0
$\begingroup$

In my regression analysis, I have 1 dependent and 5 independent variables. The analysis showed that fifth independent variable's p-value (Sig.) is 0.734, which is 0.05, thus statistically insignificant for analysis/model.

The regression method I am using is: Backward.

My question is: 1) how should I state this result in my paper? Also what should I write about hypotheses? should I reject the hypotheses or just say that it was not-significant for the model, therefore the hypotheses could neither be rejected, nor approved?

Please advise, what should I write

Improve this question
$\endgroup$
2
  • $\begingroup$ The only thing you can say is that the null hypothesis can't be rejected. $\endgroup$
    – George
    Commented Dec 26, 2015 at 19:26
  • $\begingroup$ My comment has nothing to do with the question. "A high p-valued variable" just caught my attention. I don't know whether English speakers realize how their language so easily allows some phrasings which are meaningful but whose French translations would be totally hilarious. $\endgroup$
    – Stéphane Laurent
    Commented Dec 26, 2015 at 19:31

1 Answer 1

Reset to default
0
$\begingroup$

It really depends on the context, ie. what does non-signifigance imply.

If you model, as an example, says that in fact average age of the population does not impact the level of health care expenditure - that you have some serious explanation to do - because it goes against all well established results.

On the other hand, if it is a borderline (non)-important variable that you included then, perhaps, it does not really matter.

The point I am trying to get at is this: the readers want some form of explanation/story as to why these results makes sense (or why they don't) and you need to supply that.

As for what you should write, I suggest following the advice of @George - the only thing this particular signifigance level tells you is that you cannot reject the null hypothesis.

Improve this answer
$\endgroup$
3
  • $\begingroup$ that means that I can reject the H1 hypotheses as I cant reject H0, right? $\endgroup$
    – Geo Drawkcab
    Commented Dec 26, 2015 at 22:55
  • $\begingroup$ No it means exactly the opposite $\endgroup$
    – Repmat
    Commented Dec 26, 2015 at 22:57
  • $\begingroup$ Your description of your situation is not clear. Are you doing one regression with all of the "independent" variables (I'll call them regressors), or are you doing five simple regressions, one on each regressor? If the former, then a high p-values for the fifth regressor only tells you that the fifth regressor doesn't add any detectable information to the regression beyond what the first four provide. $\endgroup$
    – Martha
    Commented Dec 28, 2015 at 5:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.