Research Hypothesis

Updated 17 Sep 2026

•
What it is
- •
  a testable prediction about a relationship, difference, or pattern
•
Statistical formulation
- •
  $H_0$ describes the null condition being tested
- •
  $H_1$ describes the alternative
- •
  specify the parameter, direction where appropriate, and analysis before inspecting results
•
Example
- •
  for paired performance measurements, let $\Delta$ denote the population mean improvement
- •
  a directional comparison can use $H_0:\Delta\leq0$ versus $H_1:\Delta>0$
•
Caution
- •
  “statistically significant” is an analysis outcome, not the substantive phenomenon being predicted
- •
  failure to reject $H_0$ does not prove $H_0$
- •
  exploratory and qualitative studies need not force their questions into null-hypothesis tests
•
Related
- •
  Research Question
- •
  Analysis Plan

Linked references 3