Well that kinda failed, it only shows the picture in my reader. Here's the plain text as well.
Yea I did notice, I guess because sometimes it doesn't make sense to me to have large coefficients based on a small sample size. But it gets harder at the edges. I have an example here, just wondering if you could possibly see what you think about this? The Wald isn't impossibly high here, but I can't remember which of my models out of the tens had the very high Walds! But this is a good example I think because the size of the small dichotomous group is only 62 yet the coefficients are high.
Just thought I'd also point out I usually add the variables in steps, starting with the SES because they're of main interest to me then adding other sociodemographics. But here I've just presented the final model with everything in. I *think* I've included all the relevant information! http://i55.tinypic.com/1565kyh.jpg
|