T-test or correlation coefficient - (May/22/2011 )

Hi,

I am analysing the viral load in some samples from patients and I think that the viral load decreases when the age increases. How I can prove the significance of the relation between viral load and age. Shall I use T-test or correlation coefficient and if t- test which numbers I should use for (tail) and (type) in the test.
I am totally new to statistics so please help me.

Thanks

The type of analysis depends on the type of data you have. If you can plot it on a graph and fit a line to it appropriately, then use the correlation coefficient. Fitting the appropriate line is tricky though - you need to be able to examine the residuals and see if they are randomly distributed or not... if they are not you will need to try another curve.

A t-test (or chi squared test) can be used to compare two groups of samples. ANOVA can be used to compare multiple samples. However, there are specific conditions that make each analysis the correct one to use.

I suggest going to your supervisor and asking what would be appropriate, and if all else fails, try asking a statistician.

If you want to prove the correlation between viral load and age then test the correlation coefficient. The t-test or any other test that compares statistics between groups will not give you a correlation. It will tell you only if there is a significant difference between the mean value between two groups. My personal opinion is that the t-test is a bit over-used.