Audit intermediate Works with: Any AI

Usability Metrics Calculation Check

Table output Has example

Recompute task success, time and satisfaction figures from raw data.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

Verify the numbers in a usability report. Raw data by participant and task, with completion, time in seconds and any rating: [RAW_DATA]. Figures reported in the document: [REPORTED_FIGURES]. Number of participants: [PARTICIPANT_COUNT].

Recalculate the success rate as participants completing divided by participants attempting, the median time (and mean for comparison), and the average rating, showing every step. Compare with the reported values and mark each as matching, close or wrong. Check that failed tasks are handled consistently in the time calculation, that outliers are explained, and that percentages are not quoted from tiny groups without a count. Add an approximate interval for each success rate using 1.96 times the square root of p times 1 minus p divided by n, and say what it implies.

Return a table of recalculated against reported values, corrected statements, and a note on the limits of small samples.

Add a worked example with invented raw data for ten participants on one task. Completed: P1 yes 62 seconds, P2 yes 48, P3 no, P4 yes 95, P5 yes 51, P6 no, P7 yes 58, P8 yes 71, P9 yes 300, P10 yes 55. Successes = 8 of 10, so the success rate is 80 percent. Times of successful participants sorted: 48, 51, 55, 58, 62, 71, 95, 300; median = (58 + 62) divided by 2 = 60 seconds; mean = (48 + 51 + 55 + 58 + 62 + 71 + 95 + 300) divided by 8 = 740 divided by 8 = 92.5 seconds, pulled up by one 300-second outlier, so the median describes the typical time better. Approximate margin of error for 80 percent with n = 10: 1.96 times the square root of 0.8 times 0.2 divided by 10 = 1.96 times 0.126 = about 25 points, so the plausible range is roughly 55 to 100 percent. If the report said 90 percent success or a mean of 60 seconds, mark it as wrong and give the corrected statement. Add checks for how failed tasks are treated, for rounding, and for whether the satisfaction scale was scored consistently, such as a single ease question from 1 to 7 averaging 5.4. Provide the corrected sentence: 8 of 10 participants completed the task; the typical time was about a minute; the sample is small, so results are indicative.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

Invented data, ten participants. Completed: P1 62 s, P2 48, P3 no, P4 95, P5 51, P6 no, P7 58, P8 71, P9 300, P10 55.
Success = 8 of 10 = 80 percent.
Sorted times: 48, 51, 55, 58, 62, 71, 95, 300. Median = (58 + 62) / 2 = 60 s. Mean = 740 / 8 = 92.5 s, pulled up by one 300 s outlier.
Margin of error = 1.96 x sqrt(0.8 x 0.2 / 10) = about 25 points, so roughly 55 to 100 percent.
Report said 90 percent success and a mean of 60 s: both wrong. Corrected: 8 of 10; typical time about a minute.

• audit