Regression Hypothesis

#1
I am not good at statistics that is why I am seeking help in this forum. Thanks for any help in advance.

My research hypothesis is
Student will score high on optimism, this optimism will be highly correlated with more positive attitude towards fat children, and it will further enhance their friendship. All these variables are measured on continous scales.

One of my friend who is good at statatistics has suggested me to use (moderated) hierarchical regression analysis. He has asked me to use optimism and friendhisp (in step 1) with attitude as dependent variable. And interaction term (optimism*friendhip) in step 2. He has asked to check whether coefficient of interaction term (in step 2) and optimism (in step 1) are significant.

Is it right approach to check this hypothesis? I will appreciate any help.
 

CB

Super Moderator
#2
From your hypotheses, it sounds more to me like to you may have just one independent variable (optimism) that you think has an effect on two different dependent variables (positive attitude, friendship). Does that sound right?

(Though I'm not sure exactly what it means to measure "friendship"?)

I just want to highlight something here:
Student will score high on optimism, this optimism will be highly correlated with more positive attitude towards fat children
You probably need to think carefully about what it means to score "high" on optimism (what's a high optimism score?); and how large a correlation you would need to see to support the hypothesis that optimism and attitude to fat children are "highly correlated". Note that a statistically significant correlation doesn't necessarily imply that the correlation is high, and vice versa.