0

I'm doing some data analysis on a small dataset (35 individuals) and I'm wondering whether my findings are statistically significant. Individuals are asked questions before and after a treatment, so it's a paired dataset.

Therefore, I've put the data in Libreoffice Calc and applied the TTEST function. Unfortunately, the information online about this function is really sparse. The docs say "Returns the probability associated with a Student's t-Test.", but I can't find any examples how to interpret this value.

Am I right in assuming that the returned value is the probabilty that the null hypothesis is true (p-value). I.e. a TTEST value 0.0011 would mean that there is a 0.11 % chance that the treatment wasn't actually helping, but the results are just random?

I'm looking for an answer that is specific for Libreoffice Calc's TTEST function.

Thanks for your help!

Screenshot 1

Dataset 1 Screenshot 2 Dataset 2

Fabian
  • 143
  • 4
  • http://stats.stackexchange.com/questions/31/what-is-the-meaning-of-p-values-and-t-values-in-statistical-tests is related, but it does not answer my question what Libreoffice Calc's TTEST actually returns. The result is /not/ a t-value (that I could look up in a t-table with DF=34), it is a probability. – Fabian Jul 07 '15 at 18:02
  • It seems to me that LibreOffice Calc's TTEST returns the p-value. I used data from an [example data set](http://www.statstutor.ac.uk/resources/uploaded/paired-t-test.pdf) and TTEST's value matched the p-value mentioned there (p = 0.004). [Screenshot](http://imgur.com/CqL5O01) – Fabian Jul 08 '15 at 22:18

0 Answers0