Exercise 12: Showing equivalence of regression, t-tests, and ANOVA

Part 1: Do a dichotomous coding of variable "observation time" below to find out when pool (i.e., billiards) play is most effective: early in the evening or late in the evening. Imagine you've conducted a variety of independent polls drawn from observers of pool play using a 1-7 Likert scale (which we will treat as a continuous variable). Dichotomize "observation time" and run the simple regression model. Interpret the coefficient with reference to your early/late variable. Write it up the way we have in past exercises, in the regression style.

Person,ObsTime,PoolPlay
1,early,5
2,early,6
3,early,3
4,early,4
5,early,5
6,late,2
7,late,3
8,late,3
9,late,4
10,late,1
(Indeed, part of this exercise is figuring out how to get weird data formats into SPSS...)

Part 2: Also run this as an independent-samples t-test and make reference to some of your SPSS output to show the equivalence of the regression model to the t-test. (And yes, report some numbers.) In a separate paragraph: Write it up fully (but concisely) in the t-test style, as well.

Part 3: Run the above analysis as a one-way ANOVA and write it up in the ANOVA style. In a separate paragraph: As in Part 2, also make some casual notes (with numbers) showing some equivalence in your output to Parts 1 and 2.

Part 4: Finally, with reference to one of these sets of results, offer a sentence or two in plain English explaining why the results look the way they do (hypothetically, of course, as these data are thoroughly artificial...). I'm making reference here to the pool play itself, not the equivalence patterns or anything else -- i.e., why the results are such that they are because of some hypothetical explanation of early/late contexts of pool play.

Send to psyc7302@gmail.com with subject line "7302 Exercise 12."