Statistics

Part One: Confidence Intervals

1. State your research question in one sentence.

2. Present the point estimate for the population mean of Y.


3. Present and interpret the 95% interval estimate for the population mean of Y.

4. Present the point estimate of the population mean of X if it is a continuous variable, the population proportion if X is a nominal variable.

5. Present and interpret the 95% interval estimate for the population mean if X is continuous, the proportion if the independent variable is nominal.


Part Two: Hypothesis Testing

6. Present a hypothesis test for the dependent variable (Y) in your research project. For some variables there will be a logical test to conduct. For example, suppose Y is the unemployment rate in Michigan. One test that comes to mind is whether Michigan’s population mean unemployment rate is different from the national rate. Another test would to be see if it is higher the natural rate of unemployment. For some variables, however, a logical test won’t be readily apparent and you will need to just make up some test to demonstrate that you know how to do a hypothesis test of the mean. I know this seems weird and artificial, but remember the point is to demonstrate that you know how to conduct a hypothesis test.


7. Present a hypothesis test for the independent variable (X) in your research project. For some variables there will be a logical test to conduct. For example, if you have a nominal variable (e.g., X = 1 if a college graduate vs. X = 0 if not a college graduate) you can test if the majority of the population has a college degree (is the population proportion greater than 50 %?). However, for some variables a logical test is not readily apparent and you will need to make up some test just to demonstrate that you know how to do a hypothesis test (of either the population mean or the proportion). Again, I know this is weird and artificial, but remember the point is to demonstrate that you know how to conduct a hypothesis test.

8. Test the theory using a two-tailed test of the population correlation. Does the sample support the theoretical model’s prediction? Is the correlation coefficient significantly different from zero? This test actually gets at whether or not there is evidence that the theoretical model works (not artificial!).