Final Exam | Psychology 543 | Fraley http://www.uic.edu/classes/psych/psych543/

The exam will be due on Friday Dec 13th by noon. Please make sure your answers are typed, well written, coherent, and thoughtful. Some of the questions clearly have "right" answers, but others will involve a bit of subjectivity. Please provide thoughtful answers to such questions in order to maximize the number of points you receive. The final exam will constitute 40% of your grade for the course.

1. On the first day of class, you took a pop quiz regarding some basic statistical concepts (e.g., standard deviations, covariances). Take a look at that quiz--and your answers--and, for the questions 1-6, explain why your answers were wrong if they were wrong. Provide a new answer in light of what you've learned this semester. Do not worry about questions you originally answered correctly. (I'm assuming you still have a copy of your quiz in your files. Just in case you don't, I have placed a copy of your original answers to the quiz in your mailbox.)

2. Please explain what a p-value is. In addition, please return to your original quiz and look at question number 7. Each of those interpretations is false. Explain why each of those statements is false.

3. Reeder, Kumar, Hesson-McInnis, and Trafimow (2002) recently published an article in the Journal of Personality and Social Psychology which they examined the perceived/inferred morality of aggressors and how perceptions of morality may vary as a function of the perceived motives of the aggressors. In Study 1 the researchers instructed subjects to read scenarios about aggressive acts and continuously rate the morality of the aggressors. The researchers manipulated two variables: (a) whether aggression was encouraged (rewarded) or discouraged (punished) in the scenario and (b) whether the scenario involved aggression that was reactive (i.e., in reaction to some kind of provocation) or instrumental (i.e., in the service of a tangible outcome).

Here are the mean levels of perceived morality obtain in each of the four cells of the experiment (standard deviations are reported in parentheses):

(a) Create a single regression equation that will allow you to model the means obtained in each cell of this table. (Hint: This will require some dummy coding.) In your answer, be sure to (1) write the full regression equation (i.e., place the actual parameter values in the equation), (2) show me how the equation can be used to model each mean in the table, (3) explain whether there are main effects of each factor and an interaction (ignore significance testing issues for this part), and (4) create a barplot or bargraph that depicts the means for this study.

(b) The authors employed a sample size of 80 (20 people in each cell of the design). Use your knowledge of S-Plus to estimate the authors' statistical power to detect (1) a main effect of reactive vs. instrumental, (2) a main effect of encouragement vs. discouragement, and (3) an interaction between these two variables. For the purposes of these computations, assume that the population means are identical to those observed in the sample. (Note: For simplicity, please pool the various standard deviations. In other words, assume that, while the population means may vary from one cell to the other, there is only one sigma.)

(c) Although the authors did not hypothesize a main effect of the encouraged vs. discouraged variable, how many people would they need in each cell in order to detect this main effect with statistical power of 80%?

(d) Using "backwards inference", calculate the 95% confidence interval associated with each mean.

4. Sometimes our theoretical models do not explain the data as well as we might hope. Think of a simple theoretical model in your area of research and discuss, with this model in mind, three reasons why that model may be incapable of explaining the data perfectly. To illustrate your points, create (i.e., make up) a simple dataset (or datasets) that is pertinent to your theoretical example. A good answer will not be a simple imitation of the examples discussed in class. Please provide a sufficient degree of detail--including real numbers, real analyses, and simulations.

5. R. A. Fisher and Pearson and Neyman advocated dramtically different perspectives on the role of p-values in science. Explain at least two ways in which these two groups of individuals differed in their opinions. Also, with respect to each issue, take a position and explain/defend your own thoughts on the matter.