exam 2 Flashcards
manipulation
changing one variable (independent variable) to determine its affect on another (dependent variable)
measurement
collection of data that allows you to determine a quantifiable change/difference in variables of interest
within-subjects experimental design
comparing two or more sets of scores within one group of individuals to see how dependent scores have changed across independent variable groups
between-subject experimental design
comparing two or more sets of scores across two or more groups of individuals to see how dependent variables scores across the subjects
Why are between-subjects good?
it removes the effects of contamination across conditions
- practice or experience
- fatigue or boredom (situps)
- contrast reactions or effects (testing two drugs on same subject)
- necessity
faults of between-subjects experiments
- need for a large sample group
- confounding variables (8am vs 2pm groups)
- variability variance
randomization
restricted random assignment (equal groups)
How do we minimize variability issues?
- standardizing procedures and treatments (keep everything constant)
- controlling for individual differences (keep the groups balanced)
- larger sample size
central limit theorem
as our sample size gets larger, our group averages from the populations are less varied
unequal (differential) attrition
effectiveness of highly involved treatment versus control group (being nice or mean to the average students)
ways of diffusion of effects
- shared information (horoscope experiment where the groups talked to each other and then knew what was going on)
- compensatory rivalry (video game experiment where everyone wanted to be the best: texters stopped texting, callers stopped listening)
- resentful demoralization (the control group becoming mad they arent being helped)
advantages of within-subject designs
- smaller sample requirements
- control over differences between groups
- control over differences between individuals
dangers of within-subject experiments
- overall
- third-variable interaction effects
- internal validity issues
- enviromental factors
- time-related factors
- order effects
- carryover effect
counterbalancing
switching the order of condition presentation in order to test and control for order effects
(have some people part of the control group first and then some people part of the experimental group first)
matched subject designs
pairing individuals that are considered equivalent and placing them into separate conditions
nonexperimental research stragety
comparisons across groups where we make no attempts to avoid threats to internal validity
ex. no manipulation
quasi-experimental research stragety
approximation of a true experiment due to its attempts to avoid threats to internal validity
ex. how people react after earthquakes but you can’t control when the earthquake happens
internal validity
validity established if the study produces a single, unambiguous explanation for the relationship between variables