Sternberg Memory Scanning Test — Free Working Memory Search Speed Task (28 Trials)

The Sternberg task is a classic experimental paradigm for measuring how quickly you can search the contents of working memory. A memory set of 1 to 6 digits appears for 1.2 seconds, followed by a 1-second blank and then a single probe digit. You press Yes if the probe was in the set and No if it was not, as quickly and accurately as you can. Half of the probes were in the set and half were not.

In Saul Sternberg's 1966 experiment, reaction time grew by a nearly constant amount for each item added to the memory set, a finding interpreted as comparing items in memory one at a time. That is why the key metric here is not accuracy but the time added per item.

There are 28 trials in total with set sizes 1 to 6 evenly mixed. The headline score is accuracy in percent, shown with mean reaction time in milliseconds computed from correct trials only. The detailed result shows mean reaction time for each set size and the regression slope of reaction time on set size in ms per item, which is how much slower you respond, on average, for each additional item in memory.

The smaller the slope, the faster your memory search. Adults are generally reported to slow by roughly 30–40 ms per item, but 28 trials is a short run for estimating an individual slope precisely, so trends across repeated sessions are more meaningful than any single value. Accuracy of 90% or higher indicates stable searching; when accuracy is low, reaction time and slope become less reliable to interpret.

As a reaction-time test it is sensitive to your device and input method. Touch screens, mice and keyboards differ in latency, and browsers and display refresh rates can shift results by tens of milliseconds, so use the same device when tracking trends. Sleep loss and fatigue slow overall reaction time, while caffeine can speed it up slightly. The slope tends to increase gently with age, and the first few trials can be treated as a warm-up while you adapt to the rules.

Repeating the test on the same device at a similar time of day, a few weeks apart, lets you follow mean reaction time and slope over time. If accuracy is repeatedly low even when you feel well and you find it hard to hold on to what you just heard in daily life, consider a professional evaluation of attention and memory. This test is a self-check, not a diagnostic tool.