PSYCH 210 Lecture 25 Outline of Last Lecture I. Factorial ANOVAa. Calculational example with hypothesis testing stepsOutline of Current Lecture I. Factorial ANOVAa. Interpretation of interactionb. AssumptionsII. CorrelationCurrent LectureI. Hmwk 10 due Friday at 1:00pma. Practice problems on Learn@UWb. Review Session Saturday 1-2:30c. Exam Sunday 2:45pm in room 121 Psychologyi. SPSS Assignment can be turned in during examII. Factorial ANOVAa. Hypothesis Testing Stepsi. For the interaction1. Look at the pattern by making a line graph of the cell meansa. Remember that ‘nonparallel’ line rule is just a guess!i. To determine whether the interaction is significant, lines must be significantly nonparallel (i.e. the p-value is the definite determinant of significance. However, sometimes we only have the graph to use)b. (For table) Marginal Means go with Main Effects!2. Post-Hoc Testa. When do we need post-hoc tests?i. For interaction, run post-hoc if:1. Overall F is significantb. Underlying concept: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.i. Is the pattern we got the pattern we predicted?c. One existing option:i. Simple main effects test1. Specifically for Interactions2. Breaks Interaction down into smaller piecesa. Are individual pairs significantly different?3. Describe pattern shown in grapha. Describing Interactions (conclusion)i. Include:1. IVs/Tx/Levels2. Specific DV (measurement)3. Full Comparisona. Talk about all conditions!4. Systematic descriptiona. Whether significant or non-significant!b. Assumptionsi. Interval or ratio dataii. Independent observationsiii. Underlying Normal distributioniv. Homogeneity of variance (in all conditions)III. Correlationa. Inferential or descriptive?i. Investigates the relationship between two DVs (x and y)1. Note that up until now, we’ve been looking at differences betweenmeansii. Correlation itself is a descriptive statistic1. Why are so many correlational studies done, if they cannot tell us about causation?a. When variables are impossible to manipulateb. When variables are unethical to manipulateb. Scatter Plotsi. How to construct1. 2 measures (can be of different scales) from each participant plotted on coordinateii. They can tell us about1. Direction of relationshipa. Ex) Positivei. As x increases, y increasesb. Ex) Negative (Inverse)i. As x increases, y decreases2. Degree of relationshipa. Regression Line or ‘Line of Best Fit’i. Weak relationship1. Lines very scattered, far from the regressionlineii. Strong Relationship1. Points not very scattered, close to lineiii. No scatter at all = perfect correlationc. Calculational example: Pearson r Correlationi. This single value can tell us1. Direction of relationship and degree of relationshipii. Ranges from +1 to -11. Sign indicates whether relationship is positive or negative2. Absolute Value describes strength of relationshipa. Values closer to 1 are stronger, values closer to 0 are weakeriii. Definitional formulas1. r= [Degree to which x&y vary together]/[ Degree ‘’vary separately]2. Similar to…a. t= [Actual diff. btwn Ms]/[Diffs. expected by
View Full Document