Random slopes

Random slopes over some predictor p are

• the way the model adjusts the average (fixed) slope over p to better match the data from each level of the grouping variable(s).
• one number per level of the grouping variable, which is added on to the fixed slope over p to produce the group-specific slope over p.
• also called “varying slopes” or “slope adjustments”.
• represented by the + p in the random effect lme4 syntax: (1 + p | g), where g is the grouping variable.

To work out whether your model can contain random slopes over any of its predictors, see Identify possible random effects.

Linked flash cards

Outgoing links

• TODO

Backlinks

• TODO