In multilevel modelling, what does it mean if the intercept is fixed?

In multilevel modelling, what does it mean if the intercept is fixed?

• A. Intercept is fixed and does not vary across groups.
• B. Intercept varies across groups.
• C. It is the slope that varies.
• D. The model has no intercept.

Answer: (A)

In multilevel modeling, saying the intercept is fixed means there is a single intercept parameter that applies across all groups; the baseline level of the outcome does not vary by group. For example, if groups are schools, the expected starting point when predictors are zero is the same for every school. If the intercept were allowed to vary, you’d have a random intercept model, meaning each group could have its own baseline. The other ideas—intercept varying or the slope varying—refer to different model specifications (random intercept vs random slopes). Having no intercept is a separate modeling choice and not what fixed intercept implies.

In multilevel modeling, saying the intercept is fixed means there is a single intercept parameter that applies across all groups; the baseline level of the outcome does not vary by group. For example, if groups are schools, the expected starting point when predictors are zero is the same for every school. If the intercept were allowed to vary, you’d have a random intercept model, meaning each group could have its own baseline. The other ideas—intercept varying or the slope varying—refer to different model specifications (random intercept vs random slopes). Having no intercept is a separate modeling choice and not what fixed intercept implies.