1 of 6 Exam 1 PS306, Spring 2005 1. Define both internal and external validity. Which type of validity is more important? Why? Using the Higgins and Marlatt study (drinking alcohol as a means of reducing tension) and the Argyle study (glasses and intelligence), tell me what Mook has to say about the importance of external validity. [10 pts] Respond using Mook article notes. 2. Dr. Juan Moore decided to extend some of Darley and Latané's results by replicating their original experiment on bystander intervention with the addition of a couple of conditions. As you might remember, the original experiment involved 3 conditions (Group Size 2, Group Size 3, and Group Size 6). Dr. Moore decides to include Group Size 4 and Group Size 5 to see if he can determine the trend of delay in helping behavior. He decides to use 6 participants in each of his 5 conditions. Unlike the original experiment, Moore uses response time in min utes as his DV. Other than the additional conditions, all else remains the same (communication study in cubicles with different numbers of "participants," talking over telephone, "victim" sounds like he's having a seizure, measure how long it takes real participant to go to help the "victim"). Some incomplete information from the replication conducted by Dr. Moore is seen below. a. Complete the source table and tell Dr. Moore what he can conclude on the basis of his experiment. b. Suppose that Dr. Moore tells you that he wants to make his experiment more powerful by conducting a repeated measures design. What are two things that you might tell him about using that procedure in his quest for power for this study. [20 pts] 4 334.533 83.633 100.360 <.0001 401.440 1.00025 20.833 .833DF Sum of Squares Mean Square F-Value P-Value Lambda PowerGroup SizeResidualANOVA Table for Time to Respond 6 1.667 .816 .3336 3.500 1.049 .4286 8.333 1.366 .5586 9.667 .516 .2116 9.667 .516 .211Count Mean Std. Dev. Std. Err.23456Means Table for Time to RespondEffect: Group Size You would reject H0 [µ2 = µ3 = µ4 = µ5 = µ6], and conclude that there is a significant effect of group size, F(4,25) = 100.36, MSE = .833, p < .001. To determine which groups differ, you would need to compute Tukey’s HSD: ! HSD = qMSErrorn= 4.16.8336= 1.55 Thus, any means that differed by 1.55 or more would lead you to conclude that the means were sampled from populations with different means. 2 3 4 5 6 2 ----- 3 1.833 ----- 4 6.663 4.83 ----- 5 8.0 6.167 1.334 ----- 6 8.0 6.167 1.334 0 -----2 of 6 We could conclude that respond significantly faster in Group Size 2 compared to all other group sizes. Furthermore, people in Group Size 3 respond significantly faster than groups of size 4, 5, or 6. Once the group reaches a certain size, there seem to be no additional increments in response times. b. This study would not need the additional power of a repeated measures design because there are many significant effects. Moreover, given the deception involved, you would not be able to conduct this study with a repeated measures design (participants would catch on). 3. Dr. Kip Werkin is an industrial/organizational psychologist who is interested in the impact of environmental factors (such as noise) on productivity. He has a group of workers experience each of a set of background noise levels (70 dB, 80 dB, 90 dB, and 100 dB) as they work on a project that involves creating delicate instruments. The dependent variable is the number of errors made in the construction of the pieces. Complete the source table and tell Dr. Werkin what he should conclude from this study. [10 pts] 9 4.000 .4443 13.900 4.633 22.339 <.0001 67.018 1.00027 5.600 .207DF Sum of Squares Mean Square F-Value P-Value Lambda PowerSubjectCategory for SPLCategory for SPL * SubjectANOVA Table for SPL 10 .200 .422 .13310 .200 .422 .13310 1.000 .667 .21110 1.600 .516 .163Count Mean Std. Dev. Std. Err.SPL 70 dBSPL 80 dBSPL 90 dBSPL 100 dBMeans Table for SPLEffect: Category for SPL First of all, you would reject H0 [µ70 = µ80 = µ90 = µ100], concluding that the noise level had an impact on number of construction errors made, F(3,27) = 22.339, MSE = .207, p < .001. To determine which specific groups differed, you would compute Tukey’s HSD: ! HSD = qMSErrorn= 3.86.20710= .555 Thus, the 100 dB group produced significantly more errors than all other groups. The 90 dB group made more errors than the 70 dB and the 80 dB groups. 3a. If the same data were analyzed with an independent groups design, what would the source table look like? Under which conditions would a repeated measures analysis of a data set not lead to a larger F-ratio than an independent groups analysis? If the SSSubj is relatively small, then the repeated measures ANOVA will not yield a larger F. [5 pts] Source df SS MS F Treatment 3 13.9 4.633 17.37 Error 36 9.6 .267 Total 39 23.53 of 6 4. Dr. Bing Bada is convinced that sleep deprivation influences aggression. To test this assumption, volunteer participants are randomly assigned to sleep-deprivation periods of 0, 24, 28, or 72 hours and subsequently tested for aggressive behavior in a controlled social situation. The dependent variable (aggression) is operationally defined as the number of times the participant hits or curses at a confederate who attempts to annoy the participant. Complete the StatView output below and then tell Dr. Bada what to do next. [10 pts] 3 .275 .092 .264 .8509 .792 .09536 12.500 .347DF Sum of Squares Mean Square F-Value P-Value Lambda PowerSleep DeprivResidualANOVA Table for Agg Act 10 .200 .422 .13310 .300 .483 .15310 .400 .699 .22110 .400 .699 .221Count Mean Std. Dev. Std. Err.0242872Means Table for Agg ActEffect: Sleep Depriv First, of course, you would retain H0 because p > .05. Next, you would try to address the reason(s) that your study did not result in a significant effect. In other words, you will look for ways to increase power. Dr. Bada used a no-deprivation control group, as well as a fairly wide range of periods of sleep deprivation. As a means of increasing the treatment effect, of course, you could always increase the range of sleep deprivation (96 hours, etc.). You might also look for a more sensitive DV than number of times that the participant hits or curses at someone else (e.g., include other forms of verbal abuse, or other forms of aggression). You should always consider increasing the sample size (only n = 10 here). Finally, you could reduce