Experimental Design: One-Way Correlated Samples Design

1
Chapter 12

• Experimental Design One-Way Correlated Samples
  Design

2
Experimental Design One-Way Correlated Samples
Design

• Advantages and limitations
• Comparing two groups
• Comparing t-test to ANOVA
• Comparing more than two groups

3
Advantages and limitations

• One-way correlated samples
• One-way 1 IV
• Correlated samples no random assignment
• Each score in one group (condition) is paired
  with a score in the other group(s) (condition(s))
• Advantages
• Can reduce systematic error (confounding)
• Can reduce random error (due to indiv diff)
• Limitations
• Creating pairs of participant scores may be
  difficult
• Repeated measurements can create methodological
  concerns

4
Advantages and limitations

• Natural pairs
• Participants scores paired for some natural
  reason
• Matched pairs
• Participants scores paired because researcher
  matches them on some variable
• Repeated measures
• Participants scores paired because they come
  from the same participants
• Objective is to reduce sources of extraneous
  variability

5
Advantages and limitations

• Advantages of repeated measures design
• Controls EVs due to individual differences
• Requires fewer participants
• Appropriate for studying questions that involve
  repeated exposure/testing
• Appropriate for longitudinal research

6
Advantages and limitations

• Methodological issues of repeated measures design
• Carryover effects
• Transient
• Permanent
• Sensitization
• Carryover effects can often be controlled by
• Randomized order of conditions
• counterbalancing

7
Advantages and limitations

• Comparing repeated measures design to independent
  samples design
• Effect on random error and inferential statistic
• Effect on degrees of freedom
• Consider the net effect

8
Comparing two groups

• Random sampling
• Paired assignment to 2 groups (conditions)
• 1 IV with 2 levels
• Lets try an experiment involving the Stroop
  effect
• Go to the following website
• http//faculty.washington.edu/chudler/java/ready.h
  tml

9
Comparing two groups

• variability within groups (error variability)
  random error (extraneous variables)
• variability between groups systematic error
  (confounds) systematic variability (effect of
  IV)
• Goals
• Reduce random error
• Eliminate systematic error
• Maximize systematic variability through
  manipulation of IV

10
Comparing t-test to ANOVA

• Correlated samples t-test
• Limited to 2 groups
• Independent samples Analysis of Variance (ANOVA)
• 2 or more groups
• Both parametric tests
• Require assumptions of
• Normality
• Homogeneity of variance

11
Comparing t-test to ANOVA

• Correlated samples t-test
• difference between the 2 group means
• t ----------------------------------------------
  ------------
• standard error of the difference
  between means
• t values when null hypothesis is true
• t values when null hypothesis is false
• Larger the t (pos or neg), the lower the
  probability that the difference is simply due to
  chance
• Alpha level and decision-making

12
Comparing t-test to ANOVA

• Correlated samples ANOVA
• variability between the two groups
• F ----------------------------------------------
  ------------
• error variability
• F values when null hypothesis is true
• F values when null hypothesis is false
• Larger the F, the lower the probability that the
  difference is simply due to chance
• Alpha level and decision-making

13
Comparing more than 2 groups

• Addition of groups often clarifies relationship
  between IV and DV
• ANOVA to determine effect
• A priori specific comparison test
• Does not require significant F
• post hoc specific comparison test
• Does require significant F

14
Summary

• Correlated samples design
• Random sampling
• Paired assignment
• Natural pairs
• Matched pairs
• Repeated measures
• Paired assignment designed to reduce random error
• Manipulation of IV
• Analyzed with t-test or ANOVA