freetest
我用spss做的问卷分析共796例,28个项目,用因子分析提取了9个因子,KMO值0.692,累计贡献率
58.436%。
我想请问:
1. 在书上看到作因子分析 KMO和累计贡献率最好在0.8以上,但是有些文献根本达不到这个标准
我该如何提高 KMO和累计贡献率.
2. 我用现有的因子分析计算出9个因子的因子的分,以此为基础进行了Logistic分析,这合适吗?
anita_jiu
1. According Kaiser (1974), >0.5 - barely acceptable, 0.5-0.7 is moderate. Also, you could argue that your sample size which is relatively big, your 28 variables could be remained for factor analysis. Howevre, I do recommend you examine your correlation matrix (i.e. try to find varibles that have low R²), communality table (e.g. search for low values) for a further consideration regarding whether or not to delete some variables that are not doing a 'good' job.
2. The decision of using Logistic analysis depends on the nature of your dependent variables (e.g. are they continous, discrete, dichotomous or a combined measures?).
freetest
I do recommend you examine your correlation matrix (i.e. try to find varibles that have low R²)
correlation matrix( find varibles that have low R²)请问在哪里看这个
anita_jiu
Analyze - data reduction - factor analysis - descriptives - correlation matrix, tick boxes 'coefficients, significance levels'.