Course contentsShow
← Data Science
Lesson 356 of 2,14510. Hypothesis Tests for MeansFree lesson

Making Decisions and Stating Conclusions

Comparing p-values to significance levels and translating statistical decisions into practical conclusions about the population mean.

Making Decisions and Stating Conclusions

You'll learn how to translate a p-value and significance level into a clear statistical decision and practical conclusion about your population mean.

The Decision Moment

After calculating your test statistic and finding the p-value (the probability of observing data this extreme if the null hypothesis were true), you face a simple yes-or-no decision:

Compare your p-value to your significance level (α):

  • If p-value ≤ α: Reject the null hypothesis (H₀)
  • If p-value > α: Fail to reject the null hypothesis

Think of it like a courtroom: you set a standard for "reasonable doubt" (your α, say 0.05) before the trial. When the evidence (p-value) meets that standard, you reject the presumption of innocence (H₀).

From Statistical Decision to Practical Conclusion

The decision is mechanical, but the conclusion requires context:

When you reject H₀: "There is statistically significant evidence that [describe the alternative hypothesis in context]."

Example: "At the 0.05 significance level, we have sufficient evidence to conclude that the mean customer satisfaction score differs from 7.0."

When you fail to reject H₀: "There is insufficient evidence to conclude that [alternative hypothesis]."

Example: "We do not have sufficient evidence to conclude that the mean customer satisfaction score differs from 7.0."

Critical Points

  • Never say "accept H₀" — only "fail to reject"
  • Always reference your significance level
  • Connect conclusions back to the real-world context, not just numbers
  • A non-significant result doesn't prove H₀ true; it means the evidence wasn't strong enough

Key Takeaway: Compare your p-value to α to make a binary decision (reject or fail to reject H₀), then translate that decision into a clear, contextual statement about what the data suggests regarding your research question.