Null Hypothesis
Definition
The null hypothesis is a formal assumption that there is no effect, no difference, or no association between variables being studied.
Correct Scientific Usage
Researchers explicitly state null and alternative hypotheses before conducting studies. Statistical tests evaluate whether observed data are sufficiently inconsistent with the null hypothesis to justify rejecting it. Researchers do not prove hypotheses true; they test whether data contradict a specific null assumption.
Common Misunderstandings
Many people assume that rejecting the null hypothesis proves an alternative explanation, when it only provides evidence against the null.
Similarly, failing to reject the null hypothesis is often misinterpreted as proof that no effect exists, rather than simply insufficient evidence of an effect.
Why It Matters
Because much of scientific inference is framed around rejecting the null hypothesis, results can appear more decisive than they are. Understanding this framing clarifies why statistical results often require cautious interpretation.
References
- Hypothesis Testing, P Values, Confidence Intervals, and Significance, StatPearls
- Absence of evidence is not evidence of absence, Journal of Applied Oral Science
Related Terms
Related Articles
- What Statistical Significance Actually Tells Us
- What 'Backed by Science' Really Means
- Why One Study Is Almost Never Enough