ID# Exam 1 PS 306, Spring 2002You should recognize the above cartoon as a fanciful way of reminding you that the SkidmoreHonor Code is in effect. Work your way through the exam quickly and carefully, answering eachquestion as completely as you can. Show as much of your work as possible, so that you can besure of getting as much credit as possible. Don’t hesitate to comment on particular designs, etc.,even if it’s not called for explicitly in the question (some questions are implicit). I think of apoint as a minute, so you should expect to spend about 10 minutes on a 10-point question (forexample). If you spend more time on a question than it is worth, you may not be able to completethe exam. There is a class following ours, so I will collect all the exams promptly at the end ofthe allotted time. Good Luck! Have a wonderful Spring Break!1. Briefly provide an example of a nonmanipulated characteristic of a participant. Then, providea clear explanation why you could not make a causal claim about such a variable if it were usedin a study. (Be very explicit!) [5 pts]A nonmanipulated characteristic of a participant is some aspect of the person that he orshe brings to the study (age, personality, IQ, race, etc.). So, imagine a study with twogroups (Low IQ and High IQ). Suppose that you use a DV of reaction time and you find asignificant difference between the two groups. Would you be comfortable concluding thatIQ caused the differences in RT? I would nope not! People with High IQ may be moreconfident, which makes them more prone to respond rapidly (so it’s actually confidencethat is producing the observed difference). Or people with Low IQ may be more cautiousgenerally, which leads them to respond more slowly.2. Hypothesis testing is essential to the research enterprise in psychology. Briefly define Type IError, Type II Error, and power. Then, tell me why power is so important (or alternatively, why aType II Error is so bad). Finally, tell me as many ways as you can to increase the power of astudy. [5 pts]Type I Error (a) = probability of incorrectly rejecting H0Type II Error (b) = probability of incorrectly retaining H0Power (1 – b) = probability of correctly rejecting H0Power is important (or one wants to avoid Type II Errors) because it’s wasteful of one’stime to look for differences (effects of a treatment) with little likelihood of actually findingsuch differences. You should do everything you can to ensure that your study has sufficientpower to detect the sort of treatment effects that are typically investigated by psychologists.Such strategies include:“large” sample size“large” treatment effects“small” error terms (by reducing individual differences or random error)using a repeated measures design3. Briefly explain why it is essential to develop an error term (MS) for the repeated measuresdesign that includes only variability due to random factors? How is it computed? [ 5 pts]The MSTreatment in a repeated measures design has no individual differences in it (becausethe same people contribute to each level of the factor). Thus, you need an error term thathas no individual differences in it. As a result, your F-ratio will tend toward 1.0 when notreatment effect is present. You can think of computing MSError for a repeated measuresdesign by removing the individual differences from the MSError you would have in anindependent groups design. The error term in a repeated measures design indexes theinteraction between participants and treatment. Thus, if each individual responds similarlyto the levels of treatment, your error term will be small.4. External validity is important in some circumstances, but Mook tries to argue that it’s notparticularly important in a lot of psychological research. First of all, define external validity.Then, using at least two of the studies cited in the article, explain circumstances in which Mookargues that external validity is not important (and why he thinks that way). Finally, using a singlestudy that Mook discusses, indicate why a manipulation check would have been appropriate (orwhy a manipulation check would not have been needed). [10 pts]Clarity and detail would lead to a good response to the first part of this question.A manipulation check would be useful when the IV is not directly observable. (Also, youprobably need human participants to conduct a manipulation check, so the Harlow studywouldn’t work.) Thus, you might consider the Higgins & Marlatt study, where participantswere made more or less anxious by a manipulation of threat of shock. But were the peopleactually more or less anxious as a result of the manipulation? No real way to know unlessyou do a manipulation check. At the end of the study, you might ask each participant aseries of questions, in which you embed a question asking the participant to rate howanxious he or she had been during the earlier phase of the experiment. You would hopethat the high anxious group would give you a higher anxiety rating than the low anxiousgroup.5. Dr. Buster Gutt believes that laughter is a good antidote to depression. To test his hypothesis,he randomly samples 20 people and asks them to wear a counter device that they press everytime they laugh during a given randomly selected day. At the end of the day of the study, hegives participants a device that tests their level of depression. Rather than use the BeckDepression Inventory, he chooses to use the less well known Degree of UnmanageableDepression Evaluation. Scores on the DUDE run from 0 (no depression) to 20 (very depressed).Dr. Gutt has analyzed his data as seen below.191.819.670.6514.234CountNum. MissingRR SquaredAdjusted R SquaredRMS ResidualRegression SummaryDepression Score vs. Times Laughing1 619.666 619.666 34.567 <.000117 304.755 17.92718 924.421DF Sum of Squares Mean Square F-Value P-ValueRegressionResidualTotalANOVA TableDepression Score vs. Times Laughing15.412 1.542 15.412 9.993 <.0001-.797 .135 -.819 -5.879 <.0001Coefficient Std. Error Std. Coeff. t-Value P-ValueInterceptTimes LaughingRegression CoefficientsDepression Score vs. Times Laughing02468101214161820Depression Score-5 0 5 10 15 20 25 30Times LaughingY = 15.412 - .797 * X; R^2 = .67Regression PlotBased on these analyses, interpret his results as completely as you can. (Be explicit!) If a personlaughed 10 times a day, what would you predict that person’s DUDE score to be? If a personlaughed 20 times a day, what would you predict