fix

Author: strengejacke

vignettes/practical_context_effect.Rmd

@@ -179,11 +179,11 @@ estimate_contrasts(mixed, c("phq4_within", "phq4_between", "time"))
 
 This pairwise comparison table adds a crucial statistical caveat to our visual and descriptive observations. The Difference column here represents the mathematical change in the size of the context effect between two specific time points (e.g., the context effect grew by 0.99 points from Time 1 to Time 2).
 
-However, looking at the statistics, we can see that none of the differences between the time points are statistically significant (all p-values = 0.183).
+However, looking at the statistics, you will notice that the estimated difference between any two adjacent time points is exactly identical, yielding the exact same standard error and p-value.
 
-This lack of significant differences between time points is not surprising, as our model assumes a *strictly linear time trend*. Under a linear assumption, the rate of change (the slope) is constrained to be constant from one time point to the next. These point-to-point comparisons would only differ if we had modeled time non-linearly (e.g., by adding a quadratic term).
+This uniformity is a direct mathematical consequence of our model assuming a *strictly linear time trend*. Under a linear assumption, the rate of change (the slope) is constrained to be constant. Therefore, the estimated change in the context effect from Time 1 to Time 2 is forced to be exactly the same as the change from Time 2 to Time 3. These point-to-point comparisons would only vary if we had modeled time non-linearly (for instance, by adding a quadratic term).
 
-Therefore, to test whether the context effect meaningfully changes over time in a linear model, it is more appropriate to evaluate the overall average contrast of the slopes across the entire study period. To do this, we calculate the contrast between the within- and between-effects without stratifying by time.
+Because a linear model yields constant step-by-step changes, testing these identical individual intervals is often less informative. To test whether the context effect meaningfully changes over time overall, it is more appropriate to evaluate the average contrast of the slopes across the entire study period. To do this, we calculate the contrast between the within- and between-effects without stratifying by time.
 
 ```{r}
 estimate_contrasts(mixed, c("phq4_within", "phq4_between"))