EDB 9080 Assignment 2

Dr. Satish Nargundkar

 

Part A [Something old, Something New]

Repeat the hypothesis tests done in class with the Helicopter data, with these small modifications. First, do it all in SPSS. For each of the following, clearly write down the following:

a.       Null and alternate hypotheses,

b.       The level of alpha you chose, and

c.       Your interpretation of the results. First indicate if you reject the null or not, and then say in plain English what that means in the context of helicopter flight times.

 

1.       Pick the data for any one of the helicopters. Test the hypothesis that its flight time is greater than 2.50 seconds on average (not 2.0 as we did in class).

2.       Pick any two of the helicopters, and test whether their average flight times are equal or different from each other.

3.       Perform an ANOVA to test for differences across all 4 helicopters in terms of average flight time. Remember that ANOVA will only tell you if at least one of them is different, but not which combinations of the four are different from each other. For that, select the Post Hoc testing option in SPSS under ANOVA, and select Tukey. Click here for a video that explains this.

4.       Pick any two helicopters. Perform an F-test for variances to test if they are equal or different.

Answer everything in a Word document - copy and paste the output from SPSS as needed.

 

Part B [Something borrowed (from the web), and Something (that might turn you) Blue]

1.       We discussed Alpha, the limit of Type 1 error (typically set at 0.05) in class, but did not discuss Beta, the Type II error. Write a paragraph or two explaining in your own words what Alpha and Beta are, how they related to the Null and Alternate hypotheses (there are plenty of online sources to help you here is one). What is the Power of a test? See the video on Power and Sample size I posted on the class website (or Click here).

 

2.       Critical Thinking: In the context of the criminal justice system (Innocent until proven guilty), how are both type I and type II errors sought to be controlled?