ANOVA Example: Econ 222
ANOVA Example: Econ 222
ANOVA Example: Econ 222
Three training methods were compared to see whether they led to greater productivity after training. The following are productivity measures for individuals trained by each method. These productivity data are the dependent variable for this study. The training method, which is a categorical variable with three categories, is the explanatory variable. The categorical variable, Method, is also referred to as the explanatory 'factor' and the categories are also called 'factor levels' or 'treatments.' The term 'treatment' is sometimes used because the experimental units, here the individuals being trained, are exposed to three different 'treatments' before the productivity measure is taken. If this data were collected and presented in an electronic data set, it would be viewed as a data set with 18 observations of two variables. The data set might look like this: Obs Prod Method 1 45 A 2 40 A 3 50 A 4 39 A 5 53 A 6 44 A 7 59 B 8 43 B 9 47 B 10 51 B 11 39 B 12 49 B 13 41 C 14 37 C 15 43 C 16 40 C 17 52 C 18 37 C Note that the letters used for the Method variable (A, B, and C) are identifiers. Numbers could have been used instead or any other distinct identifiers such as Meth1, Meth2, and Meth3. It is important to understand that if numbers are used as identifiers (actually category labels), they cannot be used as if they were highlevel numeric values in any statistical analysis of the data set. If the data were presented as above, in order to use Excel for the ANOVA, the data would have to be rearranged into a column or row table in the same manner that the problems are presented in our text.
Analysis of Variance Do these data provide sufficient evidence to conclude that three training methods lead to different mean productivity?
45 59 41
40 43 37
50 47 43
39 51 40
53 39 52
44 49 37
Anova: Single Factor SUMMARY Groups Method A Method B Method C All Data Combined
Count 6 6 6 18
Sum Average Variance 271 45.17 30.1667 288 48.00 47.6000 250 41.67 31.0667 809 44.94 39.1144
ANOVA Source of Variation SS Between Groups 120.7778 Within Groups 544.1667 Total 664.9444
df
45.17
50
2 Training Method