Hello,
I have ran a regression where one of the terms is quadratic
command:
reg log(y) x1 x2 x2^2 x3 x4 x5 x6
Then I want to plot it such that the relationship between y and x1 is shown in that it increases in at a diminishing rate (concave)
So, I used this command:
twoway qfit y x
Indeed, it showed the trend, however, I am wondering if this was the correct method to do so.
Also, the y variable is in logs, but I want to show them in units, which in this case is done by making the log to the power of e (e^log(y)). But, when I do the graph using y instead of log(y), the concave trendline is gone and instead it is convex and upward sloping.
Many Thanks
I have ran a regression where one of the terms is quadratic
command:
reg log(y) x1 x2 x2^2 x3 x4 x5 x6
Then I want to plot it such that the relationship between y and x1 is shown in that it increases in at a diminishing rate (concave)
So, I used this command:
twoway qfit y x
Indeed, it showed the trend, however, I am wondering if this was the correct method to do so.
Also, the y variable is in logs, but I want to show them in units, which in this case is done by making the log to the power of e (e^log(y)). But, when I do the graph using y instead of log(y), the concave trendline is gone and instead it is convex and upward sloping.
Many Thanks