Table 4.
True standardized mean differences (PI DIFF ) and impact of non-invariant items on latent (ES SI-PI ; lower triangle) and composite (ES PI-ALL , ES PI-RES , upper triangle) mean differences
| Disease group | Ortho | Rheu | Asthma | COPD | Cancer | ||
|---|---|---|---|---|---|---|---|
|
Positive and active engagement in life | |||||||
|
Ortho |
|
|
0.32a |
0.27a |
0.31a |
0.24a |
ES
PI-ALL
|
| |
0.09 |
0.17 |
0.06 |
0.03 |
ES
PI-RED
|
||
|
Rheu |
PI
Diff
|
0.59 |
|
0.08 |
0.02 |
0.10 |
ES
PI-ALL
|
|
ES
SI-PI
|
0.27a |
|
0.14 |
0.05 |
b |
ES
PI-RED
|
|
|
Asthma |
PI
Diff
|
0.20 |
-0.43 |
|
0.06 |
0.02 |
ES
PI-ALL
|
|
ES
SI-PI
|
0.28a |
-0.02 |
|
0.09 |
0.04 |
ES
PI-RED
|
|
|
COPD |
PI
Diff
|
0.47 |
-0.13 |
0.27 |
|
0.08 |
ES
PI-ALL
|
|
ES
SI-PI
|
0.27a |
-0.13 |
0.01 |
|
0.12 |
ES
PI-RED
|
|
|
Cancer |
PI
Diff
|
-0.03 |
-0.64 |
-0.24 |
-0.52 |
|
|
| |
ES
SI-PI
|
0.28a |
> - 0.01 |
0.01 |
> - 0.01 |
|
|
|
Self-monitoring and insight | |||||||
|
Ortho |
|
|
0.11 |
0.22a |
0.12 |
0.10 |
ES
PI-ALL
|
| |
0.02 |
0.06 |
0.02 |
0.04 |
ES
PI-RED
|
||
|
Rheu |
PI
Diff
|
0.24 |
|
0.32a |
0.22a |
0.21a |
ES
PI-ALL
|
|
ES
SI-PI
|
0.08 |
|
0.13 |
0.08 |
0.10 |
ES
PI-RED
|
|
|
Asthma |
PI
Diff
|
-0.34 |
-0.56 |
|
0.09 |
0.13 |
ES
PI-ALL
|
|
ES
SI-PI
|
-0.05 |
-0.13 |
|
0.03 |
0.02 |
ES
PI-RED
|
|
|
COPD |
PI
Diff
|
-0.16 |
-0.38 |
0.16 |
|
0.03 |
ES
PI-ALL
|
|
ES
SI-PI
|
-0.03 |
-0.10 |
0.03 |
|
0.02 |
ES
PI-RED
|
|
|
Cancer |
PI
Diff
|
-0.31 |
-0.55 |
0.04 |
-0.15 |
|
|
| ES SI-PI | <0.01 | -0.08 | 0.06 | 0.02 | |||
Notes: Ortho: orthopedic condition; Rheu: rheumatism; PIDiff: Estimations of latent mean differences in partial invariance models; ESSI-PI: Difference in latent mean differences between strict and partial invariance models; ESPI-ALL: Difference between latent mean differences in partial invariance models and composite mean differences using all items of a scale; ESPI-RED: Difference between latent mean differences in partial invariance models and composite mean differences using only items with pairwise non-invariant parameters; arelevant misestimation (ES > |0.2|); bno item with DIF between groups.