Help with regression model

#1
Hi,

I am trying to understand whether there are any problems with a regression model I'm using for research. I'm interested in the effect of quality of information about the wealth and income distributions on people's stated preferences for government redistribution of wealth and income. I have two main variables: estgini (an individual's perceived gini coefficient for the US wealth distribution, so it will be a number between 0 and 1) and idealgini (an individual's stated 'ideal' wealth distribution, if they could choose what the wealth distribution looks like ideally). People generally underestimate wealth inequality in the States. The actual gini (for my purposes) is approx 0.7, while the average estgini is 0.4, and the average idealgini is 0.26.

I then define two new variables: i) eiginidiff = estgini - idealgini; in words, the extent of wealth redistribution away from their perceived status quo that an individual prefers, and ii) infogap = 0.7 - estgini, i.e. how accurate a person's information about the wealth distribution is.

In principle, I'm interested in the effect of infogap on eiginidiff, i.e. whether better informed people tend to prefer more redistribution.

The correlation between estgini and idealgini is 0.36, and the correlation between eiginidiff and infogap is -0.58. Basically what I'm worried about is that with this model I won't be able to make any actual inferences about the effect of 'infogap' on 'eiginidiff', i.e. whatever coefficient I get will just be capturing the correlation between estgini and idealgini.

Not sure if my fears are well founded? If anyone has any ideas or experience with a similar situation would love to here from you.

Thanks