Getting advice on data analysis for your research question can be tricky. Sometimes it can be unclear to yourself what you are asking. Sometimes you are so in the weeds that you forget what someone outside of your project needs to know to understand the problem that you are asking for advice on.

To help make discussion flow better in stats beerz and to help everyone (asker, listeners, and advisors) get the most out of a question, here is a template of asking for statistical advice. Of course your question could come up at any of the steps here and don’t feel that you need to conform to this but it’s a potential place for you to start! Especially if you’re a bit nervous about asking a question.

One key is brevity as we want to start with the high level and delve into the weeds as needed. By laying just the groundwork the group will be better prepared to discuss.

1) What is your ecological question? (1-2 sentences –> See Bolker pg 23)

  • Can you describe your question conceptually? e.g., Are ecological tipping points labile?
  • Can you describe your question specifically? e.g., Does the interaction between temperature and light limitation shift the 10% mortality threshold in seagrass species X?
  • Do you have a hypothesis?

Note: this is not your entire thesis, this is the conceptual/specific question that you are directly interested in asking

2) What is your experimental design? (2-3 sentences)

  • Are your data hierarchical?
  • How many samples do you have?
  • Would it help to sketch it on the white board?
  • How many variables do you have?
  • What kind of data do you have? discrete? continuous?

3) What is your statistics question?

  • What statistical method are you using (or thinking of using) and why?
  • If you have a model, can you write your model on the board (in words)? e.g., Bird abundance ~ temperature + territory size + salmon biomass
  • What is your goal for the study: prediction or description - e.g., choosing top model, is the purpose prediction or description