A comparative study of R functions for clustered data analysis

Table 2 CI’s of the fixed effects and the variance-covariance parameters

	Intervention effect	ρ	σ² or σ	σ_u	σ_ε
geese	2.161 (1.060, 3.263)	-0.076 (-0.125, -0.027)	101.768 (93.518, 110.019)
gls	2.252 (0.597, 3.907)	0.012 (3.7×10⁻⁴, 0.405)	10.093 (9.727, 10.473)	1.107	10.032
	2.252 (0.595, 3.908)
lme	2.252 (0.602, 3.901)	0.012	10.093	1.107 (0.589, 2.081)	10.032 (9.669, 10.409)
lmer	2.252 (0.604, 3.900)	0.012	10.093	1.107 (0.234, 2.111)	10.032 (9.673, 10.414)
	2.252 (0.525, 4.086)

The “geese” method finds the point estimate and CI of the parameter σ² instead of σ as the other methods. The order of the CI’s of the intervention follows the order of the CI’s of β in Table 1

ISSN: 1745-6215