2
$\begingroup$

I am building qq plots with a statistics package, however I notice my output is different than what I found in the econometric literature. For example, here is a qq plot from a publication I came across:

enter image description here

In this one the standardized residuals are on the Y axis. However, when I ran my package's built-in method for this kind of qq plot I got the axes switched (standardized residuals on the X axis). As seen above the labels on the literature's is simply "Standardized Residuals", whereas in the following Graph the label is "Quantiles of Standardized Residuals".

enter image description here

My software does not have any arguments for which axis it goes on.

My Question is: Is there a conventional rule for which axes the theoretical quantiles/ standardized residuals should go on? Or is it trivial? Also, is it conventional to include "Quantile" on the standardized residual axis space, or is it implied? Or are they two completely different things?

$\endgroup$
  • 4
    $\begingroup$ It's trivial in that if you label your axes it is clear what you've plotted. On the other hand, there's certainly a clear convention that I expect with plots (as described in the second-last sentence of this answer -- that when one quantity is random and the other is fixed, the random quantity goes on the y-axis). I find it somewhat surprising that so many programs ignore this convention when doing QQ plots but I expect they have some reason for doing it. $\endgroup$ – Glen_b Oct 14 '17 at 6:11
  • 3
    $\begingroup$ I agree with @Glen_b and note further that the axis choice you desire matches also a common convention that observed responses or outcomes are plotted on the vertical axis. My customary software, Stata, which is popular in economics and econometrics, does it as you wish. A similar small issue is how one interprets versus: I was brought up to say $y$ versus $x$, temperature versus time, or whatever, but the opposite wording is frequent. stats.stackexchange.com/questions/146533/… $\endgroup$ – Nick Cox Oct 14 '17 at 6:45
  • 3
    $\begingroup$ But, back to the point, if it's not trivial to write a few lines of code to get the graph you want, your software is not a good choice. $\endgroup$ – Nick Cox Oct 14 '17 at 6:45
  • 1
    $\begingroup$ I know you've said you are trying to use the stats package, but why not try ggplot2? sthda.com/english/wiki/… $\endgroup$ – Travis Oct 14 '17 at 8:51
2
$\begingroup$

It's trivial in that if you label your axes it is clear what you've plotted. On the other hand, there's certainly a clear convention that I expect with plots (as described in the second-last sentence of this answer -- that when one quantity is random and the other is fixed, the random quantity goes on the y-axis). I find it somewhat surprising that so many programs ignore this convention when doing QQ plots but I expect they have some reason for doing it.

-Glen_b

$\endgroup$
  • $\begingroup$ I've copied this comment by Glen_b as a community wiki answer because the comment is, more or less, an answer to this question. We have a dramatic gap between answers and questions. At least part of the problem is that some questions are answered in comments: if comments which answered the question were answers instead, we would have fewer unanswered questions. $\endgroup$ – mkt Jun 21 at 20:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.