Midterm Chapter 6 Flashcards
- Compare between-subjects and within-subjects designs and explain the control problems (potential confounds) unique to each type of design and how they can be addressed.
- between subjects designs are where participants will be placed in treatment level A OR B but will not receive both. Whereas in a within subjects study, participants participate in both/all levels it is a within subjects design
For between subjects designs a potential confound is having inequivalent groups ie) one group is smarter than the other. A way to address this potential confound is to deliberately create equal groups using random assignment or matching.
For within subjects designs, order effect is the major potential confound, can be controlled using counterbalancing: (testing once per condition) complete, partial, (testing more than once per condition) reverse or block randomization.
name and describe the 4 types of counterbalancing and identify when they are used:
(when you are only testing each condition once)
1. complete counterbalancing: using every possible sequence at least once
2. partial counterbalancing: taking a random sample of orders and randomizing the order of the conditions for each subject (Latin Square)
(when you are testing more than once per condition)
1. reverse counterbalancing: the experimenter simply presents the conditions in one order, and then presents them again in the reverse order
2. block randomization: basically every condition must occur at least once before any condition can be repeated
what is a latin square, and why is it beneficial
a system that “creates” orders to ensure that each condition will occur only once and in a different position each time
a.) latin square ensures every condition of the study occurs equally often in every sequential position and
c.) ensures every condition precedes and follows every other condition exactly once
What is the essential control problem (potential confound) to be solved in a between-subjects design and what are two ways of solving the problem
Unequal groups
solution 1: random assignment
Solution 2: matching
Briefly define and explain the relationship between participant effects, experimenter effects, single-blind experiments, and double-blind experiments.
Participant effects: good subject effect, evaluation apprehension, and demand characteristics
Experimenter effects: experimenter bias, experimenter expectancy effect
Single Blind experiments are experiments where the participants are naive to what is being studied, but the researchers are not.
Double blind experiments are where both testers and participants are kept in the dark regarding what is going on in the study
Interaction: If the experimenter gives away their hypothesis or acts in a way to sway the participant, it may reveal demand characteristics meaning the participant may pick up on things encouraging them to act in a way to confirm the hypothesis, or they may want the researcher to think they are a good participant. Simply the presence of an observer can also influence participant behavior contributing to a participant effect
what is the hawthorne effect?
when participants know they are in an experiment and therefore act differently because they know they are important to the studies success.
What is the essential control problem to be solved in a within-subjects design and what is the general procedure for solving this problem?
the essential problem is order effect and these can be mediated using different counterbalancing methods
Describe two procedures for accomplishing counterbalancing for studies in which each participant encounters each level of the independent variable just one time.
- Complete CB
- Partial CB
Describe two procedures for accomplishing counterbalancing for studies in which each participant encounters each level of the independent variable more than one time.
- Reverse counterbalancing
- block randomization
