5

I have two graphs that are part of a multi-panel figure. It is clear from both the caption and the graphs that the axes mean the same thing. Should I repeat the axis labels in both graphs, or should I leave them out in one graph to avoid redundant information?

enter image description here

2
  • 7
    I should be able to tell at a glance what the graph is showing. The caption is not sufficient. The text may well be on some other page, so is not useful at all. And, frankly, in your example there is little to indicate that the right panel has anything to do with the left panel, that is itself quite busy.
    – Jon Custer
    Apr 23 at 17:38
  • 2
    "It is clear from both the caption and the graphs that the axes" wrong assumption: it is not clear from the graphs themselves...
    – EarlGrey
    Apr 24 at 12:25

3 Answers 3

28

Keep the axis labels. It may be obvious to you that they mean the same thing, but it may not be obvious to a reader. Also, if somebody wants to excerpt one of the panels to discuss in isolation, it's good if the axis label goes with it.

The one exception to this that I would give is if you are packing together an array of graphs (as in the example below, excerpted from this paper), in which case the graphs can be treated as a single panel and the "interior" labels omitted.

Example of a 5x2 array of graphs in a single figure panel

15

It's often largely a style thing, but since your two y-axes don't share the same scale, the fact that they're the same variable isn't as clear as it normally would be. For this particular case, I recommend including a label.

An alternative would be to be very descriptive in the caption of the figure.

4
  • 2
    On the other hand, if OP were to go without the axis label, I would definitely suggest adjusting the scales so that they match Apr 24 at 10:27
  • @user2390246. I considered that, and it would deemphasize the rising slope, which is probably why the scales are different. Apr 24 at 12:02
  • 2
    @ScottSeidman In some/many(?) cases, if you need to change scale to emphasise a rising slope, you've not got a rising slope that's worth emphasising.
    – TripeHound
    Apr 24 at 15:44
  • @TripeHound you're right about that. If I were the referee, and that slope was the basis of any interpretation, I'd expect to see a p-value for the slope. Apr 24 at 15:48
2

To add to the reasons why should label the figure axis: you (or other people) may wish to use the RHS figure in other contexts (e.g., in a talk) where it does not appear alongside the LHS one.

2
  • 1
    Bugbear of mine: if you are using a figure in a talk, make a new version! If figures are copied straight from a paper into a talk, they usually end up with unreadable font sizes. Apr 24 at 10:30
  • 1
    @user2390246 That's because a lot of people make them with unreadable font sizes in the papers too --- let your figures breathe, people!
    – jakebeal
    Apr 24 at 13:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .