Are public-school instructors' educational backgrounds (1=bachelors, 0=masters/doctorate) related to the amount of mail they receive from publishers? To answer this question, a researcher asks randomly-selected teachers from a large number of randomly-selected school districts to donate all of the mail they received from publishers over an 8-month period. The researcher counts each instructors pieces of mail and records the result into ten groups (0 = 0-8 pieces, 1= 9-16 pieces,  2 = 17-24 pieces, etc.).

Data File:  preview
Variables: ED, PIECES

PERS    ED    PIECES
1        0         5
2        0         5
3        0         4
4        0         3
5        0         3
6        0         3
7        0         4
8        0         1
(data continue)

Answer the following questions:

1. Do the data meet the assumptions of ANOVA or Kruskal-Wallis? Explain in full.
2. What kinds of data do the two variables of interest represent?
3. How many independent variables and levels do the data have?
4. Explain which statistic you ran and why.
5. Copy the F-statistic or the H-statistic (whichever is relevant) and write whether or not the .05 level of significance is attained.
6. Does the result of the comparison of means tell you whether or not there are differences between groupings? If not, what would you have to do in order to determine this information?
7. If the results are significant, what is the direction of the differences?