Multiple Regression, Analysing Attachment Dimensions

#1
Hi,

I am trying to do my analysis for my work but getting very confused.

I am using multiple regression (SPSS version 17) and have my dependent variable, which is my own measure of 'rebound'. My two predictor variables are attachment (ECR: Brennan et al, 1998) and basic psychological need satisfaction (Need Satisfaction Scale; La Guardia et al, 2000).

I have used the Experiences in Close Relationships questionnaire (ECR; Brennan et al, 1998) and there are two scores/dimensions for each participant; anxiety and avoidant. I was told to work out the average for each, but I have no idea how to analyse them in SPSS or how to enter them into a multiple regression model! Do I enter them as individual 'predictor variables' or is there another way? Or do I do two completely different analyses, one for each dimension?

I am looking at the extent to which attachment styles (anxious, avoidant) and the three basic needs; autonomy, relatedness and competence (BPNS) can predict an individual entering a 'rebound relationship'. I have asked individuals about how their needs were met in their previous relationship (with the Need Satisfaction Scale) and given them the attachment questionnaire.

I have entered my data into SPSS so I have average scores for autonomy, relatedness and competence for each individual and I have two scores from the attachment questionnaire. One is the average score for anxiety and the other is the average score for avoidance (the average scores are for each individual). I then have my 'rebound' score for each participant. Rebound score will be placed in the 'dependent variable' box, but my problem is how to analyse the others.

Do I just place everything else in the 'independent variable' box or do I have to do separate analyses for each of the attachment dimensions? Or is there something completely different that I need to do?

Thanks,

Ami
 

Lazar

Phineas Packard
#2
Take a close look at all the relationships between your variables first. Take a look at correlations but also graphical displays including scatter plots. This will allow you to take a look at what is going on and whether you are going to have any problems.

Things to look out for:
1. High correlations between IVs. Pay particular attention to high negative correlations. These can give you a snap shot of whether there is likely to be any problems later on (e.g. multi-collinearity).
2. Look for any non-linear relationships in the scatter plot.

If all is in order enter all the IVs in at once.