Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Make violin plots without having equal numbers in each group?

Hi,

I'm trying to make violin plots and am having a little trouble. I have two groups with data points across three time points. I'd like to graph the three time points (each with a different colour) and have some indication of which data points correspond with which groups. I'm having some trouble arranging my data in a way that facilitates this -- could someone explain how I can do this using the new JASP plotting tools? Also, as the second time point was completed by participants at home, the response rate is only about 50% and therefore the data points aren't even between groups -- JASP is saying this is a problem and won't allow me to split the data by group because of it.

Any help would be hugely appreciated!

Thank you!
Sarah

Comments

  • Hi Sarah,

    The box plot is a recent addition and we are still working out ways to extend it. That error message sounds like it might be a bug though (different numbers of response should not be a problem) -- can you report it on the JASP GitHub page so that the programmers can follow it up?

    While on GitHub, you can also request whether a future release can split the box plot according to more than one factor.

    Meanwhile, you could present six box plots to show all of the data, and use the plotting facilities in classical ANOVA to produce an more standard graph.

    Cheers,
    E.J.

  • Hi E.J.,

    That's really helpful, thank you! I'll report this to GitHub. In the meantime -- when I create the plots it puts the data for each time point in a separate plot (i.e., rather than having multiple in different colours across the X axis). Is it possible to arrange the data so they will all be on one graph?

    Cheers,
    Sarah

  • If you've got a variable "time" ("1", "2", "3", say) and you pick split by "time", then it should work. But it is difficult to convey these things without the benefit of a drawing, so maybe I misunderstand.
    Cheers,
    E.J.

Sign In or Register to comment.