. Author manuscript; available in PMC: 2022 Dec 23.

Published in final edited form as: J Happiness Stud. 2022 Sep 5;23(8):3713–3730. doi: 10.1007/s10902-022-00565-8

Table 3.

Model Fit for Cross-Lagged Panel Analyses

Model	χ²	df	χ²/df	CFI	RMSEA (90% CI)

M1: Model with free factor loadings for Vocabulary construct	176.13	6	29.36	0.980	0.092 (0.081–0.104)
M2: Model with invariant factor loadings for Vocabulary construct	183.11	14	13.08	0.981	0.060 (0.053–0.068)
M3: Autoregressive model	2996.12	154	19.46	0.842	0.129 (0.125–0.133)
M4: PA → Vocabulary	2995.35	152	19.71	0.842	0.130 (0.126–0.134)
M5: Vocabulary → PA	2984.14	152	19.63	0.842	0.130 (0.126–0.134)
M6: Fully cross-lagged	2983.48	150	19.89	0.842	0.131 (0.127–0.135)

Note. PA = positive affect; CFI = comparative fit index; RMSEA = root mean square error or approximation; CI = confidence interval