.. _frame_question:

{{ header }}

*****************
Frame a Question
*****************

.. currentmodule:: BPt

The very first step in any workflow, before loading data or anything else, should be to frame a research question of interest in terms of a prediction.
For example, let's say, our question of interest was to investigate age-related changes in cortical thickness.
A simple predictive re-framing could be, how well can cortical thickness predict a participant's age?
What if we had longitudinal data per participant? Then maybe we could ask, how well does cortical thickness predict
age at time point 1, what about time point 2, and so on? The key pieces of information to identify after composing a question of
interest are: What are the input variables to the prediction? What variable(s) are being predicted? And, are
there any other variables which might influence this prediction in an undesirable way, e.g., potential confounding variables.