Table 4.

Model fit for multiple regression and LASSO models based on training or test data.

Outcome	Model	RMSE	R 2
Pain interference	Multiple regression based on training data	0.78	0.38
Pain interference	Multiple regression based on test data	1.17	0.29
Pain interference	LASSO for test data	0.93	0.37
Activity engagement	Multiple regression based on training data	0.72	0.47
Activity engagement	Multiple regression based on test data	1.08	0.31
Activity engagement	LASSO for test data	0.95	0.32
Activity avoidance	Multiple regression based on training data	0.70	0.50
Activity avoidance	Multiple regression based on test data	1.01	0.35
Activity avoidance	LASSO for test data	0.85	0.37

Lower values for RMSE and higher values for R² denote better model fit. Model fit indices based on training data are only available for the multiple regression model and are known to be overly optimistic.

LASSO, least absolute shrinkage and selection operator; RMSE, root mean square error.