2
$\begingroup$

I have a doubt related to some calculation a statistician performed

  I compared the increases in a Serum Hormone due to two different therapies X and Y. The results I got are in the table. The statistician who provided the data says there is a statistical difference between the two sample.

However, even if my P is < than 0.05. as far as I know the "critical t-value" is the minimum t-value I need in order to have P < 0.05. So If my t-value is smaller than my critical t-value, then my result will be not significant. 

Am I right ?

enter image description here

$\endgroup$
2
  • $\begingroup$ Your table doesn't show the sample size. This is important because it determines the degrees of freedom for the t test. $\endgroup$ – Michael R. Chernick Apr 11 '17 at 14:13
  • 1
    $\begingroup$ @Michael The second and third lines of statistics show us the sample size is between $146$ and $167$, because those values (minus one) are the degrees of freedom for which a two-tailed $t$ statistic of $-2.7536$ yields a p-value of $0.0066$. The critical value of $1.9749$ strongly suggests the df is $160$, telling us the sample size is $161$. Regardless, all the information to answer this question is available: the sample size is not needed. $\endgroup$ – whuber Apr 11 '17 at 16:55
2
$\begingroup$

It's a two-tailed test, so you need to compare the absolute value of your t-statistic to the critical value.

You then have $|-2.7536| = 2.7536 > 1.9749$, which would lead you to reject, just as you have with the p-value.

$\endgroup$
1
  • 1
    $\begingroup$ Thank you Glen. So the statistician was right :D. Thank you for your help $\endgroup$ – Simone G Apr 11 '17 at 10:44

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.