| HYPOTHESIS TESTING FOR PROPORTIONS |
| | | | | | | | | | | Comparing two independent means: |
| 1. CLAIM concerning a population proportion, will contain <, > , not equal ¹ | | | | | | | p < "value" | | | H0: p1 - p2 = 0 |
| 2. Null Hypothesis | | H0 | statement of equality, p = "value" | | | | | | | H1: p1 - p2 < 0 or > 0 |
| 3. Alternative Hypothesis | | | H1 | same as claim |
| | | | | < | left-tailed test | | | | | p-pooled = (x1 + x2)/(n1+ n2) |
| | | | | > | right-tailed test |
| | | | | ¹ | two-tailed | | | | | SE = SQRT(p-pooled*(1 - p-pooled)/n1 + p-pooled*(1 - p-pooled)/n2) |
| | Normal probability distribution | | | | based on assumption in the Null hypothesis |
| | | bell curve |
| 4. Choose significance level | | | is the complement of the confidence level |
| | for example, a confidence level of 95% is equivalent to a significance level of 5% |
| | confidence level defined the "likely" values |
| | significance level defines the "unlikely" values, the probability of being in the tail. |
| 5. compute the value of the sample statistic p hat = x/n |
| | Compute the standard error | | | SE=sqrt(p*(1-p)/n) | | Use p from H0 |
| 6. Compute probability value or p-value |
| | p-value = P(sample statistic <= computed sample statistic) | | | | | | left-tailed test | | NORM.DIST |
| | p-value = P(sample statistic >= computed sample statistic) | | | | | | right-tailed test | | 1 - NORM.DIST |
| | p-value = twice the area of the tail defined by p | | | | | | two-taield |
| 7. Initial Conclusion | | | Reject Null or Do not reject Null/Support Null |
| | | | Reject Null and support Alternative=Claim |
| | | or | Do not reject Null and do not support the claim |
| | rejection rule: | | Reject the Null if p-value is less than significance level; otherwise, fail to reject the Null |
| 8. Final Conclusion | | | Sentence: | The sample data support the claim that…. |
| | | | or | The sample data do not support the claim that… |