0
$\begingroup$

I was told to rewrite an old java program that performs some statistical calculations. I was able to solve most of the problems until I ran into this one:

The program compares two sets of data over 100 samples. I have standard deviations for both sets of data and something called F Calculated and F from the table in the F Test section.

The original program tells if there is a statistically significant difference between standard deviations of dataset 1 and 2.

After that, there is a T-Test section, which basically tells the same.

In my understanding so far F-test and T-Test are two different things and we should not confuse them. However, I have seen claims that F value is the T value squared. Also, I have seen questions where people are confused about the different results obtained from F-Tests and T-Tests.

Looking at the code I can see T-Test code in some class that deals with F-Test. Some details of the T-Test calculations are hidden in some compiled code that I have no access to so I need to figure out and implement the correct formula all by myself. We are not sure if the existing code is correct.

Can someone please point me to a document that shows how to calculate p-value and alpha in T-test and comment about mixing F-Test and T-Test formulas?

$\endgroup$
  • $\begingroup$ Given the information we have, there is no way we can reverse engineer this for you. Will your program compare small datasets? If it will you can tell us the results of a toy example, say 5 cases per group. I'm assuming by "100 samples", you mean 100 cases in each dataset. $\endgroup$ – Heteroskedastic Jim Oct 3 '18 at 12:25
  • $\begingroup$ I had to build a scatterplot chart for the data and display some statistical summary. Because we speak the different language I think I have to confirm that there are over 100 data points on the chart. I have found a way of finding p-value of my data using a t-test distribution table for the accuracy I want and degrees of freedom I use. I assume that to validate the data the ttest score has to be greater than the value obtained from the table statsoft.com/Textbook/Distribution-Tables $\endgroup$ – ruby_object Oct 3 '18 at 13:57

Your Answer

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

Browse other questions tagged or ask your own question.