Skip to main content
. 2018 Dec 1;241:348–355. doi: 10.1016/j.jad.2018.07.078

Table 3.

Results of p-values for partial f-tests, comparing each represented life course models to a saturated model.

ACE-III
Verbal memory
Letter search speed
Letter search accuracy
Hypothesis f p f p f p f p
Saturated model Reference Reference Reference Reference
No effect 2.42 0.02 1.86 0.04 2.11 0.03 1.86 0.04
Accumulation 0.91 0.49 1.63 0.13 0.82 0.55 0.91 0.49
Time period
t1 (ages 13–15) 2.81 <0.01 1.78 0.11 1.19 0.31 1.51 0.17
t2 (ages 36–53) 1.96 0.07 1.45 0.19 1.12 0.36 1.72 0.07
t3 (ages 60–69) 1.46 0.19 1.37 0.22 0.98 0.44 1.27 0.27
Best model* Accumulation Sensitive period t3 (ages 60–69) Accumulation Accumulation

To reduce multi-collinearity of repeated measures, categories at three age spans were chosen to represent case-level affective symptoms in the life course; adolescence (ages 13–15), adulthood (ages 36–53) and later-life (ages 60–69). A higher P-value (or lower f statistic) for the life course model equals a better model fit. The ‘saturated model’ is the most complicated model that contains affective case-level symptoms at all three ages, all two-way interactions and the three-way interaction. The ‘accumulation’ model proposes that the impact of exposure is cumulative over the life course and that the longer an individual is exposed to case-level symptoms, the greater the adverse impact on cognitive function at age 69. A ‘time period’ model proposes that exposure to case-level symptoms during a particular stage in life (e.g., t1, t2 or t3) has an adverse effect on cognitive function at age 69 with little or no influence of exposure to symptoms outside this specified time period. Bold values indicate p-value > 0.05.

*The best model in this case was determined by the lowest f value.