CMN 3102 Lecture Notes - Lecture 9: Null Hypothesis, Statistical Hypothesis Testing, Confidence Interval

• School: University of Ottawa
• Department: Communication
• Course: CMN 3102
• Professor: Elizabeth Dubois

Document Summary

Cmn3102 - hypothesis testing and chi-square - 28/02/20. You need to formulate two alternative hypotheses for your research. An alternative hypothesis is the prediction that there is a relationship or there is a difference that has not occurred by chance or random error. Null hypothesis is the counterpart of the alternative hypothesis where there are zero relationships or differences. Clumping the dependent and independent variables together allows us to test mathematically the hypothesis. Relying on people and numbers is tricky and it"s often impossible to study a whole population so researchers have to rely on samples instead. For something to be statistically significant, it should be accurate at least 95% of the time when tested. You can create a confidence interval by using multiple variables. Testing for significance: determine if the difference or relationship has occurred by chance. Step 1: set the probability level find the confidence level. Step 2: conduct a statistical test to find the calculated value.