In an experiment done in Ontario, Canada by entomologist Geoffery Beall, 72 independent blocks (areas to be sprayed) were created in tobacco fields and then 6 types of insect sprays were randomly assigned to the different blocks. The six different insect sprays were tested to see if there was a difference in the number of tobacco hornworms, Manduca quinquemaculata, found in the field after each spraying. This data set is already in R, just type in data(InsectSprays) and then attach(InsectSprays). Regardless of whether or not the Assumptions for the test are met, all parts below!!!

(a) Make a side-by-side boxplot showing the distribution for the number of insects for the different insect sprays. Comment on the results.

(b) Create residual plots for this data to check the normality and variability assumptions. Comment on what you see in the Residual vs. Fitted Plot and the Normal Q-Q Plot.

(c) Perform a Levene Test for Homogeneity of Variance. What can you conclude about the variation?

(d) State the null and alternative hypotheses for the one-way ANOVA on this data set.

(e) Perform the ANOVA on these data and show me your output table. What do you conclude?

(f) If the results of the ANOVA indicate that the means are significantly different, perform a multiple comparisons test (Tukey HSD confidence intervals) to determine which insect sprays differ in terms of the mean insect count. Analyze the differences and Interpret the first two significant intervals. Which of the insect spray(s) would you recommend if you want to kill the most tobacco worms?

count spray

110A

27A

320A

414A

514A

612A

710A

823A

917A

1020A

1114A

1213A

1311B

1417B

1521B

1611B

1716B

1814B

1917B

2017B

2119B

2221B

237B

2413B

250C

261C

277C

282C

293C

301C

312C

321C

333C

340C

351C

364C

373D

385D

3912D

406D

414D

423D

435D

445D

455D

465D

472D

484D

493E

505E

513E

525E

533E

546E

551E

561E

573E

582E

596E

604E

6111F

629F

6315F

6422F

6515F

6616F

6713F

6810F

6926F

7026F

7124F

7213F