Hi together,
I am looking for an answer to an SPSS related question, using only ONE dummy coded IV with two groups. Other than that my model only includes moderators and control variables.
After performing several tests, including an explorative factor analysis and receiving reliable constructs, my model does unfortunately show an insignificant effect of the dummy coded IV.
Am I correct with the assumption that this implies that there is NO difference between the two groups? And can the intercept (=reference category) be interpreted in a meaningful way if the dummy coded IV is not significant ?
Hello Franziska,
it was not quite clear what role mentioning the factor analysis and moderator has in your question. Is this right: You have an effect of the IV that is insignificant. In this case, the intercept is the predicted Y-value for the X-group coded as 0 an the B-coefficient is the difference between X=1 and X=0. The fact that it fails to reach significance simply means that it is no strong enough to rule out the H0 as a probable scenario.
HTH
Holger
Hello Holger,
thanks for your answer. I understand, that my intercept represents my Y-value for the reference group coded as 0. I was just not sure if it makes sense to even interpret the intercept since the hypothesis which is based on this dummy variable has to be rejected as p>5%.
Franziska,
the intercept is a parameter like any other. If you like, you could use the confidence interval to judge the uncertainty with regard to its population value if you want.
Best,
Holger
Thanks very much Holger! One (last) questionl. If in general a dummy coded variable is not significant, could it be the case that the moderators, which include the dummy coded IV, are significant? I have 3 moderators for my insignificant IV and none of them is significant.
Hi Franziska,
you mean that there is no main effect/difference of the dummy IV but that the relationship with the outcome is moderated. Yes of course. Test that by introducting product terms consisting of the multiplied (uncentered/raw) dummy IV and the respective moderator.
Best,
Holger
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability