Hi Karen,
I'm almost always on the receiving end of advice here, but this
sounds like a straightforward homogeneity of slopes/ANCOVA problem.
If dependent is English Scores, independent is Verbal Ability, and
you have 2 groups (Females; Males), then you would test for
homogeneity of slopes between the sexes (i.e.; do the 2 sexes have
the same slope of VA with ES). If slopes are nsd, then test for
differences in elevations of group English Scores evaluated at mean
Verbal Ability.
Does this describe what you are trying to test?
regards,
Ian
Ian D. Martin, Ph.D.
Tsuji Laboratory
University of Waterloo
Dept. of Environment & Resource Studies
On 14 Apr, 2010, at 12:44 AM, Karen Wood wrote:
> Hi all. I'm confused.
>
> This time it concerns the interpretation of the beta coefficient and
> zero-order correlation for the interaction terms.
>
>
> Example: Dependent variable = English scores
>
> males = 0; females = 1
>
> (interaction term) verbal ability X sex: beta = 0.375; zero-order
> = 0.436
>
> If the data is re coded (in order to obtain significance levels and
> beta coefficients for females)
> males = 1; females = 0
>
> verbal ability X sex: beta = -0.319; zero-order = 0.043
>
> The zero-order correlations between English scores and verbal ability
> are r = 0.126 and r = 0.587, for males and females, respectively.
>
> Thank you
>
> Karen
>
> =====================
> 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
=====================
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
Locked
