Multiple Comparisons

In-class Practice

An agricultural researcher wants to investigate whether different fertilizer treatments affect corn yield (measured in bushels per acre). Four treatments are considered:

  1. Control: No fertilizer.
  2. Nitrogen: Nitrogen-only fertilizer.
  3. Organic: Organic fertilizer.
  4. Combination: A combination of nitrogen and organic fertilizer.

The researcher randomly assigns 10 plots of land to each treatment and measures the resulting corn yield. Data for the experiment is on our course webpage as fertilizer-experiment.csv.

The researcher is interested in answering the following questions:

  1. Does applying fertilizer produce a different average corn yield than applying no fertilizer?
  2. Does nitrogen fertilizer produce a different average yield than organic fertilizer when applied individually?
  3. Does combining nitrogen and organic fertilizer produce a different average yield than using either fertilizer individually?

For each research question:

Using the R functions that we discussed, perform the following operations:

  1. Fit a one-way ANOVA model to investigate differences in average corn yield among the four fertilizer treatments.
  2. Obtain the estimated marginal means using the emmeans package.
  3. Define and implement your three planned contrasts.
  4. Conduct the hypothesis tests at (= 0.05).
  5. Interpret the results using the context of the research question.