Hi Cookies
Yes, you do need to resubmit with the correct data - it would damage your reputation, and your supervisor's, if you publish an article with incorrect stats which are then picked up by a reader. You don't want that! You say that your sup is also a co-author - if I was you I would draft an email for your sup to send to the editor with the revised article, explaining that on closer examination some of the results of the data need to be amended slightly (don't mention you made a mistake), but that they don't change the discussion. The editor may not send it out again to the reviewers - this doesn't sound like such a big deal, and if everything else is good, they might just accept your revised stats. It's not so bad!