PSCH 343 Lecture Notes - Lecture 6: Null Hypothesis, Type I And Type Ii Errors, Statistical Hypothesis Testing
29 Feb 2016
School
Department
Course
Professor
Document Summary
Hypothesis testing, again: null and alternative hypotheses mutually exclusive: Null hypothesis: there is no difference between your populations, the difference in means is just due to random sampling. Reject null hypothesis support the alternative (research) hypothesis. The procedural steps: complete the 5 steps to calculate a test statistic (e. g. , z-test) and compare it to a previously set cutoff (e. g. , p < . 05, if the statistic exceeds the cutoff, reject null hypothesis. Results support research hypothesis: if the statistic does not exceed the cutoff, retain null hypothesis. Results could be due to sampling error: cutoff = critical value. What will change: the way we phrase null/alternative hypotheses, the comparison distribution and its characteristics, the statistical test to compare the means of the populations. Restate the question as a research hypothesis and a null. Step 2: determine the characteristics of the comparison distribution: this involves some calculations, depending on the specific test hypothesis about the populations.
