Parameter estimation and model fit for “univariate” latent change score models of anxiety symptom severity, affect and communal behaviour
Measure | Social anxiety | Positive affect | Negative affect | Agreeable behaviour | Quarrelsome behaviour |
---|---|---|---|---|---|
Parameter, B (95% CI) | |||||
β* | −0.56 (−0.68 to −0.44)‡ | −0.49 (−0.80 to −0.11)§ | −0.49 (−0.66 to −0.29)‡ | −0.77 (−1.06 to −0.42)‡ | −0.79 (−1.27 to −0.24)§ |
Mint | 5.81 (5.62 to 5.99)‡ | 2.18 (1.88 to 2.48)‡ | 1.33 (1.11 to 1.55)‡ | 1.05 (0.78 to 1.33)‡ | −1.44 (−1.69 to −1.20)‡ |
s2int | 0.43 (0.38 to 0.90)¶ | 0.85 (0.77 to 1.42)‡ | 0.52 (0.34 to 0.84)‡ | 0.50 (0.19 to 0.99)¶ | 0.52 (0.27 to 0.92)‡ |
Mslope | 2.26 (1.63 to 2.92)‡ | 1.50 (0.52 to 2.34)¶ | 0.19 (−0.02 to 0.40) | 1.42 (0.88 to 1.95)‡ | −1.36 (−2.20 to −0.47)¶ |
s2slope | 0.36 (0.32 to 0.63)‡ | 0.36 (0.32 to 0.95)¶ | 0.08 (0.03 to 0.16)‡ | 0.49 (0.44 to 1.04)‡ | 0.37 (0.33 to 0.97)‡ |
Covariance (int. with slope) | 0.08 (−0.06 to 0.25) | 0.47 (0.12 to 0.95)§ | 0.06 (−0.05 to 0.18) | 0.27 (0.03 to 0.63)§ | 0.26 (0.03 to 0.62)§ |
ea† | — | 0.23 (0.16 to 0.32) | 0.08 (0.06 to 0.11) | 0.40 (0.29 to 0.57) | 0.22 (0.16 to 0.32) |
Model fit | |||||
χ2 | χ24 = 2.49 p = 0.64 | χ27 = 6.99 p = 0.43 | χ27 = 14.86 p = 0.04 | χ27 = 4.74 p = 0.69 | χ27 = 10.03 p = 0.19 |
CFI | 1.01 | 1.00 | 0.90 | 1.05 | 0.95 |
RMSEA (95% CI) | 0 (0 to 0.20) | 0 (0 to 0.20) | 0.15 (0 to 0.28) | 0 (0 to 0.16) | 0.09 (0 to 0.23) |
CFI = comparative fit index; CI = confidence interval; Covariance (int. with slope) = covariance of the mean intercept (baseline severity) with the mean slope describing the average linear change over time; ea = error variances; Mint = the mean intercept (baseline severity); Mslope = the mean slope describing linear change over time; RMSEA = root mean square error of approximation; s2int = between-person variance of the intercept; s2slope = between-person variance of the slope describing linear change over time.
↵* Similar to multilevel models or latent growth curve models in other structural equation modeing approaches, β represents the regression of change over each month on prior level.
↵† Significance tests were not conducted for error variances; see Data Analysis for information on the time invariance of anxiety symptom severity.
↵‡ p < 0.001.
↵§ p < 0.05.
↵¶ p < 0.01.