How to interpret inconsistent Beta values in different steps of hierarchical regression analysis?

[Cyrus]

I did hierarchical regression analysis on my data due to having interaction effects in my research model.



R2 increased from .695 in model1 (main effect only) to .734 in model2 (main &interaction effects)(sig. F change = .000). All the assumptions for the regression analysis have been met. I have two problems with the "coefficients" table:



1. As u can see in the table, the insignificant beta value of ZSC in model1 became significant in model2! Is it ok? I'm confused! Which value should i consider to reject/accept the related hypothesis? B of model1 (which rejects the hypothesis) or 2 (which confirms it!!)?

2. Although the Beta value for ZSC_X_CS is significant, its positive sign is against the hypothesis! it's supposed to have a negative sign according to the literature & also logic! How should i treat this hypothesis? Accept? Reject? Partially accept?!!!

[Bruce Weaver]

I have two comments.

1. This question is not about anything specific to SPSS.  Rather, it is a question about how to interpret coefficients in a regression model that includes product terms.  Therefore, it would be more suitable (IMO) for a general statistics forum (e.g., one of the sci.stat.* newsgroups) than for this list.  

2. You need to read up on interpretation of coefficients in models that contain product terms.  The book by Aiken & West is a good one.  There is another little green Sage book by Jaccard, IIRC.  What people sometimes call the "main effects" in your second model (with the product terms included) are not really main effects.  They are partial effects.  E.g., the coefficient for ZSC in your second model gives the effect of a one unit increase in ZSC *when CS (presumably ZCS) is equal to 0*.  The significance of the interaction tells you that the effect of a one unit increase in ZSC depends on the value of CS:  for every one unit increase in CS, the effect of a one unit increase in ZSC will increase by .151.  For some variables, 0 is an out-of-range or nonsensical value to be looking at.  When that is so, it is common to center the variable on some in-range and more meaningful value.  Many people choose to center on the mean.  I often prefer to center on a round figure near the minimum.  In any case, for the centered variable, a value of 0 now corresponds to whatever centering value was chosen, and the coefficients become more interpretable.  

I suspect this is now getting into more detail than you're ready for, so I'll stop there.

HTH.

Cyrus wrote

I did hierarchical regression analysis on my data due to having interaction effects in my research model.



R2 increased from .695 in model1 (main effect only) to .734 in model2 (main &interaction effects)(sig. F change = .000). All the assumptions for the regression analysis have been met. I have two problems with the "coefficients" table:



1. As u can see in the table, the insignificant beta value of ZSC in model1 became significant in model2! Is it ok? I'm confused! Which value should i consider to reject/accept the related hypothesis? B of model1 (which rejects the hypothesis) or 2 (which confirms it!!)?

2. Although the Beta value for ZSC_X_CS is significant, its positive sign is against the hypothesis! it's supposed to have a negative sign according to the literature & also logic! How should i treat this hypothesis? Accept? Reject? Partially accept?!!!

[Valerie Villella]

Thank you for your email.

I am out of the office on July 20 and will not be accessing my emails. I will respond to your message on my return to the office on July 23.

Have a wonderful day.

Valerie Villella
Education Coordinator &
Policy and Program Analyst

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD