Size of Effect

[doctor]

I just saw something here that made me wonder about a method i used in the SPSS regarding measuring the effect each IV had on my DV in a linear model i created.
I understood that the partial eta squared is the right measure that can tell me what the effect of each IV on the total variance - In other words it is the euivalent of squared R for the model just for individual effect (of the main effects). Because i wanted this number to sum up to 100% i calculated myself the eta squared (the SPSS doesn't give it just the partial one...) -
SS for specific IV devided per SStotal(corrected). Then i could see what the size of effect each IV had in my model.
Isn't it correct ?!?!
I saw here somebody talking about the b value.....
It's Realy Important for me to know if i did a huge mistake here !!!

[Masteras]

actually, he has a point, what you need to check is the significance of the b values, see the standardised values of the significant ones and see the contribution of each according to that value. I do not think what you did sound very good.

[Shanel]

I want to calculate eta squared for an ANCOVA. I know that SPSS provides the partial eta squared value, but presentation of this statistic is highly criticised (e.g., Levine & Hullett, 2002; Olejnik & Algina, 2003) so I want to present the more highly favoured eta squared value. I know that the formula for eta squared is (SSeffect/SStotal) aka (SStreatment/SStotal), but I don't know how to get the sum of square effect value.

Any help that could be provided would be greatly appreciated!!!