Psych 225 1st Edition Lecture 6 Outline of Last Lecture 1) Paper updates2) Stroop stimuli Outline of Current Lecture 1) Stroop Results2) F distribution3) ANOVA4) T-test and descriptivesCurrent LectureExam Review~Journal Articles: preview questions in CP!Slide: Stroop Results-how many sum of squares total?: -in example: there are 15 scores ~how many square deviations: sum of squares: within scores of women who received consistent: 5; inconsistent: 5 and control:5-from each raw score, subtracting from overall mean-SSW looking at variability Slide: Stroop Results continued-under ANOVA: look at the Sum of Squares for Between Groups and Within Groups-F ratio: F=MSB/MSW (mean square between/mean square within) -if type of presentation does not effect reactant time, the F should be about equal to 1-how much variance of F constitutes and effect depends on degrees of freedom (numerator anddenominator of the F statistic) 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.EXAM: be able to~estimate P value~explain what a p value means Slide: F Distribution (CP pg. 13)-body of table has critical values -more degrees of freedom=smaller critical values and less sampling distribution -fewer degrees of freedom=less power; tougher to meet/exceed the critical value-2 and 12 degrees of freedomF (.05) =3.89 & F (.01)= 6.93~we calculated F=5.34; estimate of value of p: .03 (between .05 and .01) Slide: Stroop Study Continued-F and T are both showing the variability -for example: when comparing control versus consistent conditions, assign a 0 to the inconsistent data so that it is dropped out of the analysis; assign a 1 to the consistent condition and a -1 to the control condition to tell SPSS to compare those groups t= u(i)- u(j) divided by the square root of the Means Square within time ((1/u(i)) + (1/u(j))~example: 14.8-17.6 divided by the square root of 28.267 times the sum of (1/5 + 1/5) =.83Slide: Page 15 of CP -using a contrast t that pools degrees of freedom from all of our samples~thus look at 12 degrees of freedom (which is the same as Mean Square Within degrees of freedom) -power in contrast t requires us to use all degrees of freedom even though we are only comparing 2 of the 3 conditions Repeat process with inconsistent and control group (by assigning 0 the consistent group, -1 to the inconsistent and 1 to the control: SPSS will subtract the means/compare)*still 12 degrees of freedom because we use all of them in contrast t testsfor this condition: 2.32 is the t value: calculated t would be in a significance value of p=.03 (ish)-default do a two tailed test Slide: ANOVA (SEE CP p.23)-type of presentation affected reaction time, F (2, 12)=5.34, p=.02Slide: t-test and descriptives (SEE CP p.23)-always write in the direction revealed even if the results are disappointing -by reminding reader of our expectations, we prep them for the discussion-focus on the words over the numbers (first hypothesis)~the more error we have the tougher it is to demonstrate effects: need to try toproduce more homogeneity among the levels of IV ~larger sample size adds more variability and thus doesn’t help with the error problem TO REDUCE ERROR VARIANCE-increase degrees of freedom-have more specific restrictions of participants-could run within subjects design(second hypothesis): participants who saw inconsistent presentation state names more slowly~Only mention one mean because the other was already mentioned (avoid redundancies) LSD test-least significant difference -contrast t tests are exactly the same as LSD test when you have pairwise comparisons -LSD= t times the square root of MSW times
View Full Document