Bisector

[Federico Nutarelli]

Hi Statlist!

I have the two following variables which I need to plot one against the other:

gen tasso_crescita_lag_dummy = log(tot_sales)-log(L.tot_sales) if recalled_dummy == 1 
gen tasso_crescita_lead_dummy = log(F.tot_sales)-log(tot_sales) if recalled_dummy == 1

Summary statistics are as follows:

Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
tass~g_dummy | 203 .064279 .5498735 -4.239164 4.046771
tass~d_dummy | 180 .0340018 .717771 -6.259137 4.618601


What I would like to do is to generate a scatter plot with a proper bisector and not like the one generated by the following code which gives me a rectangular cartesian plane:

twoway (scatter tasso_crescita_lag_dummy tasso_crescita_lead_dummy)|| function y = x, ra(tasso_crescita_lag_dummy) clpat(dash) title(Growth rate of sales:) subtitle("tasso crescita t; t-1 vs tasso crescita t, t+1")  ytitle(Tasso crescita "t-t+1") xtitle(Tasso crescita "t-t+1") legend(order(1 "Scatter points" 2 "Bisector") col(2))

Is there a way?


Thank you,

Federico

[Nick Cox]

What's a "proper bisector"?

[Federico Nutarelli]

Oh I meant a 45 degree line in a square plane (while usually STATA do it rectangular..I guess it also depends on the title length)

[Nick Cox]

If you ensure the same range on both axes and aspect(1) then function y = x will yield a 45 degree line.