chi-square & t-test

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

chi-square & t-test
Message posted by Anne on September 29, 2000 at 12:00 AM (ET)

When do you use a chi-square test and when do you use a t-test?

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

Re: chi-square & t-test
Message posted by phil on September 29, 2000 at 12:00 AM (ET)

t-tests are usually used to test hypotheses comparing the means of two samples and assumes the underlying populations are normally distributed and the variables are continuous.

The Chi-square "Goodness of Fit" test is used to test whether or not a "distribution" or any of its parameters follow some distribution. The Chi-square tests for homogeniety and independence are for discrete, (countable) data and tests for significance of relationships between factors.

Re: chi-square & t-test
Message posted by jg on September 30, 2000 at 12:00 AM (ET)

The t-test means that you have a test statistic that is normally distributed and the standard deviation is estimated.
The chi-squared distribution has many different uses. One possibility is that you have a binomial or hypergeometric situation and that you are using the normal as an approximation. In the chi-squred test for independence you are baically approximating a multinomial distribution with a normal approximation, etc.

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