Does not include dependents, groups into families. May include retirees though I could not verify this either. (you're welcome to dig through the data linked in the article yourself, let me know if you find anything). I think the general point would still stand even excluding retirees, but agree that would be a more telling cut of data.