Quantitative analysis plan

Warning

The information given on this page concerns only students who will carry out quantitative analyses (so it does not apply to students doing qualitative analyses).

Make sure you plan ahead and think about the quantitative analysis of the data you will collect for your study. The best time to do so is when you are planning your study/project. Do not wait until after you started collecting data.

Tip

We have compiled a checklist for you and your supervisor to go over at the beginning of your supervision, while thinking about the project you will carry out.

It is important that you are able to answer all of these questions if you wish to make the most out of the statistical support we offer and more importantly if you want to avoid irreparable issues later on.

1 The checklist

During the early stages of your project, bring this list of questions to your supervisor and discuss them. If, after talking to your supervisor, you are still unsure about any of these points, please book an appointment with Stefano.

  1. Can you clearly state, in simple language, what the research question(s) of your project is/are?
  2. If you are testing a specific hypothesis, did you formulate it so that it is a falsifiable statement? Note that it is OK if you don’t or cannot formulate a hypothesis! As long as you have one meaningful research question, you are all set.
  3. Have you clearly defined the concepts/objects of study in your research question and hypothesis?
  4. Have you operationalised the concepts/objects of study into clearly measurable variables?
  5. Which are your outcome (dependent) variable(s) and your predictor (independent) variable(s)?
  6. Your outcome variable(s) is of which type?
  7. Which statistical model or test will you use to answer your research question or test your research hypothesis? Note that you should consider statistical modelling or testing only if you have attended a course on statistics/quantitative methods.
  8. Have you specified and justified the minimal sample size you would need to obtain the required estimate precision or statistical power (if you will be doing statistical modelling or testing)?