How to Fit Non-Linear Regressions with Scatter Plot

[Xinghuan Luo]

Dear Statalists,

I really need help now. Thanks in advance.

In the below code chunk, I would like to fit a non-linear regression that has 2 variables on a scatter plot. That is to say, I would like to put a scatter plot and a regression plot into the same graph. But, my question is that twoway lfit can only be used with one independent variable when fitting on a scatter plot. But I actually has two independent variables, length and log_length. I also want to add a pweight option in the regression.

The x axis of the graph in my mind would be length and the y axis would be price, I just do not how to achieve this. The error I got is always "too many variables specified: price length log_length".

Thank you so much!

sysuse auto.dta, clear

reg price length

gen log_length = log(length)

local idpt_list length log_length

twoway scatter price length || lfit price `idpt_list', pweight=headroom

[Maarten Buis]

sysuse auto.dta, clear
gen log_length = log(length)
reg price length log_length
predict yhat
scatter price length || line yhat length, sort