I created boxplots for 5 traits to show the spread of true age for each variant of each trait in both samples between 2 observers:

enter image description here

A professor suggested that I should use scatterplot of Price vs Kim with symbols representing certain age ranges. Would this be a better choice?

  • 1
    $\begingroup$ Boxplots and scatterplots convey different information. What do you want to learn from your plots? $\endgroup$
    – Dave
    Mar 11, 2023 at 3:13
  • $\begingroup$ @Dave Basically the spread of true age for each variant of each trait in both samples $\endgroup$
    – user34930
    Mar 11, 2023 at 3:54

1 Answer 1


A scatterplot would be most useful for assessing for well the two observers agree (and draw the attention to this comparison).

It's hard to say which one would be more useful in your case. Try looking at scatterplots and seeing whether they convey what you need. You would need to replace each one of your five panels by four scatterplots (please keep the scaling constant within each variant). If you have a gold standard (true age), indicate that within each scatterplot.

If you decide to stick with this visualization, I would consider removing the boxplots, since you do not really have a lot of data (again, try it and see what works). However, reduce the horizontal spread between points within a variant and observer, and increase the spread between adjacent variants. And again, if you have a "true" value, indicate it in the plots.


Your Answer

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

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