Skip to main content

Table 4 Selecting the final frailty model

From: Revisiting the hypothesis of syndromic frailty: a cross-sectional study of the structural validity of the frailty phenotype

    # of   Differences   Bootstrap Observed
  # of Classes Test parameters BIC in BIC LL p = 0,05 −2*(LL [H1]-LL [H0])
A. Within LCA Models
1. H [a.1]:LCA-uv1 4 classes H [a.1]|H [a.0] 42 65,264 115 −32,476,9 43,6 269,6*
2. H [a.0]:LCA-ev2 3 classes   21 65,379   −32,611,7   
B. Within Strong Measurement Invariance (SMI) models
3. H [b.1]:SoMI-uv3 4 classes H [b.1]|H [b.0] 36 65,232 120 −32,482,7 36,3 260,8*
4. H [b.0]:SiMI-ev4 2 classes   17 65,352   −32,613,1   
C. Within Weak measurement invariance (WMI) models
5. H [c.1]:WMI-uv5 3 classes H [c.1]|H [c.0] 38 65,189 61 −32,453,8 19,6 149,4*
6. H [b.0]:WMI-ev6 3 classes   26 65,250   −32,528,5   
D. Within Null Measurement Invariance (NMI) models
7. H [d.1]:NMI-uv7 3 classes H [d.1]|H [c.0] 45 65,211 82 −32,438,8 38,4 238,0*
8. H [d.0]:NMI-ev8 2 classes   25 65,293   −32,557,8   
E. Testing the LPA-uv, WMI-uv and NMI-uv models against the null model (SMI-uv)
9. H [d.1]:NMI-uv7 3 classes H [d.1]|H [b.1] 45 65,211 21 −32,438,8 282,3 87,8
10. H [c.1]:WMI-uv5 3 classes H [c.1]|H [b.1] 38 65,189 43 −32,453,8 267,5 57,8
11. H [a.1]:LPA-uv1 4 classes H [a.1]|H [b.1] 42 65,264 −32 −32,476,9 306,1 11,6
The null model:
12. H [b.1]:SMI-uv4 4 classes   36 65,232 N.A.** −32,482,7 N.A. N.A.
Notes:         
* Null rejected         
** N.A. Not Applicable        
1. LPA-uv: LPA with unequal variances       
2. LPA-ev: LPA with equal variances       
3. SoMI-uv: FMM with Strong Measurement invariance - unequal variances    
4. SiMI-ev: FMM with Strict Measurement invariance - equal variances     
5. WMI-uv: FMM with Weak Measurement invariance - unequal variances     
6. WMI-ev: FMM with Weak Measurement invariance - equal variances     
7. NMI-uv: FMM with No Measurement invariance - unequal variances     
8. NMI-ev: FMM with No Measurement invariance - equal variances