Psych 225 Lecture 9Outline of Last Lecture 1) Impact of Intervention on Depression Study2) Decrease in error variance3) Stroop differences between Study 1 and 2Outline of Current Lecture 1) Applying Power Principles 2) Between Subjects DesignCurrent LectureExam -look at IE Practice Exams Applying Power Principles (Course Packet p.34)Slide 2: Which researcher is a large N more important? Why?Answer: Study 1 with 4 levels of IV because you have a lower degree of freedom between so you want a larger degree of freedom betweenSlide 3: Which researcher should use a stronger manipulation of IV? Why?Answer: Study 1. There is less degrees of freedom withinWhich study would be in tested in sampling distribution with more variability? Why? Study 1 less degrees of freedom? These notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.Slide 4: Which researcher use a within subjects effect?For a subtle effect or for a large effect?Researcher 1 (with subtle effect) you are lessening variation of error ~within subject designs are better able to detect small differences between means Slide 5: Which experiment has:Larger SSB: Expt. 2 where manipulation is stronger; level of IV cover a broader range Larger Fcalc: Expt. 1 where manipulation is weaker where means wont vary as much; magnitudeof MSB small makes F rather small and P rather largeLarger p: Slide 6: Which studyLarger F calc: Study 2 (minimize error or MSW-bottom of F ratio)Smaller SD’s: Study 2Larger MSW: Study 1Larger p: Study 1 (anticipate more error, shrinking value of F, which is accompanied by larger p)Slide 7: Which researcher should use fewer experimenters?-Study 1 (already concerned about error)Slide 8: Which study?F calc larger? Study 2T calc larger? Study 2Mean diff with smaller p value? Study 2Smaller diffs between means be significant? Study 2HSD larger? Study 1Slide 9: In which study do we have a better chance of demonstrating that IV affected DV?Answer: (same dfs 60-6 and 59-5=54); thus Study 1 Slide 10: Study 2Slide 11: Study 2Slide 12: How would p’s compare?Answer: SEE PG 19 and 37 for help understanding thisResearcher with hypotheses will have p values that are smaller (a priori more powerful) Between Subjects: In Class Example (CP. 41) Monkey example (24 monkeys)-referring to partitions on pg. 11 -reward size is a quantifiable type of variable: thus order the levels from small to large -if you can make order of levels meaningful, than do it!-do not have overlap between names of variable and names of levels -single factor design partition SST into 2 pieces: SSB and SSW same with degrees of freedom total partitioned to: dfB and dfW-2 explanations for why scores differ-SSB: variability of the cell mean by focusing on each IV at a time: SSF1 (reward) SSF2 (drive) SSF1xSSF2 (Reward x Drive)-dfT= 23 (24-1)-dfW= 6 cells x3 df in each=18 OR 24-6-dfB= k-1 = 6-1=5 (k is number of conditions/combinations/cells) -dfB: dfF1 (reward) , dfF2 (drive), dfF1xF2 (reward x drive) dfF1 (3-1=2) (column-1) dfF2 (2-1=1) (row-1), how many df remain? 5-2-1=2 dfF1xF2 (how manycolumns-1 x how many rows-1) Packet pg. 41 continued-how many sampling distributions are used for these above 3 f tests?df (2, 18) Rewarddf (1, 18) Drivedf (2, 18) Reward x Drive *Thus we will use 2 sampling distributions-more degrees of freedom is less variability Where would you look to see if there is a main effect for reward size?(did it affect discrimination learning)~compare the column means-did reward size have a different effect for monkeys in different condition, low and high drive (hungry vs. satiated)Looks like reward size impacted discrimination only if the monkeys were satiated; had positive impact among satiated animals, but not on hungry animals (numbers represent number of correct response discrimination monkey made; higher score the better the monkey did) Course Packet pg. 38 and 39FIGURE 1-no drive main effect: only one line (lines are on top of each other) performance of hungry monkeys were the same as satiated monkey-SSB: variability of cell means; answer: 20, absolutely no variable in cell means so there is nothing to partition Figure 2-reward size on discrimination learning -as reward size increased (from small to moderate to large) discrimination learning also increased Figure 3-reward size on discrimination learning: no effect -however, there is a main effect for drive magnitude on discrimination learning-those who were full did better on discrimination learning -same magnitude of improvement (from hungry to satiated) across reward sizeWhy aren’t all cell means the same? ~some monkeys were full others were hungryFigure 4-effect for reward size on discrimination learning -no interaction; two slopes are of the same magnitude Figure 5-no main effect on reward size (look at average of two lines) -no main effect on drive (look at average of two lines-opposite way) -BUT there was an interaction Figure 6-yes main effect on reward size (average two lines) qualify that reward has an effect when monkeys are hungry -no main effect on drive -yes interaction Figure 7-no main effect on reward size-yes main effect on drive -yes interactionFigure 8-different than figure 4 because it is not parallel, slopes are different-main effect for reward-main effect for
View Full Document