Skip to main content
. 2021 Oct 19;18:136. doi: 10.1186/s12966-021-01204-w

Table 4.

Multilevel regression model predicting physical activity behaviour

Variable Model ICC -2ll (dfp-value) u0j u2j e0ij β(SE) p-value
Null Model .46 3649.48 2.04 2.37 3.42(.08)
Model 1 3560.32 (df = 15 p = .001) 1.29 2.54
Model 2 3558.84 (df = 15p = 1.00) 1.29 .021 2.42
Intercept (constant) 3.55(.169) .001
Age .004(.011) .36
Multiple deprivation index .017(.027) .26
Mental health and wellbeing −.016(.012) .09
Social isolation −.019(.165) .45
Quality of Life: Physical Function −.468(.195) .008
Quality of Life: Feelings −.187(.196) .17
Quality of Life: Daily Activities −.142(.213) .25
Quality of Life: Social Activities .207(.216) .17
Quality of Life: Pain .048(.191) .40
Quality of Life: Changes in Health Status −.018(.218) .46
Quality of Life: Health Status −.548(.198) .002
Quality of Life: Social Support −.061(.197) .37
Quality of Life: Overall perceptions .557(1.41) .34
Employment Status (reference: not employed vs. employed) −.093(.183) .30
Partnership (reference: alone vs. partner) −.148(.18) .20

Note: ICC = interclass correlation. -2ll = −2*loglikelihood (deviance in IGLS estimation). Region level (u0j = intercept, u2j = slope). Member level (e0ij = intercept