Test claims
Determine whether a treatment truly works or the result is just chance.
Dental research
Compare fluoride treatments, evaluate new materials, assess therapies.
Make decisions
Use p-values and significance levels to draw reliable conclusions.
Be clear about the outcome, the comparison, and the population. A precise question makes the rest of the analysis much easier.
Match the test to the outcome type, study design, and whether the data are paired, independent, categorical, or continuous.
A small p-value suggests evidence against H0, but it does not tell you whether the effect is large, useful, or clinically important.
Good reporting includes the test statistic, p-value, confidence interval, and a sentence that explains what the result means for dental practice or research.
Think about a t-test when comparing average plaque index, bond strength, or gingival scores between two groups.
ANOVA is usually the starting point when several treatment groups are compared at the same time.
Use chi-square style thinking when the question is about categories such as disease present vs absent or implant success vs failure.
Before-after measurements on the same patient need paired methods, not independent-group tests.
Click any card to explore with interactive examples.
Choose a dental scenario and watch the full testing process step-by-step.
Drag the slider and see how p-value interpretation changes in real time.
Click each cell to see dental examples. Understand how decisions can go wrong.
Run a simulated clinical trial and watch every step of the hypothesis testing process.
| Concept | Symbol | Meaning | Common Values |
|---|---|---|---|
| Null Hypothesis | H₀ | No effect or difference exists | - |
| Alternative Hypothesis | H₁ | A real effect or difference exists | - |
| Significance Level | α | Max acceptable probability of Type I error | 0.05, 0.01, 0.001 |
| P-value | p | Probability of data given H₀ is true | p < 0.05 = significant |
| Type I Error | α | Rejecting a true H₀ (false positive) | 5% when α = 0.05 |
| Type II Error | β | Failing to reject a false H₀ (false negative) | 20% when Power = 0.80 |
| Statistical Power | 1 - β | Probability of detecting a real effect | 0.80, 0.90 |
Always report
Effect size + p-value + confidence intervals for transparent research.
Sample size matters
Small samples reduce power. Plan adequate n before your study.
Not proof
Failing to reject H₀ does not prove it true — only insufficient evidence.
Statistical significance says the finding is unlikely under H0. It does not automatically mean the effect is large enough to matter clinically.
A study can miss a true effect because of low sample size, high variability, or weak measurement quality. Always look at confidence intervals too.
Choosing alpha after seeing the result makes the conclusion less trustworthy. Pre-specification protects against flexible interpretation.
Even when a difference is statistically significant, ask whether it changes patient outcomes, treatment decisions, or policy in a meaningful way.