What I have understood is that it is a graphical technique used to check if 2 data sources (even if not from the same population) have similar data distribution. In that case the scatter plot will be kind of a straight line. I am not sure when the slope will be exactly 1 (45 degrees). Also is it possible to make a qq plot between 2 continuous variables with different ranges (e.g. the one provided). If so : will the x-axis (age) range from 20 to 70? Or 0 to 70? will the y-axis (salary) range from 5k to 10k? Or 0 to 10k?

  • 3
    $\begingroup$ You could draw such a quantile-quantile plot: in the simplest case, with the same number of values, plot. minimum versus minimum, second lowest versus second lowest, etc. The data will plot as a straight line if variables have the same distribution shape. Whether that helps any research goals in this case is an open question. You lose all information of which salary goes with age. There is no sense in which it's expected for salary and age that you have unit slope; at a minimum that's an expectation only for variables measured in the same units. $\endgroup$
    – Nick Cox
    Mar 16, 2018 at 12:22
  • $\begingroup$ Do you mean variables should be in the same unit or same range? $\endgroup$
    – Ad94
    Mar 16, 2018 at 12:33
  • 1
    $\begingroup$ Unit slope is a meaningless reference without the same units, so far as I can see. Conversely two variables can have the same distribution shape without having the same range. Consider variables uniformly distributed on two different intervals, for example. $\endgroup$
    – Nick Cox
    Mar 16, 2018 at 12:51
  • $\begingroup$ "What is a qq-plot": stats.stackexchange.com/questions/111010/… $\endgroup$ Jan 4, 2019 at 16:56


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.