Is it suitable to use t-test?

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Is it suitable to use t-test?
Message posted by Kathy on July 28, 2000 at 12:00 AM (ET)

First of all, thanks for answering my question last time. I repeat my question once now.I want to test a hypothesis about two independent means. For one sameple size is 19 and the sample of the second one is 29. Is the sample size large enough to perform t-test? However, I read some books that state that t-test require a normality assumption. Therefore, i plot a histogram for these 2 groups.As a result, i don't think the distribution is normal. For this reason, should i use non-parametric test instead? Can you give any suggestion to me? Thanks for answering again.

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Is it suitable to use t-test?
Message posted by Phil on July 28, 2000 at 12:00 AM (ET)

There are several non-parametric test that would be good. See: http://home.clara.net/sisa/ordhlp.htm

The test hypotheses about the median, not the mean.

Did you statistically test to see if the data was normal (suggestion normal probability plots? You might be surprised at what would pass for normality for smaller sample sizes.

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.