Statistics 744, McMaster University

When you ask a question we (JD and/or BB) will often say “Can you send me a (minimal) reproducible example?” (these are sometimes called “reprexes” or abbreviated MWE (minimal working example) or MRE (ditto, reproducible)).

• Reproducible means something we can run for ourselves to replicate your problem on our computers. Sometimes we can just look at your code and see what’s wrong, but most of the time we can’t. In particular:
  • you should be able to run your code in a clean R session (RStudio: Session > Restart R) and see the same problem.
    • make sure you don’t encounter any errors before the location of your main problem (e.g. because of unloaded data or unloaded packages causes an error before that point)
    • please tell us specifically what error/warning/wrong answer you’re getting¹
  • if you e-mail/post your code, please do so as text (i.e., cut and paste from your script file), not as a screenshot
    • screenshots are awkward if we want to cut and paste code (which, again, should be reproducible)
    • they’re also inaccessible to visually impaired people, FWIW
• Minimal means to do whatever you can to make your example short, to save us the trouble of digging through/running lots of code to find the problem.

• Some further resources.

  Many of these resources emphasize creating self-contained MREs, i.e. not depending on external data files etc.. That’s less important for this course, where JD/BB have access to your Github repository. You can tell us e.g. “my MRE is in file ‘mre.R’, it uses ‘mydata.csv’, everything is on my repository”. Self-contained MREs are better - they’re generally shorter, and anyone in the course can try them out without access to your repo if you post them to the Teams channel - but not absolutely necessary for this course.
  • R-specific Stack Overflow question about reproducible examples
  • General Stack Overflow page on MREs
  • presentation by Sharla Gelfand
  • the reprex package, an R package for creating reprexes (web page)
  • a very old (but famous) post on “how to ask questions the smart way” (mostly more general about asking advice on technical forums)