Statistics for Business: Decision Making and Analysis, 3rd Edition

1. 1. ... shown for the slope in the table. The test would not find a statistically significant difference.

   2. What is missing from this comparison of the two groups?³

      ³ Where are the plots? We need to see, at a minimum, boxplots comparing the data in these groups in order to check the conditions for the two-sample comparison.

Comparing Several Groups Using Regression

The connection between the pooled two-sample t-test and simple regression hints that we can use multiple regression to compare the means of several groups. Simply add more dummy variables.

For the wheat trials, define dummy variables for the other varieties:

$$\begin{array}{lll}\hfill D\left(\text{Hatcher}\right)& =\hfill & 1 \text{if} \text{plot} \text{grows} \text{variety} \text{Hatcher} \text{and} 0 \text{otherwise}.\hfill \\ \hfill D\left(\text{NuHills}\right)& =\hfill & 1 \text{if} \text{plot} \text{grows} \text{variety} \text{NuHills} \text{and} 0 \text{otherwise}.\hfill \\ \hfill \text{D}\left(\text{Ripper}\right)& =\hfill & 1 \text{if}\hfill \end{array}$$