Research Designs
Internal Validity • Isolation of the independent variable(s)
External Validity
1 Randomized designs
• Experiments where participants are randomly assigned to the experimental groups or conditions. • Often referred to as “true experiments.”
Randomized designs
• Provides a safeguard against biased assignment of sampling units to the different treatment groups. • Distributes the characteristics of the sampling units over the different conditions to prevent biased outcomes. • Permits the use of statistical analyses that require certain data characteristics.
2 Ways of achieving random assignment
• Presorting booklets or questionnaires • Blindly drawing names • Flipping a coin • Consulting a table of random numbers
Non factorial designs (one independent variable: one way): • Between subjects • Within subjects
Factorial designs: • Between subjects • Within subjects • Mixed
3 Between subjects designs
• Subjects are Condition A Condition B exposed to one condition each. Subject 1 Subject 2 • Also called nested Subject 3 Subject 4 designs. Subject 5 Subject 6 Subject 7 Subject 8 Subject 9 Subject 10
Within subjects designs
• Subjects are Condition A Condition B Subject 1 Subject 1 exposed to each Subject 2 Subject 2 condition. Subject 3 Subject 3 – Also called: Subject 4 Subject 4 • Repeated-measures Subject 5 Subject 5 design Subject 6 Subject 6 • Crossed design Subject 7 Subject 7 • Importance of Subject 8 Subject 8 counter-balancing Subject 9 Subject 9 Subject 10 Subject 10
4 Threats to internal validity
Between or within subjects designs
Threats of internal validity: • Researcher expecta***betweention • Participant expectati***betweenon • Participant selectio***betweenn • Maturation and historical factors • Lost of participa***betweennt • Habituation and fatigue • Statistical regression
5 Threats to internal validity: • Researcher expectan***withincy • Participant expectanc***withiny • Participant selection • Maturation and historical f***withinactors • Lost of participan***withint • Habituation and fatig***withinue • Statistical regressio***withinn
Controls:
• Constancy
• Systematic variation —counter-balancing: Latin Square
• Random variation
6 Latin Square
Latin Square
Order of administration 1 2 3 4 Sequence 1 A B C D
Sequence 2 B C D A
Sequence 3 C D A B
Sequence 4 D A B C
7 Within subject (repeated measures)
pre -test – V. I. – post test
pre -test – V. I. – post test pre-test post test
Solomon design
1. pre-test – I. V. – post test 2. I. V. – post test 3. pre-test – – post test 4. – post test
8 1. pre-test – I. V. – post test Solomon 2. I. V. – post test Design 3. pre-test – – post test 4. – post test Pre-test yes no
yes 1 2 I. V. no 3 4
Factorial design: • Between-subject • Within-subject • Mixed
9 What is the research design?
1. Find the independent variables
2. For each I. V.: • Within-subject or between subject? • levels (names & numbers)
2 X 2 mixed design with the independent variables x (within) & y (between)
Factorial designs
Hypotheses:
• Main effects (= number of I. V.) • Interaction
10 Main effects
English
English French
French
Males Females Age Sex
Interaction English French English
French
Males Females Age Sex
11 BURNHAM (1966) Expectancy control design
Expectancy Brain state Lesioned Unlesioned Totals
Lesioned 46.5 49.0 95.5
Unlesioned 48.2 58.3 106.5
Totals 94.7 107.3
I. V.? • Brain state Within- or between-subject ? Levels? 2 Lesioned vs. Not lesioned • Researcher Expectancy Within- or between-subject ? Levels? 2 Lesioned vs. Not lesioned
12 2 X 2 Between-subjects Design
Brain state
Lesioned Not-Lesioned
Belief: Belief: Belief: Belief: lesion No lesion lesion No lesion
13