Chapter 6—Methodological Control in Exp. Research Flashcards
(18 cards)
What is a between-subjects design? When are they used?
A genre of research designs where participants are assigned only to one level of the IV. Thus, we need one equivalent group per level of the IV.
They are used if the IV is a subject variable, or if the experience gained in one level of the condition would make it impossible for them to participate in other levels (ex. deception)
What is the main weakness of a between-subjects design? How can we accommodate for it?
Differences between the conditions may have to do with the differences between the individuals in those groups.
We can accommodate for this weakness by creating equivalent groups—groups equal to each other in every important way except for the levels of the IV
What are the two ways of creating equivalent groups, and what are their conventional procedures?
- We can simply use RA with increasing proficiency as group sizes increase (regression effects) to spread confounding factors evenly through the groups; the main procedure is Blocked RA
- For a smaller number of subjects, RA can fail to create equivalent groups. So we go through a process called matching to help ensure the groups are similar on the key extraneous variables.
- Measure participants on a relevant variable
- Pair/group people who have similar scores
- Randomly assign an equal amount of people in that
group to each condition
What is a matching variable?
In the matching procedure, this variable is an extraneous variable especially important to control. People are matched on their scores for this variable.
What is a within-subjects design?
A genre of research designs where each participant experiences all conditions in the study.
Used for convenience, if the population is scarce, and if participating in one condition doesn’t make it impossible to participate in another.
What is the main disadvantage of within-subjects designs? How can we accommodate for them?
Order effects: any change in performance caused just by the order in which conditions are presented, not the conditions themselves.
—> Progressive effects: performance changes steadily over time NOT because of the condition, but because of a period of time (ex. practice/fatigue).
—> Carry-over effects: how one experiencing one condition affects the next; a condition has a residual effect.
We can accommodate for them through counterbalancing
Describe the 3 types of counter balancing and when they’re used
Participants are tested once per condition:
- Complete counterbalancing: Every possible order of conditions is used, and participants are assigned to each order equally.
—> Used when you have few conditions (e.g., 2 or 3), since the number of orders grows factorially - Partial counterbalancing: Only some of the possible orders are used, not all.
—> Used when you have many conditions and complete counterbalancing would require too many participants
—> Common method: Latin Square
Participants are tested more than once per condition:
- Reverse counterbalancing: Each participant receives the conditions in one order, then again in the reverse order.
—> Used when you want to control for progressive effects (like fatigue or practice)
—> Common method: Block randomization
Describe the Latin Square method of partial counterbalancing
Ensures each condition appears once in each position and follows every other condition equally often.
Describe the Block Randomization method of reverse counterbalancing
Every condition occurs once randomly but cannot be repeated until all conditions occur. Then, they repeat in a random order and so on.
What are error bars?
Bars on a line graph or bar plot that indicate the amount of variability in the mean; often reflects sd or confidence interval; mean +/- sd
What is a cross-sectional study? What approach does it take?
A quasi-experimental research design that compares different groups of people at a single point in time to examine differences or relationships.
It takes a between-subjects approach
What is the major detriment to cross-sectional studies?
Cohort effects: differences between cohorts could result from the effects of growing up in different times
Cohort: a group of people born at the same time
What is a longitudinal study? What approach does it take?
A single group of subjects is studied over time
It takes a within-subjects approach
What is the major drawback to longitudinal studies?
Attrition affects: the effects felt when participants drop out of the study; if certain types of people are more likely to leave (e.g., those with lower motivation, poorer health, etc.), the group that remains may no longer represent the original sample.
Threatens internal and external validity
What is a cohort sequential design? What approach does it take?
Researchers study multiple cohorts over shorter periods, and then piece together their results to simulate a longer developmental timeline.
Within-subjects: Each cohort is followed over time (like a longitudinal study).
Between-subjects: Different cohorts are compared (like a cross-sectional study).
What is experimenter bias? When does it occur?
When the researcher treats participants in different conditions differently
Occurs when an experimenter’s expectations for a study affect its outcome
What are the 3 ways we control for experimenter bias?
- Mechanizing procedures as much as possible
- Using research protocols: a step-by-step plan that ensures all participants are treated the same
- Double blind: neither the experimenters nor the participants know which condition is being tested on any trial
What is participant bias? What are the three main aspects that contribute to it?
A bias that occurs when the subjects’ behaviour is influenced by their beliefs about how they are SUPPOSED to behave in a study.
- Hawthorne effect: when behaviour is influenced by the knowledge that one is in an experiment and is therefore important to the study’s success
- Evaluation apprehension: participants want to be seen in a positive light, so they behave as they think the ideal person should
- Demand characteristics: aspects of the study that reveal the hypothesis being tested; if these features are too obvious, participants may no longer act naturally.
