Dependent Samples

1. Dependent Samples Within-Subjects Experimental Design

2. Within-Subjects Design

3. An Example Chocolate and Sustained Attention Task

4. An Experiment IV ? • Chocolate Level of IV? • 1 unit of chocolate • 0 units of chocolate DV? • Sustained attention (as measured by the # of A’s found in a letter canceling task)

5. An Experiment Hypotheses? • HA: Participants will differ in sustained attention before eating chocolate vs. after eating chocolate. • H0: Participants will not differ in sustained attention before eating chocolate vs. after eating chocolate.

6. Between-Subjects Design Different groups of participants receive different levels of the IV Each participant serves in only one condition of the IV Independent samples or matched samplesare used in each condition Within-Subjects Design One group of participants receives every level of the IV Each participant serves in all conditions of the IV The same sample is used in each condition Comparison of Between-Subjects and Within-Subjects Designs

7. Two Sample Designs

8. Within-Subjects (a.k.a. Repeated-Measures) Design • Within-Subjects – comparison of the treatment effect involves examining changes in performance within each participant across treatments. • Participants serve as their own controls • compare their response at one level of the IV to their response at a different level of the IV • Repeated-Measures – participants’ behavior is measured repeatedly

9. The Ultimate Matched Pair

10. Advantages of Within-Subjects Design • Controls for many of the participant variables that might systematically vary with the DV • Reduces variability (error variance) • More powerful design • Requires fewer participants

11. Disadvantages of Within-Subjects Design • More demanding on participants • Participant attrition/mortality • Loss of data • Carryover (a.k.a. Order) Effects

12. Carryover Effects Carryover effects: • Occur when a previous treatment alters the behavior observed in a subsequent treatment • scores in later conditions are higher or lower as a result of having been exposed to previous conditions • Become a confounding variable if not controlled

13. Carryover Effects • Practice effects/Learning • Fatigue effects

14. Experiment 1 • Suppose a researcher compares two different study methods, A and B. • Condition A: Participants read 10 pages of text and use a highlighter to mark the important points. Participants then take a test on this material. • Condition B: Participants read 10 pages of similar text and make up sample test questions and answers. Participants then take a test on this material.

15. Experiment 1 Continued • Suppose all participants in this experiment first experience Condition A (use a highlighter while reading) and then Condition B (write sample questions and answers). • Results: Participants perform better on the test of Condition A text material than on the test of Condition B material.

16. Experiment 1 Continued • Can we conclude that using a highlighter is a better study method than writing sample questions and answers?

17. Experiment 1 Continued • Not necessarily. Two possibilities exist. • The study method in Condition A (highlighting) is better than writing test questions and answers or • By the time they did Condition B, participants were tired or bored. • There is a confound due to carryover effects • This threatens internal validity

18. Experiment 2 • You are conducting a cola taste test to determine which of two brands of cola is preferred. The two brands are in identical containers and will be poured into identical cups. The participants will drink the same amount of each cola and will drink cola A before cola B. • Any problems with this design?

19. Counterbalancing Counterbalancing: • exposing different participants to different orders of conditions • helps prevent carryover effects from accumulating in one particular treatment condition

20. Types of Counterbalancing • 1. within-subject counterbalancing – presentation of different treatment sequences to the same participant. • 2. within-group counterbalancing – presentation of different treatment sequences to different participants.

21. Within-Subject Counterbalancing • Each participant experiences each condition of the experiment at least twice using different orders each time. • Used when each condition is brief • ABBA • Balance the order effects by presenting one sequence and then its opposite

22. Within-Group Counterbalancing 3 requirements of within-group counterbalancing: 1. each treatment must be presented to a participant an equal number of times 2. each treatment must occur an equal number of times at each position 3. each treatment must precede and follow each of the other treatments an equal number of times

23. Within-Group Counterbalancing 1 2 Participant 1 A B Participant 2 A B Participant 3 A B Participant 4 B A Participant 5 B A Participant 6 B A 1. each treatment must be presented to a participant an equal number of times 2. each treatment must occur an equal number of times at each position 3. each treatment must precede and follow each of the other treatments an equal number of times

24. Within-Group Counterbalancing 1 2 3 Participant 1 A B C Participant 2 A C B Participant 3 B A C Participant 4 B C A Participant 5 C A B Participant 6 C B A 1. each treatment must be presented to a participant an equal number of times 2. each treatment must occur an equal number of times at each position 3. each treatment must precede and follow each of the other treatments an equal number of times

25. Relationship between # of Conditions and # of Possible Orders • The number of possible orders is a multiple of the number of conditions • 2 conditions (A, B)  2 possible orders • 3 conditions (A, B, C)  6 possible orders • 4 conditions (A, B, C, D)  24 possible orders • 5 conditions  120 possible orders • 6 conditions  720 possible orders

26. Relationship between # of Conditions and # of Possible Orders How do you calculate the # of possible orders? • The number of orders is the number of conditions (N) factorial (N!) • N * (N-1) * (N-2) *… • 2! = 2 x 1 = 2 • 3! = 3 x 2 x 1 = 6 • 4! = 4 x 3 x 2 x 1 = 24

27. Complete Counterbalancing Complete counterbalancing: • all possible treatment sequences are presented • at least one participant receives each order of the conditions

28. Complete Counterbalancing Advantage: • Offers the best control for order effects Disadvantage: • Requires a large number of participants when have more than 4 conditions of the IV • usually only used for experiments with 4 or fewer conditions of the IV

29. Incomplete Counterbalancing Incomplete counterbalancing: • Only a portion of all possible treatment sequences are presented

30. Incomplete Counterbalancing How do you select the sequences? 1. Randomly 2. Random starting order with rotation

31. When is Counterbalancing Useful? • Counterbalancing can control carryover effects only if the effects induced by different orders are of the same approximate magnitude • Use counterbalancing when you have equal carryover effects • Do not use counterbalancing when you have differential carryover effects

32. Equal vs. Differential Carryover Effects Equal carryover effects: • scores in one condition are affected to the same extent regardless of which condition preceded it • Use a within-groups design with counterbalancing

33. Equal Carryover Effects Treatment Without Create Questions Highlighter Actual Tx Effect (diff=40-30=10) A then B B then A Carryover Tx Effect A then B B then A 20 + 10 = 30 20 + 20 = 40 20 Tx effect for highlighting Tx effect for creating ?s 20 + 20 = 40 20 + 20 = 40 20 for creating ?s + 10 for carryover 10 for highlighter + 10 for carryover 20 + 10 = 30 20 + 30 = 50 10 in B 10 in A Because the carryover effects are equal for each condition, they cancel each other out

34. Equal Carryover Effects • counterbalancing is effective when carryover effects are equal

35. Equal vs. Differential Carryover Effects Differential carryover effects: • scores in one condition depend on which condition preceded it • Use a between-subjects design (either independent or matched samples)

36. Differential Carryover Effects Treatment Without Create Questions Highlighter Actual Tx Effect (diff=40-30=10) A then B B then A Carryover Tx Effect A then B B then A 20 + 10 = 30 20 + 20 = 40 20 Tx effect for highlighting Tx effect for creating ?s 20 + 20 = 40 20 + 20 = 40 20 for creating ?s + 20 for carryover 10 for highlighter + 10 for carryover 20 + 10 = 30 20 + 40 = 60 10 in B 20 in A Because the carryover effects are NOT equal for each condition, they DO NOT cancel each other out

37. Differential Carryover Effects • Counterbalancing will NOT eliminate carryover effects when the carryover effects are NOT equal for each condition • Use a between-subjects design (either independent or matched samples)