Causality and inference:In May 1998, the New York Times carried an article by Adam Bryant describing a research project to relate two rather diverse phenomena: the performance of a Fortune 500 company, and the golf game of its CEO. Before we go further, you should read this article.
Bryant, Adam (May 31, 1998). Duffers Need Not Apply, The New York Times. Section 3; Page 1; Column 1 You may use our library to retrieve it . You may click here for a quick access.
Are you convinced now? If so perhaps it is time to pack up the statistical tools in their tool bag and break out the golf bag. Or perhaps it's just time to sharpen those statistical tools a little, and see just what underlies this case.
The data described in the Bryant study can be found here using this SPSS data set already created.
The Bryant article reports its conclusion based on a subset of 44 cases from the 51 in the data set. This time we will use the
whole dataset. Let us seperate them into three groups. This involves identifying 14 low performing companies, 25 middle performing companies, and 12 high-performing companies based on StockRate.
The next step is to decide whether these differences are real or not. The article makes the claim that they are; see what you find. A good first step would be to calculate confidence intervals for the means of the three groups. Prepare graphs of confidence intervals for the means of handicap for each of the groups. Interpret the results.
Now let us try the t-tests. You should conduct independent samples t-tests to decide whether the differences between the groups are supported. Let us work on group 1 vs. group 2 for
Stockrate. Then group 2 vs. group 3 for Stock rate.
When we want to test multiple groups, we would need to use ANOVA test. Let us using ANOVA (and follow up pair-wise t tests if necessary). Interpret the results. You are welcome to draw the confidence intervals and compare them.
Transfer your output to Word® and format it in an appropriate manner. In your paper, address the following issues.
- the value of the Bryant article as a description of credible research on management
- The nature of the hypothesis being tested, and the appropriateness of the statistical procedures employed in that test. You may illustrate your points using the tests you conducted.
- 7 out of 51, that is about 14% of the data. Is removing 14% of the data an acceptable approach?
- How should we remove the outliers? Is t-test or ANOVA sensetive to outliers?
- Your evaluation of the discussion of the research, as reported by Bryant. Why people may think that there is a relationship between golf and firm performance. If you are going to study such relationship, how would you collect data and carry out the study?
- Your evaluation of why the New York Times published the article, and why it continues to get attention
The paper is usually between 5- 12 pages long.