Based on the results in Table 1 above, what might be a recommendation for improving the internal validity of the stepping test?
. Measurement Internal and External Validity
Below is an example of a study to determine the reliability of a new “step test” to measure functional ability. The step test involved having each subject step forward as far as they could with one leg three times. The time (seconds) it took to perform the 3 repetitions was recorded. 241 healthy subjects aged 18-65 completed the study. One rater tested all subjects. Distance (cm) is a continuous variable so ICC and paired t-test was used to compare reliability between trials.
Table 1. Average Time (seconds) and Standard Deviation (SD) by Trial and p-values between Trials.
| Trial 1 mean seconds (SD) | 5.72 (1.57) |
| Trial 2 mean seconds (SD) | 5.29 (1.45) |
| Trial 3 mean seconds (SD) | 5.06 (1.37) |
| Paired T-test, Trial 1 vs 2 p-value | <.0001 |
| Paired T-test, Trial 1 vs 3 p-value | <.0001 |
| Paired T-test, Trial 2 vs 3 p-value | <.0001 |
- Based on mean changes from trial 1 to trial 3, and the paired-test p-values, describe internal validity concerns in terms of the type of bias that may be present?
- Based on the results in Table 1 above, what might be a recommendation for improving the internal validity of the stepping test?
