In: Statistics and Probability
Decide and state whether a one-sample z test, a two- sample z test, or a t test is appropriate. Whichever test you choose, state your null hypothesis, then obtain a P -value and formulate a valid conclusion.
1a.) 20 subjects who do not speak German are asked to participate in a trial for a new type of crash course in the German language. 10 subjects are randomly placed in Group A, to test the new teaching method, while the other 10 are placed in group B and given a more traditional crash course in German (though they are not told this is not the new approach). After both courses have ended, evaluators (who don’t know which course each subject took) determine that 50% of subjects in Group A are competent speakers of German, compared to 40% of subjects in Group B. Based on this, does the new method work better, or is the 10% difference due to chance?
1b.) 200 subjects who do not speak German participate in a trial for a new type of crash course in the German language. 100 subjects are randomly placed in Group A, to experience the new teaching method, while the other 100 are placed in group B and given a more traditional crash course in German (though they are not told this is not the new approach). After both courses have ended, evaluators (who don’t know which course each subject took) determine that 42% of subjects in Group A are competent speakers of German, compared to 38% of subjects in Group B. Based on this, does the new method work better, or is the 4% difference merely due to chance?