Table 4.
ANOVA Model-Adjusted Differences in Latency between TBI and Sham, by Day and Swim Per Day
| Day | Swim | Estimate | SE | P-value | Hommel p-value |
|---|---|---|---|---|---|
| 1 | 1 | 2.30 | 7.41 | 0.7566 | 0.7566 |
| 1 | 2 | 6.85 | 7.41 | 0.3575 | 0.7150 |
| 1 | 3 | 12.56 | 7.41 | 0.0932 | 0.4255 |
| 1 | 4 | 13.55 | 7.41 | 0.0704 | 0.3522 |
| 2 | 1 | 8.48 | 7.41 | 0.2553 | 0.5363 |
| 2 | 2 | 9.92 | 7.44 | 0.1850 | 0.5363 |
| 2 | 3 | 17.37 | 7.44 | 0.0214 | 0.1714 |
| 2 | 4 | 19.48 | 7.44 | 0.0101 | 0.0912 |
| 3 | 1 | 22.39 | 7.41 | 0.0032 | 0.0318 |
| 3 | 2 | 27.06 | 7.41 | 0.0004 | 0.0045 |
| 3 | 3 | 26.74 | 7.41 | 0.0005 | 0.0053 |
| 3 | 4 | 13.53 | 7.41 | 0.0708 | 0.3539 |
Hommel p-values adjust for multiple testing. This model ignores censoring associated with truncation of swim times at the 120-sec maximum, so these estimates should be considered biased. Bold = p < 0.05.
ANOVA, analysis of variance; SE, standard error; TBI, traumatic brain injury.