# Which test to use to compare means for normal and non normal data

#### Vthompson

##### New Member
Hi!
I have quantitative test scores I wish to compare. However, one group (n=27) is not normally distributed and the other one (n=21) is normal.
Thank you

#### Karabiner

##### TS Contributor
You can use the t-test (Welch test). With n(total)=48, we can expect that the
mean difference is normally distributed in the population (see central limit theorem).

With kind regards

Karabiner

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Are scores bound within 0-100%? It would be beneficial if you posted a histogram of the scores for each group.

Thx.

#### Vthompson

##### New Member
I am not sure what you mean by your first question. However, see the histogram and other descriptive statistics. Thanks so much for your help!
I am looking up the Welch test. I don't recall learning about that!

#### Attachments

• 225.9 KB Views: 6

#### hlsmith

##### Less is more. Stay pure. Stay poor.
Well, what I was alluding to was that scores can be percentages, which are bound between 0 and 1.00. If all of the scores were high (e.g. .90's but there was quite a bit of variability, your standard deviation may not represent the data coverage symmetrically. Say your mean = 95% and std = 10%.