Skip to main content
. 2008 Jun 6;38(4):348–360. doi: 10.1007/s10519-008-9212-5

Table 5.

Model fitting results for an interaction model of full scale IQ with mean real estate price as moderator

Mean real estate price −2LL Df Inline graphic Dfdiff P
Model 1 Full model 5588.163 723
          1a Equalize unmoderated parts of variance (a, c, an e) across sex and cohort 5602.742 732 14.579 9 0.10
          1b Equalize unmoderated part means (m) across sex and cohort 5607.506 735 4.764 3 0.19
          1c Drop parameter c 5607.507 736 0.001 1 0.97
Model 2 Equalize moderation parameters a′, c′ and e′ (variances) across sex and cohort 5611.287 745 3.781 9 0.93
Model 3 Drop all moderation on the variances 5616.798 748 5.511 3 0.14
Model 3a Drop all but moderation in E of older males (Inline graphic) 5611.774 747 5.024 1 <0.05
Model 4 Equalize moderation parameter means (m′) across sex and cohort 5613.055 750 1.281 3 0.73
Model 5 Drop moderation means 5616.713 751 3.658 1 0.06
Parameters final model
Means m = 103.7412 Variances A = 10.4655
m′ = 0.38 E = 4.8456
e′om = 0.4649

Note: Parameters a, c and e refer to the unmoderated parts of the variance components A (additive genetic effects), C (common environmental effects) and E (unique environmental effects) of FSIQ, while parameters a′, c′ and e′ refer to the moderation parameters of these variance components. Parameter m refers to the unmoderated part of the means model, while parameter m′ refers to the moderation effect in the means. Subscripts refer to young (y) or old (o) and males (m) or females (f), respectively