Spatio-temporal Rasch analysis of quality of life outcomes in the French general population. Measurement invariance and group comparisons

Table 3 Estimation of the parameters associated with the covariates explaining the mean of the latent trait “General Health” (β _μ ) and with the covariates explaining the variance of the latent trait (γσ ² )

	Effect	Estimate	Standard error	p-value
Variables explaining the mean of the latent trait (β_μ)	Constant (μ*)	0.8238	0.0415	<0.0001
	Women	−0.2121	0.0267	<0.0001
	Survey 2003	−0.1393	0.0413	0.0007
	Ager	−0.0349	0.0019	<0.0001
	East	−0.2039	0.0480	<0.0001
	West	0.1259	0.0344	0.0003
	South-West	−0.0666	0.0301	0.0272
	Ager*Women	0.0049	0.0013	0.0002
	Ager*sample2003	−0.0030	0.0018	0.1047
	Ager* (North, West)	−0.0046	0.0013	0.0004
	Ager* (East)	0.0062	0.0024	0.0096
	Survey 2003*(North, Eastern Paris Basin)	−0.1172	0.0233	<0.0001
	Survey 2003*(Western Paris Basin, South-East)	−0.0675	0.0266	0.0113
Variables explaining the variance of the latent trait (γσ²)	Constant (σ²*)	1.4130	0.0791	<0.0001
	Survey 2003	−0.1499	0.0837	0.0730
	Ager	−0.0068	0.0034	0.0439
	East	−0.2334	0.0513	<0.0001
	South-West	−0.1254	0.0534	0.0188
	Ager*Survey 2003	0.0068	0.0037	0.0632
Explained variance	Estimated variance without covariates: ${\hat{σ}}_{0}^{2}$ (k₀)	1.4272 (37)
	Estimated variance with covariates: ${\hat{σ}}^{2}$ (k)	1.2406 (50)
	Rate of explained variances: _τ	13%

Ager is a categorical variable taking values 0 for the 18–24, 2 for the 25–34, 12 for the 35–44, 17 for the 45–54, 24 for the 55–64, 32 for the 65–74, and 40 for the 75–84 year-olds.
Women is a variable coded 1 for women and 0 for men.
Survey 2003 is coded 1 for the year 2003 and 0 for year 1995.
Region names are dummy variables coded 1 for the cited region and 0 for other geographical areas.

ISSN: 1471-2288