PSY 355 1st EditionFinal Exam Study GuidePSY 355 Exam 4 Study GuideLecture: Mixed Factorial Designs Factorial Design—2 (or more) IV’s -Repeated measure on one Indep. Variable -Between groups measure on the other Question: How to get people to contribute more to charity event? Two Independent VariablesType of Charity (Between)--- Mental Health Relief, Animal Shelter Animal in Ad (Within)--- Dog, Cat, Rabbit How much ($) would you give? (Graph the means) Partitioning the Variance -Finding the F ratios that are important a little difficult -F’s that you want are1. Main Effect for Type of Charity (Between Groups)2. Main Effect for Type of Animal 3. Interaction of Charity X Animal (Both Within Groups) F-Ratios (divide each by own term) Between SubjectsCharityTOTALWithin SubjectsAnimalxcharity/AnimalF-Ratios (divide each by own term) (see beachlab.com)20 20 10 40 60 30Degrees of Freedom calculation Between Subjects: [Charity (2) X # Subjects in each charity (10)] – 1 = 19 Charity: [Charity (2) – 1] = 1Subjects w/in each gp: [Charity (2) X (# S’s each gp – 1)] = 18 Note: df’s for Charity & S’s within each gp add up to the df’s Between Subjects (1 + 18 = 19) Degrees of Freedom calculation Within Subjects:[Charity (2) X #S’s in each Charity (10)] X [Animals (3) – 1] = 40 Animals [Animals – 1] = 2 Animals X Charity [A – 1] X [C – 1] = 2 Charity X Subjects Within Groups [Charity X (A – 1) X (# S’s/gp – 1)] 2X2X9 =36 Note that all the Within Subject terms add up to the total df for Within Subjects 2 + 2 + 36 = 40 Mixed Factorial Design -Some Variables can be Repeated Measured while others are between groups -The difficult part is knowing which term is correct for the F ratio. -Computer program may do the analysis for you, but you need to know which variables are within versus between -Several Variations on this design -MANOVA, ANCOVA MANOVA -Suppose in the Charity study we had several dependent measures: 1. Money they would contribute2. How much money they thought someone else would contribute 3. How much they liked the ad -Probably all very correlated. Analyze each separately? Increased risk of Type 1 error. Why? -MANOVA will analyze these multiple measures together, so that their ―shared variance does not ‖create the mathematical possibility of a Type 1 error. -You definitely need a computer to do this. Don’t try this at home... Analysis of Covariance -Example: Can women be more assertive with a woman than with a man? -40 Female Participants -Assess assertiveness (score from 0 – 100) -Then randomly assign to role play situations in which a Male or Female makes an unreasonable request (―Please go 50 miles out of your way to pick up my order at REI because I don’t want to pay for the gas to do it myself. ) ‖-Examples of assertive responses...? –-Participant makes response and is scored from 1 – 5 by raters. -After 10 responses, she has a totaled score from 10 (low) – 50 (high).-How to analyze this data set? -Answer: Treat Male or Female variable as a Between Groups Variable and -Treat the Trait Assertiveness variable as a correlational variable (note, NOT IV)‖Result: -Main effect for ―male or female making request‖-Association (sort of main effect) for ―Trait assertiveness with responses ‖-Most Importantly, Interaction! -Example: Most women, regardless of their ―trait assertion score, have trouble being assertive with ‖man, but only those with low scores have trouble being assertive with a woman. SUMMARYFactorial Design—2 (or more) IV’s -Repeated measure on one Indep. Variable -Between groups measure on the other F’s that you want are1) Main Effect for Between Groups IV 2) Main Effect for Within Subjects3) Interaction of Both Variables (Both Within Groups) SUMMARY CONT. -MANOVA will analyze multiple measures together, so that their ―shared variance does not create the ‖mathematical possibility of a Type 1 error. -ANCOVA analyzes one variable as discrete and one as continuous. Lecture: One Independent Variable with Repeated Measures Designs Variations— -Sometimes called “single subject” or “small n designs” -How do you know if an Independent variable is affecting of changing dependent variable while only using a small number of subjects? -ABA Designs-Multiple Baseline Designs One Independent Variable with Repeated Measures Designs ABA Designs: A = Baseline Condition B = InterventionA = Back to baseline -Question: Does cocaine increase an animal’s activity level? -What’s the IV? What’s the DV?One Independent Variable with Repeated Measures Designs -Operational definition of “activity level”—sensors on cage floor that measure motion -So, just administer the drug and watch how active they are? (as compared to what?) -Need a baseline! 1. Observe the rat’s activity level (DV) for a baseline period (Baseline = A) 2. Administer the drug (IV) and observe activity level (Intervention = B) One Independent Variable with Repeated Measures Designs Problem: Potential confounding variables1. History (external changes in environment) 2. Maturation (internal changes in the rat) 3. Testing Effects -So we add a period in which we revert to baseline conditions (remove the IV) and measure the DV -What should we expect to happen? One Independent Variable with Repeated Measures Designs Other examples1. Depression and Medication a. Depression is episodic and naturally increases and decreases from time to time.b. How do we know if a particular medication relieves depression? c. ABA and if it’s effective, reinstate medication 2. Behavioral Intervention a. How do we know if an intervention to change a problem behavior is effective?b. Potential problem, however:What if we do not want to revert to baseline? One Independent Variable with Repeated Measures Designs How about a Multiple baseline design? -TWO or more dependent measures- DV1 and DV2 A = BaselineB = Intervention affecting DV1(only) C = Intervention affecting DV2 (only) One Independent Variable with Repeated Measures Designs-Example Behavioral Intervention—“The Man Who Would Not Brush His Teeth” -Background: VA Program to help/encourage/support long term inpatients to go to group homes in the community -Very demoralizing to spend your life in Hospital -Very expensive esp when not necessary -Challenge: Most had developed inappropriate behaviors over the years – e.g. did not make bed, dress in regular clothes, shower and so