Higher prior fatigue was associated with lower current motivation log-odds, with significant effects for both the linear ( OR = 0.01, β = −4.97, SE = 2.73, p = 0.069) and quadratic terms ( OR = 0.01, β = −4.43, SE = 2.36, p = 0.061), indicating a non-linear relationship with approaching significance at the p < 0.05 level.
In terms of hypothesis testing, the group comparisons did not reach statistical significance at the reference point of time passed (the midpoint of the study session).
Non-significant terms were removed unless they were main effects involved in significant or marginally significant interactions.
Tests of the proportional odds assumption (using a fixed-effects equivalent model) indicated no significant violations overall, although lagged fatigue ( p = 0.064) and lagged motivation ( p = 0.053) showed marginal significance, suggesting potential deviations, particularly at the lowest threshold where data might be sparser.