What this page shows
Every brain test on this site keeps its runs in this browser. This page reads those histories and shows, for each task you have run, one number from its test page: in that page's unit, with its 95% interval where the page computes one, and with the date and device of the run. Below it, your last 20 runs on one device form a chart, so a trend is visible without opening each test.
For every test but one, that number is the one its page reports as the main result. The Attention Span Test gives response speed, lapses and false starts as its main numbers, none of them with an interval, so the profile shows its median reaction time, which the page gives with one.
Practice rounds are left out. A task you stop before its end is not saved by its test page, so it never reaches the profile, with three exceptions that the profile counts: Dual N-Back keeps a session ended early with its finished blocks, the Mental Rotation Test keeps a finished first block as a short run, and the Verbal Memory Test keeps the recall of a False memory run you stop during its recognition test. Dual N-Back saves a session after every block and replaces the earlier copy of it; should an older copy remain, the profile still counts the session once.
The number shown for each test
| Test | Tasks | Number shown |
|---|---|---|
| Attention Span Test | 3 and 10 minutes | median reaction time, ms, with its 95% interval; the page's main number is response speed, which has no interval |
| Attentional Blink Test | both letters; second letter only; brief display | blink size, %; dip at lags 2 and 3, percentage points; letters available at 0 ms |
| Change Blindness Test | flicker; no blank | background details minus main objects, s; median time, s |
| Dual N-Back | dual; position; letter | mean n of the session |
| Go/No-Go Test | Go/No-Go; SART; stop-signal | commission errors, %; SSRT, ms |
| Memory Test | digits, letters, blocks | mean span (MS); block span |
| Mental Rotation Test | in depth; picture plane | slope, ms per degree |
| Multiple Object Tracking | standard; speed threshold; training | capacity m; speed in dot widths per second, for each number of targets |
| Number Sense Test | A; B | Weber fraction w, marked unreliable fit when the fit ended at the limit of its search or the interval is over 1.5 times as wide as w; subitizing range from 0 to 9 dots, marked unreliable fit when its curve explains under 70% of the differences in your times; both as on the test page |
| Stroop Test | Stroop; flanker; Simon | effect as a trimmed mean, ms |
| Task Switching Test | task switching; trails | switch cost, ms; B/A ratio |
| Time Perception Test | short and long intervals; tapping | regression index ρ; SD of asynchrony, ms |
| Verbal Memory Test | serial position; false memory | studied words written, %; hidden words written, % |
| Visual Memory Test | change detection | capacity Kmax |
| Visual Search Test | feature; conjunction; 2 among 5s | search slope with the target present, ms per item |
Each number keeps the meaning it has on its own page, explained there. Hidden words written in the false memory task is a share, not a score: writing more of them is neither better nor worse.
Devices
Your input and your screen add a delay of their own to every time, so the profile groups runs by device: the same kind of input (keyboard, mouse, touch or pen) and refresh rates within 10% of each other, or both below 50 Hz, the rule the test pages use for their own history. The chart, the table of runs and the comparison of first and last runs use one group; the buttons above the chart switch between groups. A run is marked less precise when its own test page warned about the timing of that run: a screen below 50 Hz or with uneven frames, or too many trials mistimed while the computer was busy.
Why there is no overall score
The tasks measure different things in different units, and one run of a task measures a person only roughly. For example, Hedge, Powell and Sumner (2018) tested people twice, three weeks apart: for the Stroop effect, the intraclass correlation between the two sessions was 0.60 in one study and 0.66 in the other. The guide the paper cites counts 0.6 as good reliability, and even at that level people did not keep the same order from one session to the next. A score made by adding such numbers together, or a rank against other people, would add up that chance as well, so the profile shows each task on its own and compares it only with your own earlier runs.
Practice and change over time
From ten runs of a task on one device, the profile sets the mean of your last five runs beside the mean of your first five. A difference there need not be a change in you: you also learn a test by repeating it. Woods and colleagues (2011) tested 30 young adults three times, 5 to 11 days apart: their mean digit span rose by 0.30 digits forwards and 0.45 digits backwards from the first to the third session.
Things to try
- On a computer, run the Memory Test twice in this browser, once typing on the keyboard and once clicking its keypad, then switch between the two device groups above the chart. On a phone every run is a touch run, so it stays one group.
- Run the same test on ten different days, then read the comparison of your first and last five runs: how much of the change could be practice?
- Download the CSV and open it in a spreadsheet: each row is one run, with its value, interval and device.
Limits
- This browser only. Results stay in the browser that ran the test. Another browser, a private window or another device keeps its own history, and clearing this site's data in the browser deletes all of them.
- A limited history per test. Each test page keeps at most its newest 300 runs, and fewer when its history grows too large for storage; the oldest go first. Download JSON keeps a full copy.
- Nothing is deleted here. To delete runs, use Clear history on the test page.
- Not a medical test. None of these numbers is a diagnosis or a measure of intelligence.
Controls
- Buttons: a task under Your tests shows its last result, chart and runs; the device buttons above the chart switch the device group; Download JSON saves every history as the test pages stored it, Download CSV one row per run.
- Keyboard: Tab moves between the controls, Enter or Space presses one.
- Touch and mouse: tap or click a task or a device.
- Language: the link at the end of Export switches the app to Slovak. Without a choice of your own, the profile uses the language of your newest run.
Sources
- Hedge, C., Powell, G., & Sumner, P. (2018). The reliability paradox: why robust cognitive tasks do not produce reliable individual differences. Behavior Research Methods, 50, 1166–1186. Full text (PMC5990556)
- Woods, D. L., Kishiyama, M. M., Yund, E. W., Herron, T. J., Edwards, B., Poliva, O., Hink, R. F., & Reed, B. (2011). Improving digit span assessment of short-term verbal memory. Journal of Clinical and Experimental Neuropsychology, 33(1), 101–111. Full text (PMC2978794)