Hypotheses versus expectations

Hypothesis Format

  • If you have a strong theory, you can cleanly predict what happens to Y if you change X.
  • In that case, it makes most sense having formal hypotheses in your paper.
  • For example, see Datta, Foubert and van Heerde (2015, paper here)


  • If your theory is not strong, or your predictions are frequently going in both ways (e.g., both positive and negative), it’s wiser to not write out formal hypotheses but rather expectations.

  • For example papers, see Datta, Ailawadi , and van Heerde (2017, paper here or Datta, van Heerde, Steenkamp, and Dekimpe (2022, paper here).

  • Alternatively, a paper can very well do without any hypotheses or expectations, see Datta, Knox, and Bronnenberg (2018, paper here). Curious about how to write such so-called “empirics-first” papers? See Golder et al. (2022, paper here).

Important: choose a format that fits your research!