0

Good Morning,

I have a scientific article to discuss and present in class, the article is about the positif benefit of the standard ISO 9001 on medium and large companies. By doing a multiple regression analysis they have the table below: enter image description here

For the intercept of all the variable (except the last one) they have a p-value greater than alpha, so we should reject the null hypothesis thus the model should not have an intercept,

But in scientific article they put the models as follow and add the intercept in each one even if the p-value is greater than alpha.

enter image description here

Can we accept the intercept even if it p-value is greater than alpha?

Thank you very much for you help

mohammed ilyass rahmouni
  • 1
  • 1
  • The intercepts are not significant, so from a statistical point of view they are equal to 0 and therefore should not be included in the model. However some people still include it (even if it's not significant) when some $\beta$ coefficients are significant. I don't know why. – user2974951 Nov 29 '18 at 08:54
  • Please see this Q&A https://stats.stackexchange.com/questions/7948/when-is-it-ok-to-remove-the-intercept-in-a-linear-regression-model for why you should never, well hardly ever, remove the intercept from a regression model. – mdewey Nov 29 '18 at 10:25

0 Answers0