Assessment of Path Model Fit: Evidence of Effectiveness and Recommendations for use of the RMSEA-P
回顾了潜变量模型路径拟合指标的发展，通过新模拟研究解决RMSEA-P是否有效的问题，发现其对多指标模型识别错误设定非常有效，并给出了使用建议。
We review the development of path model fit measures for latent variable models and highlight how they are different from global fit measures. Next, we consider findings from two published simulation articles that reach different conclusions about the effectiveness of one path model fit measure (RMSEA-P). We then report the results of a new simulation study aimed at resolving the questions of whether and how the RMSEA-P should be used by organizational researchers. These results show that the RMSEA-P and its confidence interval is very effective with multiple indicator models at identifying misspecifications across large and small sample sizes and is effective at identifying true models at moderate to large sample sizes. We conclude with recommendations for how the RMSEA-P can be incorporated along with other information into model evaluation.