallwood & Schooler (2006), 'The Restless Mind' — the theoretical paper that frames mind-wandering as attention decoupling from the external task and attaching to an internal stream. Cite it as the conceptual frame for SART errors.
• Cheyne, Solman, Carriere & Smilek (2009), 'Anatomy of an error' — the few responses right before a press on the 3 are actually faster. Errors do not arrive out of nowhere; they are the end of a gradual slide into autopilot. Point students to this when they look at the reaction-time trace in their own result.
• Smilek, Carriere & Cheyne (2010) — the ecological-validity paper. Reaction-time variability lines up with self-reported everyday attention lapses, which is the basis for saying that unevenness tells you more than average speed.
• Killingsworth & Gilbert (2010), 'A wandering mind is an unhappy mind' — a large experience-sampling study that pinged people on their phones at random moments. People were thinking about something other than what they were doing 46.9% of their waking time, and reported being less happy while doing so. This number ties straight back to the opening question.

Discussion questions

• Same four minutes, same task — why is our class distribution this wide? Ability, or state on the day? How could we tell the two apart?
• If responses speed up right before an error (Cheyne, 2009), how does that change what 'paying attention well' means? Is faster always better?
• If we spend nearly half our waking hours mind-wandering (46.9%), is it a defect or a function? How would you test its link to planning or creativity?
• What does it mean, and not mean, that a questionnaire of everyday slips correlates with a four-minute task? If the correlation is small, what conclusion survives?

Variations and extensions

• Prediction game: before the test, have students write down how many errors they expect, then compare with the actual result. The accuracy of metacognition (knowing your own attention) becomes the topic.
• Condition comparison: split the class — one half with music, the other in silence — and put the two groups' distributions side by side (create two group links). Random assignment becomes something they have done, not just heard about.
• Pre/post: at the four-week retest, have students note sleep hours or time of day, and you have material for a correlation exercise. Explain the practice effect first (the second run is usually slightly better).
• Measure debate: have students hypothesize which predicts everyday attention better — mean reaction time or reaction-time variability — then compare with Smilek (2010).

• 🧘Mind-wandering (SART)How often does your mind drift under monotony?Take the test →
• 🎧Auditory CPTTell high from low tones by ear and respond only to targets — auditory sustained attention.Take the test →
• 🛑Stop signalYour brain's braking distance (SSRT).Take the test →