Reply With QuoteAre the variables RCMAS, FES, and Harter supposed to be composites of those columns in parentheses?
If so, how should they be combined?
Until I know how those three variables are constructed, it is really difficult to answer your questions...
Hello.
I need help with determining whether there is sig diff. between enco 1 and enco 2 on three measures data is collected on RCMAS (coded as pa, wo, ca,lie and tot), FES (coh,exp,con,ind,ao,ico,aro,mre,org,ctl) and Harter(cog,soc,ath,physapp,job,rom,beh,frnd)
Is the power for this sample sufficient?
Pl. see attachment.
I have run this in SPSS using one-way Anova.
Anyone have any other ideas or can verify the results?
thanks for yr time.
Are the variables RCMAS, FES, and Harter supposed to be composites of those columns in parentheses?
If so, how should they be combined?
Until I know how those three variables are constructed, it is really difficult to answer your questions...
Hello.
The RCMAS is the first measure that consists of 5 subscales that include PA, WO, CA,LIE and TOT.
FES is the second measure and consists of 10 subscales including: coh, exp, con, ind, ao,ico,aro,mre,org,ctl
HARTER is the third measure in my project and includes the cog,soc,ath,physapp,job,rom,beh,and frnd subscales.
I am looking at three measures and one sample that is divided by primary and secondary encopretic children.
The three measures should not be combined but looked at separately.
I want to know if there are sig. diff. b/w primary encopretic group(coded as 1 on enc) and secondar (coded as 2) and which group has higher or lower scores on these three measures.
If specific subscales are significant and not others, I would like to qualitatively explain that in my writings later.
Age of children in months and their sex (coded as 1-male or 2-female may also be in the data columns.
Hope that clarifies things
Thanks for yr help.
OK. If you do a set of 1-way ANOVAs on each subscale, you should be able to tell which ones are significantly different between enco1 and enco2 - the pattern of significant/insignificant differences then should be related back to your hypotheses and the underlying theory...
Although your sample sizes are small, I don't think power is an issue here...
what is an acceptable statistical power number/percent?
Typically you want to be able to detect a difference 80-90% of the time if in fact a difference exists
Posting Permissions
|
|