In multilevel modeling, what does a random slope mean?

• A. It indicates a non-linear relationship.
• B. The intercept varies across groups.
• C. The slope is fixed across all groups.
• D. The slope of the model is free to vary across different groups or contexts.

Answer: (D)

In multilevel modeling, a random slope means the effect of the predictor on the outcome is allowed to differ across groups. Rather than assuming one universal slope, you estimate group-specific slopes that vary around an overall average. This captures how the relationship between the predictor and the outcome can be stronger in some groups and weaker in others, reflecting context-dependent effects.

The best description is that the slope of the model is free to vary across different groups or contexts, because that statement directly conveys that the slope itself is not constant but can differ by group.

Why the other ideas don’t fit as well: having the intercept vary describes random intercepts—differences in the starting point across groups, not the effect of the predictor. A non-linear relationship would involve the shape of the relationship changing, not necessarily the slope varying by group. Saying the slope is fixed across all groups contradicts the idea of a random slope, which by definition allows the slope to vary.