协方差分析

用年级为协变量，年龄为自变量，抑郁症状为应变量分男女两组得到男和女各年龄段的抑郁症状的边际估计平均值（Estimated Marginal Means） 及标准差。  和用年级为协变量，年龄、性别为自变量，抑郁症状为应变量，使用性别条件下的年龄比较得到男女各年龄段的抑郁症状的边际估计平均值（Estimated Marginal Means） 及标准差。哪种方法正确，或者哪种方法更合理？两种方法的结果是不一样的，第一种方法是分性别分别得到年级的控制值；而第二种方法男女的年级控制值是一样的。

数据呢，要经过检验和你分析的目的啊！

两种方法的选择是依数据选择吗？协方差分析都是用SPSS做的，本人倾向第一种方法

把两种方法都用spss做一下分析，然后看看结果和你的分析目的哪个更联系紧密，更能达到分析效果就可以了。

不会吧，总要有点统计理论的支持吧

Which model to use depends on your study aim and your understanding of your study topic (clinical experience, finding of pervious studies, etc.). If you really think gender play a important role on depression, you can explore your data as followings:
1)Do ANCOVA on male and female separately
2)Do ANCOVA with gender as a variate and considers the two-way interactions with other variates and covariates
3)Do multiple regression with gender as a predictor and considers the the two-way interactions with other varitates

Although the above three ways have different technical focuses, finally they should tell the same story

thanks for hlpcdadong! in fact, I have checked some articles about ANCOVA, and found that the way I used grade as covariate maybe unfit, so I have gave up the ANCOVA in my study.

请版主送我的20论坛币给hplcdadong，以示感谢！

Thanks and good luck with your analysis