Please Help - Regression Analysis

#1
Hi,

I was hoping that someone could help me with interpreting my spss output for my honours thesis. I feel very overwhelmed as I am not very confident when it comes to quantitative methods and statistical analysis. My supervisor is away so I am unable to run this by him. I understand if you do not have the time to help me, but I thought it was worth a try!

I have written the whole thesis but when I was proof reading it, I got confused when I got to the results and discussion section, and started to think that maybe I had not interpreted my SPSS output correctly.

The purpose of my study is to investigate the influence that spirituality, religiosity and demographic factors have on the ego integrity of elderly people.

The results of my regression showed that the combined demographic factors (gender, ethnicity and education level) accounted for 29.2% of the variance in ego integrity. Adding religiosity into the equation resulted in a significant F. Change of 9.4% of the variance above and beyond the demographic factors, F(1,45) = 6.9, p < .05. Entering spirituality into the equation also resulted in a significant F. Change, F(1,44) = 6.86, p < .05, accounting for a further 8.3% of the variance in ego integrity. The model as a whole (including demographic factors, religiosity and spirituality) explained 46.9% of the variance in ego integrity.

The Anova table indicated that, entry of the demographic variables alone, yielded a significant prediction equation, F(3,46)= 6.33, p<.01. Addition of the religiosity variable, also yielded a significant prediction equation, F(4,45)= 7.09, p<.001. Entering the spirituality variable resulted in an overall significant prediction equation, F(5,44)= 7.78, p<.001. I understand everything up until here, but then I get confused when I look at the coefficients table in the SPSS output.

The coefficient table shows that among the demographic variables, education level alone was the only significant predictor of ego integrity. There was a positive correlation between education level and ego integrity (B=5.03, p<.01), indicating that higher education levels are related to higher ego integrity scores. There was also a significant positive correlation between spirituality and ego integrity (B=.21, p<.05). Thus, higher levels of spirituality are related to higher ego integrity scores. But, the positive relationship between religiosity and ego integrity, however, is not significant (B=.04, p>.05). This confuses me, because the previous tables found that religiosity is a significant predictor of ego integrity, but now the coefficients table is claiming there is not a significant relationship. I know that I must be looking at this wrong...I thought I understood it, so I wrote the discussion section of my thesis, but then when I read over it, I just got confused, so I thought I cant submit it if I dont even know what I mean!

I was wondering if somebody could explain to me what exactly the coefficients table explains? I have tried researching this on the net, but so far I havent found anything that really helps me out.

I hope what I said here makes sense. I am sorry for being such a pain. I think I am just beginning to panick because it has to be in at the publishers office on Thursday, and since my supervisor has not had the time to read the results and discussion section (I re-sent it to him a couple of weeks ago) I am beginning to worry that I have not done the right thing!

Thank you so much,
Nomesxo
 
#3
Are you saying that the change in R^2 shows a significant effect but the t value for the coefficient doesn't?

It is a hierarchical model, right?