Linear Regression Interpretation

#1
Hi there,

I am running a Multiple Linear Regression analysis. I have 2 predictor variables (self-worth and self-acceptance) and 1 criterion variable (self-esteem).
As hypothesized, self-worth and self-acceptance reliably predicted the scores on self-esteem (eg. higher scores on the predictors would equate to higher scores on the criterion).

However, by simply looking at the means for each construct, it is evident that self-worth (predictor) has a higher mean than self-esteem (criterion).
Could this be signifying a different/more complex relationship between this predictor and the criterion? Or perhaps it makes more sense to suggest that self-worth is the criterion and self-esteem is the predictor?

I do hope this makes sense.

Thank you,
Vanessa.
 

CowboyBear

Super Moderator
#2
However, by simply looking at the means for each construct, it is evident that self-worth (predictor) has a higher mean than self-esteem (criterion).
Surely these constructs are measured on different scales? I can't imagine a way in which one could meaningfully compare their means.

Or perhaps it makes more sense to suggest that self-worth is the criterion and self-esteem is the predictor?
Which is the predictor and which is the criterion depends on your research question or hypothesis, not on the sample means.
 
#3
Yep - all 3 constructs are measured on different scales.

And yes, I have determined the predictors and criterion based on my research/literature review. Which was supported... but those means have just made me question my interpretation.
 

CowboyBear

Super Moderator
#4
And yes, I have determined the predictors and criterion based on my research/literature review. Which was supported... but those means have just made me question my interpretation.
I can't really follow your line of reasoning here, sorry. :confused: The fact that self-worth has a higher mean doesn't tell you anything about whether it should be a predictor or a criterion.
 
#5
I suppose that answers my question either way.
I was just worried that because one of my predictors had a greater mean than the criterion, that would effect my interpretation of its contribution/relationship with the criterion. Thank you!