Sorry the table did not line up, but hopefully it makes sense
Help needed,
I am trying to escape ABD hell. I am only a couple of statistical procedures away from slaying the dissertation beast and have hit a wall. I am needing a statistical procedure that would allow me to compare the means of the number of sessions (DV) per adult attachment style (IV, 4 layers) per therapeutic attachment phase (IV, 4 layers). As some people are more visual thinkers (as I am), here is a table that I am trying to explore statistically.
Means (SD)
Therapeutic Attachment Secure Anxious Avoidant Anx/Avd
Pre-Attachment 23.00 (12.73) 16.80 (25.86) 11.25 (5.32) 6.00 (1.83)
Attachment in the Making 16.75 (15.25) 11.52 (11.36) 7.50 (.71) 8.40 (5.13)
Clear Cut Attachment 11.25 (7.68) 20.43 (20.19) 11.00 (1.41)
Goal Corrected Partnership 21.83 (13.29) 22.70 (30.23) 16.50 (13.44) 8.0 (2.28)
I would like to be able to see if the means for Securely Attached (as well as the remaining attachment styles) are significantly different through the progression of the phases. The pain of the neck of it all is that I will need to run this same statistical procedure on three other DV's (I am certain my dissertation chair will want each DV run separate). I am also concerned that my base pack of SPSS will not be able to run something like this.
I am sorry for making this post into a mini dissertation but I wanted to be as specific as possible. Any assistance is appreciated.
Sorry the table did not line up, but hopefully it makes sense
I am not really sure what you mean by comparing means. If your DV is interval you can run linear regression (or ANOVA) regardless of what your IV are generally - assuming you have not violated some assumption of course. The distribution of the IV does not matter. If you are trying to compare change over time you might find it easier to do repeated measure ANOVA.
"Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995
jhayes (02-20-2014)
The means, in this case, is the mean number of sessions a client has been in therapy. But it with four different attachment styles across four attachment phases, that brings in 16 different sets of means (actually 15, because there was not a Avoidantly attached client a particular phase). I would like to be able to compare the means through a progression of phases (if I am using an ANOVA, I could get this info from the Games-Howell Post hoc analysis). At least with the base pack of SPSS, I can only list one IV (or factor, as it is called when I am entering in SPSS). Would linear regression allow me to compare the means of any DV to a particular attachment style (one level of a IV) through a progression of phases (all four levels of the other IV).
I apologize if I am not phrasing something right. I realize I have lived with this in my head for so long, it makes sense to me, but to others, maybe not so much.
I suspect the problem is the use of the base pack rather than the method although I don't know the post hoc test you refer to. Given that I am not certain if the type of analysis you are running is similar to the linear regression I know I am not sure of the answer to your question (unless by means of the DV you simply mean the effect size on the DV of a specific level of a categorical IV). It seems almost as if you were addressing an interaction effect where you are analyzing the impact of one IV on the DV at specific levels of another IV. This could be calculated through simple effects rather than the normal regression slopes. But his is not normally done through time, if that is what you are doing.
It might be better for a more savy statistical poster to help you.
"Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995
jhayes (02-20-2014)
One of the DV's I need to examine is time. I read a bit on the forum and online stats sites to figure out I can run a factorial anova to examine the time DV. The other DV's are actually ordinal. And if I have figured some of this stuff out correctly, I would need to run an ordinal logistic regression. Trying to research that now. I have figured out how to run this in my SPSS, but still trying to understand the output results. I think you are correct in that I would like to look at the interaction results, analyzing the impact of one IV on the DV at specific levels of another IV. How would I do this using "simple effects". Any assistance would is greatly appreciated because my dissertation chair is not much of a stats guys so I have been on my own to figure this out.
Thanks
Tweet |