Hopefully a Novice Questions...

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 3
Joined: Mon Feb 10, 2014 9:13 pm

Hopefully a Novice Questions...

Postby Scott_L » Fri Feb 21, 2014 6:10 am

I have some variables that have 4 ordinal choices, and they are negatively skewed over 1. With only 100 respondents, adding a constant and than trying to clean it up simply isn't working. I tried adding 1 and then square rooting it, and I tried adding 2 and 3 and then log10 them. It does nothing to change the shape of normal distribution, which is basically like a 3 bar graph with 6 in the first column, 4 in the second, and 2 in the last.

I want to run T-Tests on this based on a variable that will split the respondents damn near down the middle. Should I accept defeat and move on to other statistical tests, and if so, any suggestions? Any ideas??

Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Hopefully a Novice Questions...

Postby GerineL » Fri Feb 21, 2014 8:51 am

In this case you can use the mann-whitney u test

Who is online

Users browsing this forum: No registered users and 1 guest