Fit indices for general sleep disturbance scale gmm solutions over seven assessments, with dyad as a clustering variable.

*p** ***p.05.aRandom coefficients latent growth curve model with linear and quadratic components; Chi2 = 108.81, 26 df, pb2-class model was selected, based on its having the smallest BIC, the largest entropy, and a significant VLMR. Further, the VLMR is not significant for the 3-class model, and the 3-class model estimated a class with only 4% of the sample – a class size that is unlikely to be reliable.cThis value is the Chi2 statistic for the VLMR. When significant, the VLMR test provides evidence that the K-class model fits the data better than the K-1-class model.Abbreviations: GMM = Growth mixture model; LL = log likelihood; AIC = Akaike Information Criteria; BIC = Bayesian Information Criterion; VLMR = Vuong-Lo-Mendell-Rubin likelihood ratio test; CFI = comparative fit index; RMSEA = root mean square error of approximation.

*p** ***p.05.a 包含线性与二次成分的随机系数潜增长曲线模型（Random coefficients latent growth curve model）；卡方（Chi2）=108.81，自由度（df）=26，两类别模型（2-class model）被选中，原因在于其具备最小的贝叶斯信息准则（BIC, Bayesian Information Criterion）、最大的熵值，且Vuong-Lo-Mendell-Rubin似然比检验（VLMR, Vuong-Lo-Mendell-Rubin likelihood ratio test）结果显著。进一步而言，三类别模型的VLMR检验结果并不显著，且三类别模型估计出的类别仅包含4%的样本——该类别规模难以保证可靠性。该数值为VLMR检验对应的卡方统计量。当VLMR检验结果显著时，即可证明K类别模型相较于K-1类别模型对数据的拟合效果更优。缩写说明：GMM = 增长混合模型（Growth mixture model）；LL = 对数似然值（log likelihood）；AIC = 赤池信息准则（Akaike Information Criteria）；BIC = 贝叶斯信息准则（Bayesian Information Criterion）；VLMR = Vuong-Lo-Mendell-Rubin似然比检验（Vuong-Lo-Mendell-Rubin likelihood ratio test）；CFI = 比较拟合指数（comparative fit index）；RMSEA = 近似误差均方根（root mean square error of approximation）