Psyc 242 Lecture 7.2 Repeated Measures Design PDF

Psyc.242 Lecture 7.2 Methodological issues with within-subject design Advantage of repeated measures designs  No need to balance individual differences across conditions of experiment  Fewer participants needed  Convenient and efficient  More sensitive design Sensitivity  A sen...

Psyc.242 Lecture 7.2 Methodological issues with within-subject design Advantage of repeated measures designs  No need to balance individual differences across conditions of experiment  Fewer participants needed  Convenient and efficient  More sensitive design Sensitivity  A sensitive experiment  Can detect effects of IV even when IV has small effect  “Error variance” is reduced  Same people participate in each condition  Variability due to individual differences eliminated Disadvantage of repeated measures design: practice effect  People change as they are tested repeatedly.  Performance may improve over time.  People may become bored or tired as number of “trials” increases.  Practice effects become a potential confounding variable if not controlled. Practice Effects, example  Suppose a researcher compares two different study methods, A and B.  Condition A: Participants use a highlighter while reading a text, then take a test on the material.  Condition B: Participants read a text and make up sample test questions and answers, then take a test on the material.  Suppose  All participants first experience Condition A and then Condition B  Results: test score A > test score B  Potential Confounding variable: participants are bored when they work on Condition B  What if the result is A < B? Solution to practice effect  Balance practice effects across conditions.  Counterbalance the order of conditions  Half of the participants do Condition A, then B  The remaining participants do Condition B, then A  Random Selection presentation.  Randomly select the conditions and present to participants.  Can be applied when there are many conditions Techniques with Counterbalancing  Complete Balance design  Incomplete Balance design  All possible order  Select order (Latin square design) Complete Balance Design  Balance practice effects within each participant.  Each participant experiences each order of condition, at least once – everyone will experience each condition several times.  Each participant forms a “complete experiment.”  Use different orders each time  Use when each condition is brief, and the number of conditions are not too many  e.g., simple judgments about stimuli For example  1. There are two conditions: A and B. Participants will go though the conditions many times, in different orders. Such as: ABBAABBAABABBABA, etc..  2. There are three conditions: ABC. Participants will go through each order of conditions at least once. Such as: ABC, ACB, BAC, BCA, CAB, CBA Incomplete Balance Design  Each participant experiences each condition once (not each order of conditions).  Balance practice effects across participants.  General rule for balancing practice effects  Each condition must appear in each ordinal position (1st, 2nd, 3rd, etc.) equally often. Incomplete Balance Design: all possible order  All possible orders  Use with limited number of IV conditions (3 or 4)  3 conditions (A, B, C) → 6 possible orders  Randomly assign participants to one of the six orders  4 conditions (A, B, C, D) → 24 possible orders  Randomly assign participants to one of the 24 orders.  Need at least 1 participant randomly assigned to each order Incomplete Balance Design: select order (Latin square)  Selected orders  Select particular orders of conditions to balance practice effects  Each condition appears in each ordinal position exactly once.  Randomly assign each participant to one of the orders of conditions.  This is applicable to situations when the conditions are longer or more complicated to execute Latin square example: for conditions ABC (remember for three conditions, the total possible combination is 6) Ordinal position Participants 1st 2nd 3rd #1 A B C #2 B C A #3 C A B Latin square design example  https://www.youtube.com/watch?v=2FQZnybGrTY &t=7s In class exercise: can you create a Latin Square design for 4 participants over 4 conditions ABCD? Order Participants 1 2 3 4 #1 #2 #3 #4 Comparing the different designs  Complete Balance design vs. Incomplete Balance design ?  Counterbalance design vs. random selection presentation design? Complete vs. Incomplete balance design  Here, “Complete” means whether each participant goes through ALL conditions and ALL orders of the conditions.  In a Complete Balance design, the participant goes through ALL conditions and ALL orders of the conditions.  In an Incomplete Balance design, the participant still goes through ALL conditions, but just one order  All possible order vs. select order: whether all possible orders of the conditions will be administered When to use the two designs?  Complete balance design: applicable when the conditions are very brief  Incomplete balance design: applicable when the conditions are longer or more complicated to administer. (Especially the Select order design – conditions might be very complicated)  Number of conditions should not exceed 4 Counterbalance vs. Random selection presentation  In Counterbalance design, each participants goes through ALL conditions  In Random Selection presentation, each participant goes through some randomly chosen conditions – NOT necessarily ALL conditions  When there are more than 5 conditions, usually only the Random selection presentation design is appropriate.  When there are 4 or less conditions, either a counterbalance design or a random selection presentation can be used. Data analysis  Repeated measures ANOVA (the conditions are entered as the “repeated measure”) Hypothesis of repeated measure ANOVA  Null hypothesis: all of the conditions are equal  Research hypothesis: at least one of the conditions are different from the others 23 Results reported by a graph  Note that the columns 30 show the difference S 25 among the conditions, e T 20 NOT the order of testing a i 15 r m c e 10 h 5 0 10 15 20 Number of Distracters Assignments  Connect_Chapter 7  Quiz 7 available on Blackboard now.