Confidence intervals in SPSS

[Piltdown13]

Hi folks,

I have a question regarding confidence intervals in SPSS. My understanding is that a 95% confidence interval (the one I'm planning to use) is bounded by values that are 1.96 standard deviations above and below the mean. However, when I have SPSS calculate the 95% confidence interval for my data, the bounds shown in the descriptives table don't make any sense. Sometimes the bounds will be less than one standard deviation from the mean; sometimes they will be several standard deviations from the mean. My data are all clean, so I doubt that's the problem.

I'm wondering what's going on here...how does SPSS calculate confidence intervals? Is there something I'm missing?

Thanks so much!!

[Piltdown13]

OK, I figured it out -- first, I somehow got it in my head that you use the standard deviation in calculating confidence intervals, but of course it is the standard error. Second, I had assumed that SPSS was using a critical value of 1.96 for a 95% confidence interval, but it is in fact using the t-score for a p of .05 and degrees of freedom=n-1...which, given my sample sizes, is quite a bit higher than 1.96!