0
$\begingroup$

One of my demographic variable is age. Age is measured as continues data not categorical. If I want to test differences between two groups to determine whether there is significant difference between them or not, shall I use independent t-test, or I have to covert it to categorical variable then calculate chi-square.

Thank you

$\endgroup$
  • $\begingroup$ Please edit the question. Difference between two groups of what? What has age got to do with any of the things you mention? $\endgroup$ – Sid Aug 7 '14 at 5:34
2
$\begingroup$

You should make sure you fulfill the assumptions made in a independent t-test (e.g. normality). Check Field or other literature/the internet concerning assumptions.

If the assumptions are met, you can indeed do an independent t-test.

If not, you can compute a Wilcoxon rank-sum test, or Mann-Whitney test.

Converting your variable to a categorical variable (and using chi-square test consequently) will make you lose power, so that would not be advisory.

Hope this helps

$\endgroup$
2
$\begingroup$

I'll take an example to undestand your requirement.

If you have two variables in your data, Pregnant (binary variable taking values 'Yes' and 'No') and Age (continuous). If you want to test whether there is a significant difference in the mean age between those who are Pregnant and those who are not, you might consider 't.test' in R under the assumption of unequal variance.

Your null will be Ho : the true difference in the age is zero against H1 : true difference in mean ages is not equal to 0

Then you can state you assumption based on the p-value.

As for your doubt, no you dont have to convert age to a categorical variable.

I hope I've understood your requirement correctly.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.