Download presentation
Presentation is loading. Please wait.
Published byWidya Irawan Modified over 6 years ago
1
Factor model of Ordered-Categorical Measures: Measurement Invariance
Michel Nivard Sanja Franic Dorret Boomsma Special thanks: Conor Dolan OpenMx “team”
2
Ordinal factor model This model works for ordered categorical indicators. It assumes these indicators are indicators of an underlying continuous (and normally distributed) trait.
3
Ordinal indicators Q: do you get angry? 0: Never 1: Sometimes 2: Often
1 2
4
Ordinal indicators The latent distribution is restricted too mean=0 and variance = 1. Given these constraints we can estimate the the threshold parameters indicated in red and blue here. Either: μ and Σ Or: the Thresholds are identified.
5
Ordinal indicators factor model.
Since the item variances are restricted to 1 the labda’s are scaled. This means the labdas squared reflect the variance explained by the common factor F1 in the item Xi
6
Ordinal indicators factor model.
Expected (polychoric) covariance: Cov(Xi) = Σ = i it+ i Expected means/threshold: E(Xi)= μ = τi + iκ Restrictions for identification (alternatives, Millsap 2004) = 1 diag(22t+ 2)=diag() μ = 0 = τi + iκ = factor loadings = factor (co)variance matrix = residual variances ε1 to εn τ = intercepts κ = factor mean
7
Model restrictions for measurement invariance.
Configural invariance: Group 1: 1 = 11 1t+ 1 Group 2: 2 = 22 2t+ 2 Restrictions: Thresholds are equal over groups. diag(11t+ 1)=diag(22t+ 2)=diag() Indicator means: Group1: 0 Group2: FREE 2 = 1 = 1
8
Model restrictions for measurement invariance.
metric invariance: Group 1: 1 = 1 t+ 1 Group 2: 2 = 2 t+ 2 Restrictions: Thresholds are equal over groups. diag(11t+ 1)=diag(22t+ 2)=diag() τ : group1: 0 group2: FREE Κ: set to 0 1 = 1 (however 2 = FREE)
9
Model restrictions for measurement invariance.
Strong invariance: This concerns the means model. E(Xi)= μ1 = 0 E(Xi)= μ2 = κ and τ2 = 0 Thresholds are equal over groups. diag(11t+ 1)=diag(22t+ 2)=diag() τ : group1: 0 group2: 0 Κ: set to 0, free in group 2. 1 = 1 (however 2 = FREE)
10
Model restrictions for measurement invariance.
Strict invariance: Can we equate the matrix over groups. However at this point is not a free parameter: diag(11t+ 1)=diag(22t+ 2)=diag() So we must chose arbitrary values for that are equal for both groups. Replace the residuals algebra (diag(22t+ 2)=diag()) with a residuals matrix.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.