PSYC*2360 Chapter 10: Within-Subjects Design Flashcards
What is a behavioural diary?
A self-report data collection strategy where participants record their behaviours and associated feelings as they occur
T or F: When completing a behavioural diary, behaviour is typically recorded at predetermined times.
True
What are the three common within-subjects designs?
- Pretest-posttest
- Repeated-measures
- Longitudinal
What is the simplest within-subjects design?
Pretest-posttest
What occurs during a pretest-posttest design?
Participants are measured before and after exposure to a treatment or intervention
What type of measurement is taken as an initial assessment of a participant at the onset of a study, prior to any treatment or intervention?
Baseline measurements
What is the “pretest” measurement in a pretest-posttest design?
Baseline measurements
What occurs during a repeated-measures design?
Participants are exposed to each level of the IV and are measured on the DV after each level
How do repeated-measures designs differ from within-subjects design?
- Repeated-measures: Don’t take baseline measurements
- Within-subjects: Take baseline measurements
What occurs during a longitudinal design?
Participants are repeatedly measured on the DV over an extended period of time
Do within-subjects or between-subjects design require a smaller sample size?
Within-subjects
Do within-subjects designs have a statistical power advantage or disadvantage over between-subjects designs?
Statistical power advantage
Why do within-subjects designs have a statistical power advantage over between-subjects designs?
The reduction of random error between groups makes it easier to isolate the IV’s impact on the DV
What are three disadvantages of within-subjects designs?
- Potential for lower external validity
- Potential for lower internal validity
- Potential logistical challenges
Why might within-subjects designs have the potential for lower external validity?
Observed effects may be the result of multiple exposures rather than a single expose to the IV
What are six potential threats to the internal validity of a within-subjects design?
- Attrition
- The testing effect
- Instrumentation problems
- The threat of history
- The threat of maturation
- Order effects
What is attrition also known as?
Mortality
What is attrition?
The differential dropping out of participants from a study
Does attrition represent a threat to the internal or external validity of a within-subjects design?
Internal
What is one way to minimize attrition?
By making continued involvement in the study appealing and non-threatening to participants
What is the testing effect?
When participants’ scores change on subsequent measured due to increased familiarity with the instrument
Does the testing effect represent a threat to the internal or external validity of a within-subjects design?
Internal
What are three ways to minimize potential testing effects in a within-subjects design?
- Use distractor items to make it more difficult to guess the study’s purpose
- Employ an unrelated distractor task between conditions to decrease the likelihood of participants remembering their previous answers
- Increase the time between conditions to also decrease the likelihood of remembering previous answers
What are instrumentation problems?
A change in how a variable is measured or administered over the course of a study
Do instrumentation problems pose a threat to the internal or external validity of a within-subjects design?
Internal
How can instrumentation problems be minimized?
By keeping measurement instruments and how they are administered consistent throughout the study
In a within-subjects design, what is the threat of history?
The possibility that an external event will influence participants’ behaviour during the study
Does the threat of history represent a threat to the internal or external validity of a within-subjects design?
Internal
In a within-subjects design, what is the threat of maturation?
The possibility of either short- or long- term physiological changes that would naturally occur, regardless of participation in the study
Does the threat of maturation represent a threat to the internal or external validity of a within-subjects design?
Internal
What is one way to determine the potential role of maturation in a within-subjects design?
Use a comparison group that doesn’t receive any intervention, but can be assessed over the same time period to observe natural changes
In which type of within-subjects design is the threat of maturation particularly problematic?
Longitudinal studies
What are order effects?
The influence that the sequence of experimental conditions/ IV levels have on the DV
Do order effects present a threat to the internal or external validity of a within-subjects design?
Internal
What are four types of order effects?
- Practice effects
- Fatigue effects
- Carryover effects
- Sensitization effects
What are practice effects?
Changes in responding or behaviour due to increased experience with the measurement instrument
How can practice effects be minimized?
Provide participants with practice completing the measures or doing the task before starting the experiment
What are fatigue effects?
Deterioration in the quality of measurements due to the participants becoming tired, less attentive, or carless during the course of the study
How can fatigue effects be minimized?
Make experimental tasks interesting and relatively brief
What are carryover effects?
When exposure to earlier experimental conditions influences responses to subsequent conditions
How can carry over effects be minimized?
By giving participants breaks between conditions
What are sensitization effects?
When continued exposure to experimental conditions increases the likelihood of hypothesis guessing
How can sensitization effects be minimized?
Mislead participants about the study’s purpose
What are three ways to minimize order effects overall?
- Randomizing the order of levels
- Counterbalancing
- Using a latin square design
Why is randomizing the order of levels in a within-subjects design not ideal for controlling order effects?
Some sequences may occur less frequently than others or not at all
How do researchers counterbalance to avoid order effects?
Identify and randomly assign all participants to potential treatment sequences
How is the number of orders for a given number of levels of the IV calculated?
N! where N= the number of levels of the IV
In which counterbalancing procedure does each experimental condition appear at every position in the sequence order equally often?
Latin square design
How does a latin square design reduce the number of conditions needed in a within-subjects experiment?
By creating a subset of all orderings with each condition in each position only once
T or F: Latin square designs only use as many sequences as there are experimental conditions/ IV levels.
True
Latin square designs are typically used when there are more than how many levels of the IV?
4 or more levels
Latin square designs are particularly useful for counterbalancing sequence order in which type of within-subjects design?
Repeated-measures
The list of students maintained by a psychology department who will receive credit in their class if they participate in a research study is an example of what?
A research participant pool
What is a t-test for dependent means also known as?
- Dependent means t-test
- Paired-samples t-test
In a within-subjects design with two levels of the IV and a continuous DV, what is the appropriate statistical analysis?
A t-test for dependent means
What does a t-test for dependent means determine?
If there is a statistically significant difference between two related sets of scores
When are repeated-measures ANOVAs used?
To test a hypothesis from a within subjects design with three or more conditions
What is the difference between a type one and a type two error?
- Type 1: Null hypothesis rejected, but it was actually true (there was no difference)
- Type 2: Null hypothesis failed to be rejected, but it was actually false (there was a difference)
Does alpha represent the type one or type two error probability?
Type one
Does beta represent the type one or type two error probability?
Type two
What is the typical cut-off value for a type two error (beta)?
0.2
what does (1-beta) measure?
Power