JOURNAL CLUB: Turning tables into graphs.
Bob O'Hara
Tuesday, 27 May 2008 18:56 UTC
Inspired by the various comments about graphs, I thought it was worth bringing up this paper:
Gelman A., Pasarica C., & Dodhia R. Let’s practice what we preach: turning tables into graphs. American Statistician 56, 121-130.
I think it is worth reading as an encouragement to draw figured rather than write tables: they are generally easier to read. Sometimes a bit of thought is needed, and a few blind alleys explored, before a good figure is produced.
I’m curious about this, though. How easy do other people find it to interpret the figures? Would you draw any of them differently?
-
Replies
-
Bob, this paper is good reading. I knew that figures are usually preferable to tables, but the authors make the point that all tables with quantitative data can be turned into figures. I have to remember this in m next paper.
And the authors state that Microsoft Excel is just not flexible enough for their needs. Tips for the right software to create good figures would be discussion in itself. And there is also a Nature Network group for that.
-
There is also a long comment thread at Mind the Gap blog which contains some useful tips and links on alternatives to MS Excel.
-
Nice paper. But who would ever think of publishing a table as the one accompanying Figure 7?
-
Mehta et al. did.
In older stats papers particularly, the raw data is presented in tables so that other people can use it too.
-
This is sometimes done in online-only Supplementary Information, where tables can be provided in downloadable formats for other people to analyse and re-use. I am aware of one format being Excel (sorry!) but I am sure it is done in others, also.
-
