DOC PREVIEW
SKIDMORE PS 217 - PS 217 Final Exam Answers

This preview shows page 1-2 out of 7 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

1 Final Exam PS 217, Spring 2004 1. What is the relationship between power and effect size? That is, when you are considering a research design in which there is a large effect size, what are the implications for power? When you are considering a research design in which there is a small effect size, what are the implications for power? How would you typically add power to a study? [10 pts] With large effect size, you would not need a lot of power to detect the effect. With small effect size, you would need a lot of power to detect the effect. To increase power, you would typically increase sample size (n). In an experiment, you could also consider ways in which you could increase the treatment effect. You could also consider ways in which you could decrease error due to individual differences (possibly using a more homogeneous sample) or random variability (possibly by improving instructions, eliminating disturbances to participants, etc.). 2. How seriously do people take product reviews? Chaiken and Maheswaran (1992) conducted an interesting experiment in which they varied the credibility of the review source and the general message of the review. They asked college students to read a review of a new telephone answering machine. The researchers told half the participants that the review came from a flyer printed by the discount store Kmart (low credibility) or from the magazine Consumer Reports (high credibility). Each participant then read one of three types of review, an unambiguous strong review, an ambiguous review (the answering machine was better than some machines but not as good as others), or an unambiguous weak review. The researchers then asked the participants to rate on a 10-point scale their willingness to buy the answering machine for $50 (10 = very willing to buy). The results of their study are replicated below. Complete the analysis and interpret the results of this study as completely as you can (as in a Discussion section). [10 pts] {Pittenger} 1 .067 .067 .056 .8143 .056 .0562 80.433 40.217 33.618 <.0001 67.235 1.0002 58.233 29.117 24.339 <.0001 48.678 1.00054 64.600 1.196DF Sum of Squares Mean Square F-Value P-Value Lambda PowerSource CredibilityEndorsementSource Credibility * EndorsementResidualANOVA Table for Rating 10 8.000 1.155 .36510 5.900 1.101 .34810 3.000 1.155 .36510 5.400 .699 .22110 6.600 1.174 .37110 5.100 1.197 .379Count Mean Std. Dev. Std. Err.High, AmbiguousHigh, StrongHigh, WeakLow, AmbiguousLow, StrongLow, WeakMeans Table for RatingEffect: Source Credibility * Endorsement 23456789Cell MeanAmbiguous Strong WeakCellLowHighInteraction Line Plot for RatingEffect: Endorsement * Source Credibility First, I would compute Hartley’s FMax, to determine if I should be concerned about heterogeneity of variance: ! FMax=1.432.489= 2.93, so, with FMax/Crit = 7.8, I would conclude that there is no reason to be concerned about heterogeneity of variance and I would use α = .05.2 There is a significant main effect for Endorsement and a significant interaction between Source Credibility and Endorsement. To explain the interaction, I would look at the graph and see where the simple effects appear to differ, then I would compute HSD and determine if I could interpret the interaction. ! HSD = qMSErrorn= 4.181.19610= 1.45 With a Weak argument, people are more persuaded (more willing to buy) when the argument is made by a low-credibility source than when made by a high-credibility source. However, when the argument is Ambiguous, people are more willing to buy when the argument comes from a high-credibility source than when it comes from a low-credibility source. When the argument is Strong, however, participants are persuaded equally by the weak and strong sources. If you were interested in assessing the effect size, you would focus on the interaction: ! "2=58.23358.233 + 64.6= .47 3. Dr. Sally Forth is interested in studying the relationship between Locus of Control (a measure developed by Dr. Julian Rotter) and the number of different people that a person has dated. She hypothesized that there would be a positive linear relationship between locus of control and the variety of a person’s dating partners (higher locus of control leading to greater number of different people dated). Dr. Forth collected data from 50 college students on her scale of Locus of Control (0 = Low and 10 = High). Interpret her results (seen below) as completely as you can. If a person had a Locus of Control score of 7, what would be your best estimate of the number of different people that person would have dated? If a person had a Locus of Control score of 12, what would be your best estimate of the number of different people that person would have dated? Be very explicit in telling me why you would not be willing to accept the conclusion that one’s Locus of Control affected the number of different people one would have dated. [10 pts] 500.669.448.4363.957CountNum. MissingRR SquaredAdjusted R SquaredRMS ResidualRegression SummaryNumber of Dates vs. Locus of Control1 609.318 609.318 38.918 <.000148 751.502 15.65649 1360.820DF Sum of Squares Mean Square F-Value P-ValueRegressionResidualTotalANOVA TableNumber of Dates vs. Locus of Control.285 1.205 .285 .237 .81381.228 .197 .669 6.238 <.0001Coefficient Std. Error Std. Coeff. t-Value P-ValueInterceptLocus of ControlRegression CoefficientsNumber of Dates vs. Locus of Control05101520253035Number of Dates0 1 2 3 4 5 6 7 8 9 10 11Locus of ControlY = .285 + 1.228 * X; R^2 = .448Regression Plot3 There is a significant positive linear relationship between Locus of Control (Loc) score and Number of Dates, r(48) = .669, p < .001. If a person had a LoC score of 7, I would predict that the person would have had 8.88 (~9) dates. If a person had a LoC score of 12, I would have to say that I couldn’t safely predict the number of dates, because I didn’t observe anyone with that high a LoC score. If the trend continued, however, I would predict 15 dates. I wouldn’t be willing to conceive of the relationship as a causal one because it could well be that as a person has more dates, her or his LoC scores changes (causal arrow problem). It could also be that a “third” variable may be creating the relationship. That is, maybe self-esteem affects both a person’s LoC score and the number of dates that person has had. 4. Dr. Nomar Gassé was interested in the impact of varying levels of depression on a person’s


View Full Document

SKIDMORE PS 217 - PS 217 Final Exam Answers

Download PS 217 Final Exam Answers
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view PS 217 Final Exam Answers and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view PS 217 Final Exam Answers 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?