关于independent-Sample T-test,泪流求助!!

教授就是教授

9# kokkina
看sig 2tail
sig也就是P值
单纯看T值，看不出什么，他的作用是通过sig反映出来的，一般在相同的自由度下t的绝对值越大，sig越小，sig<0.05统计显著，sig>0.05统计不显著

Equal variances not assumed:
t=-0.088        df=459.581         Sig. (2-tailed)=0.930                                                   
Mean Difference=-0.002        Std. Error Difference=0.019        Lower=-0.039  Upper=0.036

For the above one since "0" in CI (your lower bound or upper bound)  the two samples you try to compare were not significantly different. Choosing two tail P-value is conservative.
Sig. (2-tailed)=0.930 is two tail p-value and another Sig. is one tail p-value. and for me you still need check CV for sampling variability. (your CV in this case around is 10%, which is less than 16.7%, so it is fine.)

13# tinalica

感谢!!正在抱着书从头看起,稍微明白点了~~汗

弱弱的问下后面诸楼，因变量答案是二分类变量，可以做T检验？
应该是完全此路不通的吧

最后决定用independent t来做了.



新的问题又出来了: 怎样在满足小于35岁(<35=1)和来自大城市(city=1)这两个条件下,对问题一的所有答案为yes(yes=1)的求和?

谁能帮忙指点下?