Comparison of observed average total time on task (TOT, min) with that predicted under length biased sampling and no additional penalty due to interruptions
| Observation sessions | Average interruption rate per hour | No of interruptions | Sample size (n) | Average TOT (min) predicted under length-biased sampling (A) | Observed average TOT (min) (B) | Percentage observed mean to predicted (B/A%) | Z* |
| 1 to 15 | 15.4 | 1 | 145 | 4.50 | 2.02 | 45 | –5.0 |
| 2 | 40 | 12.60 | 2.47 | 20 | –6.4 | ||
| ≥3† | 42 | 18.63 | 4.98 | 26 | –11.2 | ||
| 16 to 66 | 7.7 | 1 | 263 | 5.00 | 3.33 | 67 | –4.8 |
| 2 | 92 | 11.17 | 5.45 | 49 | –7.0 | ||
| ≥3† | 59 | 16.75 | 6.80 | 41 | –9.8 | ||
| 67 to 131 | 3.3 | 1 | 204 | 5.88 | 3.28 | 56 | –5.3 |
| 2 | 33 | 14.23 | 7.08 | 73 | –3.6 | ||
| ≥3† | 13 | 23.07 | 6.27 | 27 | –4.6 | ||
| All | 6.2 | 1 | 612 | 5.41 | 3.00 | 55 | –9.3 |
| 2 | 155 | 13.03 | 5.25 | 40 | –9.6 | ||
| ≥3† | 114 | 20.87 | 6.05 | 29 | –13.4 |
↵* All test statistics reject the null hypothesis of no impact due to interruptions with p-values less than 0.0005 in all cases. The negative Z values indicate that the observed TOTs for interrupted tasks were significantly lower than expected.
↵† The predicted average TOT for three or more interruptions is calculated using the formula for three interruptions.