In this section, the author compares groups of people on different doses of puppy therapy to determine whether it significantly impacts their happiness, controlling for the effects of puppy love prior to treatment. The purpose of including this covariate is that people who have an increased affinity or love for puppies might be more affected by the therapy, therefore influencing the results of the experimental study in ways that are separate from the experimental treatment. When you are ready, open the dataset up in IBM SPSS on your computer.
Follow the author’s instructions to conduct the ANCOVA using the example he provides in 13.5. Specifically, compare happiness for people receiving different experimental doses of puppy therapy using the variable dose (in this instance, happiness is your dependent variable and Dose is your categorical predictor) and controlling for Puppy_love. In other words, test whether people on a low-dose therapy, high-dose therapy, or placebo significantly differ in their happiness as a result of their treatment condition, when controlling for prior puppy love.
- Run all analyses in Section 13.5.
- When you have your output, verify that the numbers you obtained are the same as those in the textbook.
- Export your output to a word file. In your output file, highlight the relevant statistics for interpreting this analysis and annotate your output as you have been doing, and for each “NOTE”, write a few sentences explaining what the output shows and interpret the ANCOVA.
- If you need additional help annotating your output, read Section 13.6 of your textbook where the author explains the output of the analysis in detail.
Part 2: Testing the Assumption of Homogeneity of Regression Slopes
For the second part of the assignment, complete the instructions inSection 13.7, p. 439-441,to test for the assumption of homogeneity of regression slopes. In this analysis, you will include an interaction between Puppy_love and dose to test whether the regression slopes by condition are homogeneous. If not, then an assumption of ANCOVA has been violated.
- Conduct all analyses described in Section 13.7.
- Then, export the output for this analysis and highlight the relevant statistics.
- Also,annotate your outputas you have been doing, and interpret the interaction coefficient. Explain what this means.
- Turn in this output along with the output from Part 1 and the report from Part 3.
Part 3: Reporting ANCOVAs in APA Style
For the last part of this assignment, you will report on the analyses you conducted in Parts 1 and 2. To assist with this assignment, I have provided an example of an APA style table for an ANCOVA in the documentExample APA Style ANOVA Table. You may reference this as an example when completing this assignment. You should make a table for your main analysis in Part 1 (the ANCOVA) and a table for the analysis you ran in Part 2 (including the interaction term). There is also an example of reporting the results of an ANCOVA in Section13.11of your textbook.
- For this assignment, you are not required to write an introduction or method section.
- Begin with a paragraph summarizing in words the ANCOVA you conducted (describe the variables with their proper names, not the shortened names saved in the dataset). Be sure to state what the null hypothesis was and what hypotheses you were testing. Also, describe the analysis you conducted to test for the homogeneity of regression slopes.
- Next, your report should include an APA style results section reporting the ANCOVA, followed by a discussion section and the two APA style tables. In the discussion section, you should summarize the results of the analysis in words (how would you explain what you tested and what the results were to a non-statistics person?).
Next, discuss the following:
- What is an ANCOVA analysis? How is it different from an ANOVA? Why would a researcher conduct an ANCOVA?
- Explain how an ANCOVA could be conducted as a regression analysis. What do the b values mean for an ANCOVA in the linear equation?
- What is the assumption of independence of the covariate and treatment effect? What can be done to avoid or lessen the likelihood of this happening?
- Explain the assumption of homogeneity of regression slopes. If this is violated (as it was in your analysis), what does this mean? What would you recommend in this example given the results of that analysis?