Individual versus aggregate regression


Fortran must die
What is the difference between doing regression at an individual versus a aggregate level [I am particularly interested in how this influences interpretation of slopes]

This is an example of what I am talking about for individual level data.

Yisb =β0 + β1 Xisq+ β2 Xq+ eisq

Yisb is the outcome variable for individual i in WIB s in year-quarter q
Xisq are the individual attributes for person i in area s in year-quarter q
Xq is a quarter dummy variable?

The only type of regression I have personally seen is apparently the aggregated form [I have seen analysis done at the individual level, but only in ANOVA].