Switch branches/tags
Nothing to show
Find file
Fetching contributors…
Cannot retrieve contributors at this time
50 lines (48 sloc) 2.63 KB
Latin Square
* With a perfectly balanced Latin Square:
1. Every condition of the study occurs equally often in every sequential position
2. Every condition precedes and follows every other condition exactly once
* Important:
* If you use a Latin Square design, you have to use a N that is a multiple of the number of conditions you have
Reverse Counterbalancing
* The experimenter presents the conditions in one order and then presents them again in the reverse order
Block Randomization
* Similar to block randomization in between-subjects design
* Example, if we wanted 5 trials of each N-back size we would make 5 random sequences
Controlling Order Effects
* Counterbalancing assumes that the effects are going to be linear, but this isn't always the case
Carryover Effects
* Another way to do it is to use a washout period
* This is a period of time in which the effect of the independent variable can dissipate
* Can be minutes, hours, or days
Order Effects
* Our discussion assumes that we know when we have order effects
* In reality, we can make educated guesses about whether we will have them.
* The only way to really know is to use a factorial design.
* Not always bad
* Imagine you were interested in the carryover effects of a drug on a particular behavior.
* The effect of motivation of success followed by failure, and vice versa.
Characteristics of a Good Manipulation
* Construct Validity
* What you measure is what you mean to measure
* Reliability
* If you apply the manipulation you should see the same results each time
* Strength
* The conditions of the independent variable are different enough to actually create different behaviors in the group.
* Be sure not to make them too extreme as they may become unethical or not ecologically valid
* Salience
* A salient research manipulation stands out from all the other stuff going on in the study.
* If it's not salient, participants will not notice it and also are unlikely to be affected by it
Manipulation Check
* Interview participants to see if participants experienced the manipulation as intended
* You can also administer measures of the construct being measured.
* Done either:
* In a pilot study
* Post-experiment
When Manipulation Checks Fail
* It could mean the manipulation is not valid
* It could also mean it was not strong enough
Using Multiple Stimuli
* Stimulus: a person, object, or event that represents the operational definition of a condition of the independent variable.
* When the stimulus reflects a hypothetical construct, it can be represented by any number of stimuli
* So when you only use one set of stimuli, you confound the stimuli with the hypothetical construct