Page 1 of 5 ID# Exam 2 PS 306, Spring 2005 As always, the Skidmore Honor Code is in effect. Read each question carefully and answer it completely. Keep in mind that a point should roughly equal a minute, so a 10-point question should take you about 10 minutes to complete. If you spend longer than that on the question, you may not finish the exam. Good Luck! 1. As we’ve discussed there are several different types of two-factor designs. In general, it makes sense to include as many repeated factors as possible. First, tell me why one would want to run a repeated measures design whenever possible. Second, tell me the kinds of independent variables that do not lend themselves to use as repeated factors. [5 pts]Page 2 of 5 2. For the following designs, tell me the number of participants needed and the total number of pieces of data that one would have in the study. [12 pts] Design Total Participants Total Pieces of Data A 5x7 completely independent groups (between) design, so that you achieve a minimum of 30 scores per cell. A 5x7 completely repeated measures (within) design, so that you achieve a minimum of 30 scores per cell. A 5x7 mixed design, with the first factor repeated and the second factor independent groups, so that you achieve a minimum of 30 scores per cell. A 5x7 mixed design, with the first factor independent groups and the second factor repeated measures, so that you achieve a minimum of 30 scores per cell. A 5x8 mixed design, with the first factor independent groups and the second factor repeated measures, so that you achieve a minimum of 30 scores per cell. A 4x6 mixed design, with the first factor repeated measures and the second factor independent groups, so that you achieve a minimum of 30 scores per cell. 3. In our experiments, we’re always interested in power. First, define power. Second, tell me what term is the complement of power. Third, tell me about the relationship between effect size and power (i.e., with low effect size...). [3 pts]Page 3 of 5 4. People often compare themselves to other people, which has led social psychologists to study the phenomenon. In the proxy model, one estimates one’s own likelihood of success by comparing oneself to a proxy (proxy = substitute). For example, suppose that you were trying to decide whether or not to go to graduate school. If you knew someone (a proxy) who was very much like you (in perceived intelligence, motivation, grades, etc.) and that person had failed in her attempt at graduate school, that knowledge might lead you to question your own likelihood of success in graduate school. The type of tasks studied varies quite a bit, but one area studied by Martin, Suls, and Wheeler (2002) involves grip strength (though they study other kinds of tasks as well). Participants begin by squeezing an exercise hand grip as many times as they could in a 30-sec period (Task1). Their performance was recorded and related to the participants. Then participants were shown a hand dynamometer, which measures hand grip force in kg/force exerted. They were asked to predict how much pressure they could exert on the dynamometer (Task 2). To aid them in their predictions, participants were shown (fictitious) results from a participant “from last semester.” Participants were randomly assigned to one of three proxy conditions: Inferior (the proxy hadn’t squeezed as many times on Task 1 as the participant), Similar (the proxy had squeezed about as often on Task 1 as the participant), or Superior (the proxy had squeezed more often on Task 1 than the participant). In addition, based on written notes, the effort invoked by the proxy is either Ambiguous (participants were told that the experimenter did not know whether the proxy had exerted maximal effort or not) or Maximal (participants were told that the proxy had exerted maximal effort). The study was conducted as a 3x2 independent groups design. The dependent variable is a prediction difference score (PDS). Thus, if the participant expected to perform better than the proxy, the PDS would be positive. Solely from the figure below, predict what the authors found in their analyses of these data (treat any difference as significant) and interpret the results as best you can (as you would in a Discussion section). [15 pts] Main Effect for Effort: Yes/No Why? Main Effect for Proxy: Yes/No Why? Interaction: Yes/No Why? InterpretationPage 4 of 5 5. Dr. Mo Shun was interested in the impact of various dosages of a new drug (Stay Put) on the activity level of hyperactive children. She is fairly sure that, because of its chemical nature, Stay Put will be more effective for males than for females. To that end, she administers four dosage levels (None, Low, Medium, High) of Stay Put to an equal number of male and female children who exhibit similar levels of hyperactivity. The dependent variable is an activity measure, with higher numbers indicating greater activity. Analyze and interpret these data as completely as you can. [20 pts] {Johnson} 10 2.400 1.265 .40010 6.300 1.494 .47310 2.800 .632 .20010 7.100 2.846 .90010 3.200 1.874 .59310 7.200 1.814 .57310 7.200 .789 .24910 8.300 2.359 .746Count Mean Std. Dev. Std. Err.Female, HighFemale, LowFemale, MedFemale, NoneMale, HighMale, LowMale, MedMale, NoneMeans Table for ActivityEffect: Gender * DrugPage 5 of 5 6. In Lab 2, we used difference scores as the DV so that the analysis would be a single-factor ANOVA. However, now that you’re familiar with two-way ANOVA, we can think of the results from that lab in terms of actual response times in minutes (N.B. not milliseconds) as the only DV, so we’ll ignore difficulty ratings. Let’s suppose that we’d run that study as a completely between (independent groups) experiment. The two independent variables would be the same: TASK (actually solve anagrams vs. with solution present, estimate time for someone else to solve) and LENGTH (anagrams would be 4, 6, or 8 letters long). Note, however, that as a completely between design, a person would be exposed to only one of the six conditions. Complete the analysis below, interpret the results as completely as you can, and describe why these results may have arisen (as you would in a Discussion section). [25 pts] 10 .650 .108 .03410 1.135 .258 .08210 1.350 .127 .04010 1.960 .255 .08110 2.210 .256 .08110 2.720 .148 .047Count Mean Std. Dev. Std. Err.4, Estimate4, Solve6, Estimate6, Solve8,