2
$\begingroup$

For a research project I have obtained results which I have to test for significance. The two values I am comparing are the true diameter of a certain object versus the calculated diameter of that object. In this case it would be favourable if the t-test showed no significant difference between those results.

However while checking the T-table I came across this problem which just won't reach my brain. When my t-value is higher than a critical value there is a significant difference, but as the probability (alpha) increases that critical value increases. This leads to my misunderstanding that for this reason I can say that:

  1. There is a 95% probability that there is a significant difference between the true values and calculated results

But since the critical value rises with probability:

  1. There is a 99.9% probability that there is no significant difference between the two

This is obviously not right, but I cannot grasp what is going wrong.

$\endgroup$
2
$\begingroup$

When my t-value is higher than a critical value there is a significant difference, but as the probability (alpha) increases that critical value increases.

This is not the case. Conventionally, "alpha" refers to the significance level; when it gets smaller, the t critical value increases. You might be confusing alpha with its complement (1-alpha).

There is a 95% probability that there is a significant difference between the true values and calculated results

This is not a correct description.

You might say something like (if the null were true) "the probability of observing a t-value at least as large (in absolute size) as the one you obtained is less than 5%".

vs

"the probability of observing a t-value at least as large as the one you obtained is larger than 1%"

There's no inconsistency there.

$\endgroup$

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.