Standardized parameters (betas) of mixed effects Poisson and Ordinal Probit regression models of activity and participation limitations and general health perception on selected independent variables.

*Bonferroni corrected (p*4) probabilities *p‡Intercepts (in standard units) were −43.48*** for the threshold between “very bad" and the rest; −35.18*** for “very bad" and “bad" vs. the rest; −21.60*** for “very bad", “bad", and “fair" vs. the rest; 10.51*** for all other categories vs. “very good".†Intercepts (in standard units) were −12.09*** for the threshold between “very bad" and the rest; −12.98*** for “very bad" and “bad" vs. the rest; −8.86*** for “very bad", “bad", and “fair" vs. the rest; 3.99*** for all other categories vs. “very good". Intercepts (in standard units) were −22.83*** for the threshold between “very bad" and the rest; −18.30*** for “very bad" and “bad" vs. the rest; −12.42*** for “very bad", “bad", and “fair" vs. the rest; 4.72*** for all other categories vs. “very good".┐ Intercepts (in standard units) were −15.19*** for the threshold between “very bad" and the rest; −9.13*** for “very bad" and “bad" vs. the rest; −2.07 for “very bad", “bad", and “fair" vs. the rest; 10.61*** for all other categories vs. “very good". OProbit = Ordinal Probit.Note: all models have been adjusted for age and and health behaviours (alcohol consumption, smoking, physical activity).