Hi all,

I am trying to understand the non-linear relationship of age in my model using fractional polynomials.

csh_sh = b0 + b1age + b2incometh + b3educat

I ran the following code:

Code:
twoway fpfitci csh_sh age
There is theoretical explanation for the “dip” and so I want to explore this further with a fractional regression.

Code:
fp <age>, scale: reg csh_sh age incometh i.educat <age>
fp plot, residuals(none)
Q1) Is the code used correct?

Q2) What is the “component” on the Y-axis? I’m not sure i understand what this represents as it is different to “linear prediction” in the margins command. As age increases, the second graph show an upward trend but I am not sure how to quantify this.

Q2) Are there tests/code I can run to prove this curve produced is adequate? I can see already that the confidence intervals (which I assume is the grey shaded area) are very wide.

Thanks for your help in advance!