Table III. Structural invariance...

Table III.

Structural invariance analysis of constructs

	Weight
Constraint	C1	C2	χ² (df)	Δ χ² (Δdf)
Panel a. Structural invariance analysis of constructs across engagement objects (local government)
1. Fully unconstrained model			675.576 (250)
2. Factor loadings			687.582 (259)	12.006 (9)
Factor loadings and equal coefficients for:
3. Involvement →positive customer engagement	0.809*	0.765*	687.583 (260)	0.001 (1)
4. Involvement → negative customer engagement	−0.400*	−0.385*	688.214 (260)	0.632 (1)
5. Positive customer engagement → WOM	0.732*	0.791*	687.891 (260)	0.309 (1)
6. Negative customer engagement → WOM	0.529*	0.250*	690.837 (260)	3.255 (1)
Model fit:	N = 325	N = 325	N = 325
χ² (df)	355.721 (125)	319.855 (125)	675.576 (250)
CFI	0.95	0.96	0.95
IFI	0.95	0.96	0.95
GFI	0.9	0.89	0.89
RMSEA	0.07	0.06	0.05
Panel b. Structural invariance analysis of constructs across engagement objects (social networking sites)
1. Fully unconstrained model			587.249 (250)
2. Factor loadings			597.259 (259)	10.684 (1)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.684*	0.681*	597.936 (260)	0.677 (1)
4. Involvement → negative customer engagement	0.08	−0.008	598.833 (260)	0.897 (1)
5. Positive customer engagement → WOM	0.781*	0.745*	597.957 (260)	0.698 (1)
6. Negative customer engagement → WOM	0.212*	0.348*	600.205 (260)	2.964 (1)
Model fit:	N = 300	N = 300	N = 300
χ² (df)	316.041 (125)	271.208 (125)	587.249 (250)
CFI	0.96	0.97	0.96
IFI	0.96	0.97	0.96
GFI	0.89	0.9	0.9
RMSEA	0.07	0.06	0.04
Panel c. Structural invariance analysis of constructs across contexts (local government brand object vs social networking site brand object)
1. Fully unconstrained model			671.760 (250)
2. Factor loadings			692.427 (259)	20.667 (9)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.809*	0.684*	692.767 (260)	0.34 (1)
4. Involvement → negative customer engagement	−0.400*	0.08	716.464 (260)	24.027 (1)**
5. Positive customer engagement → WOM	0.732*	0.781*	692.865 (260)	0.618 (1)
6. Negative customer engagement → WOM	0.529*	0.212*	706.400 (260)	13.97 (1)**
Model fit:	N = 325	N = 300	N = 625
χ² (df)	355.721 (125)	316.041 (125)	671.760 (250)
CFI	0.95	0.96	0.95
IFI	0.95	0.96	0.95
GFI	0.9	0.89	0.89
RMSEA	0.07	0.07	0.05
Panel d. Structural invariance analysis of constructs across contexts (Local government community vs social networking site community)
1. Fully unconstrained model			591.060 (250)
2. Factor loadings			607.361 (259)	16.30 (9)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.765*	0.681*	607.700 (260)	0.339 (1)
4. Involvement → negative customer engagement	−0.385*	−0.008	615.506 (260)	8.45 (1)*
5. Positive customer engagement → WOM	0.791*	0.745*	609.174 (260)	1.813 (1)
6. Negative customer engagement → WOM	0.250*	0.348*	607.379 (260)	0.252 (1)
Model fit:	N = 325	N = 300	N = 625
χ² (df)	319.855 (125)	271.208 (125)	591.060 (250)
CFI	0.96	0.97	0.97
IFI	0.96	0.97	0.97
GFI	0.89	0.9	0.9
RMSEA	0.06	0.06	0.04

	Weight
Constraint	C1	C2	χ² (df)	Δ χ² (Δdf)
Panel a. Structural invariance analysis of constructs across engagement objects (local government)
1. Fully unconstrained model			675.576 (250)
2. Factor loadings			687.582 (259)	12.006 (9)
Factor loadings and equal coefficients for:
3. Involvement →positive customer engagement	0.809*	0.765*	687.583 (260)	0.001 (1)
4. Involvement → negative customer engagement	−0.400*	−0.385*	688.214 (260)	0.632 (1)
5. Positive customer engagement → WOM	0.732*	0.791*	687.891 (260)	0.309 (1)
6. Negative customer engagement → WOM	0.529*	0.250*	690.837 (260)	3.255 (1)
Model fit:	N = 325	N = 325	N = 325
χ² (df)	355.721 (125)	319.855 (125)	675.576 (250)
CFI	0.95	0.96	0.95
IFI	0.95	0.96	0.95
GFI	0.9	0.89	0.89
RMSEA	0.07	0.06	0.05
Panel b. Structural invariance analysis of constructs across engagement objects (social networking sites)
1. Fully unconstrained model			587.249 (250)
2. Factor loadings			597.259 (259)	10.684 (1)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.684*	0.681*	597.936 (260)	0.677 (1)
4. Involvement → negative customer engagement	0.08	−0.008	598.833 (260)	0.897 (1)
5. Positive customer engagement → WOM	0.781*	0.745*	597.957 (260)	0.698 (1)
6. Negative customer engagement → WOM	0.212*	0.348*	600.205 (260)	2.964 (1)
Model fit:	N = 300	N = 300	N = 300
χ² (df)	316.041 (125)	271.208 (125)	587.249 (250)
CFI	0.96	0.97	0.96
IFI	0.96	0.97	0.96
GFI	0.89	0.9	0.9
RMSEA	0.07	0.06	0.04
Panel c. Structural invariance analysis of constructs across contexts (local government brand object vs social networking site brand object)
1. Fully unconstrained model			671.760 (250)
2. Factor loadings			692.427 (259)	20.667 (9)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.809*	0.684*	692.767 (260)	0.34 (1)
4. Involvement → negative customer engagement	−0.400*	0.08	716.464 (260)	24.027 (1)**
5. Positive customer engagement → WOM	0.732*	0.781*	692.865 (260)	0.618 (1)
6. Negative customer engagement → WOM	0.529*	0.212*	706.400 (260)	13.97 (1)**
Model fit:	N = 325	N = 300	N = 625
χ² (df)	355.721 (125)	316.041 (125)	671.760 (250)
CFI	0.95	0.96	0.95
IFI	0.95	0.96	0.95
GFI	0.9	0.89	0.89
RMSEA	0.07	0.07	0.05
Panel d. Structural invariance analysis of constructs across contexts (Local government community vs social networking site community)
1. Fully unconstrained model			591.060 (250)
2. Factor loadings			607.361 (259)	16.30 (9)
Factor loadings and equal coefficients for:
3. Involvement → positive customer engagement	0.765*	0.681*	607.700 (260)	0.339 (1)
4. Involvement → negative customer engagement	−0.385*	−0.008	615.506 (260)	8.45 (1)*
5. Positive customer engagement → WOM	0.791*	0.745*	609.174 (260)	1.813 (1)
6. Negative customer engagement → WOM	0.250*	0.348*	607.379 (260)	0.252 (1)
Model fit:	N = 325	N = 300	N = 625
χ² (df)	319.855 (125)	271.208 (125)	591.060 (250)
CFI	0.96	0.97	0.97
IFI	0.96	0.97	0.97
GFI	0.89	0.9	0.9
RMSEA	0.06	0.06	0.04

Notes:

panel a: c₁ council brand object; c₂ community object; *p < 0.05 N = 325 across all cohorts as survey items were repeated for each object per respondent

Panel b: c₁ brand object; c₂ community object; *p <* 0.05 N = 300 across all cohorts as survey items were repeated for each object per respondent.

Panel c: c₁ brand object Local Government context; c₂ brand object SNS; *p < 0.05 **p <* 0.00;

Panel d: c₁ community object Local Government context; c₂ community object SNS; *p < 0.05

[ViewLarge]

Sharing Unavailable