Experiments Flashcards
Causal Relationships
-1.) Independent variable must precede in time the dependent variable
2.) Independent and dependent variable must go together in a meaningful way
3.) Changes observed in dependent variable must be results of changed in independent variable not other variables
Correlational Relationships
- When statistics may back up the relationship between variables but there is not a meaningful connections
Control groups
a.) exposes research participants to the independent variable
b.)rules out initial differences between conditions
c.) control for the effects of extraneous influences
Treatment Groups
The group that will receive the manipulation of the variable
Creating Equivalent groups
Random Assignments
Pretests
Radom Assignments
- Each research participant has equal chance of being assigned to a particular group
gives best possible of having an experiment with high validity
Pretests
Measure research participants on relevant variables that need to be accounted for before exposing the treatment groups to the variables
extraneous influences
1.)conducted
2.)Researcher
3.)Subjects
Conducted Extraneous Influences
1.) Treatment Reliability
2.)Threshold effects
Researcher Extraneous Influences
1.) Assistant
2.)Double Blind
3.)Confederates
Subjects Extraneous variables
1.)Hawthorne
2.)Morality
3.)Bias
Experimental designs
- Full Experiment
-Quasi experiment
-Pre-experiment
Full Experiment
-Pre-test-posttest equivalent groups
-Posttest-only equivalent groups
-Solomon four-group
Quasi-Experiments
-single group interrupted timed series
-Pretest-posttest quasi equivilant groups
-Interrupted time series quasi- equivalent
Pre-experimental
-One group post test only
-one-group pretest post test
posttest-only nonequivelent groups
lab versus field work
lab work will have more control while field work has higher ecological validity
advantages of experiments
- The researcher(s) has more control over the variables that are included in the study. You get to choose (as opposed to more inductive studies).
- There are set, standard methods to check for high quality (e.g., reliability coefficient, etc.).
- Experiments CAN be (not always) time efficient if you are looking at perceptions or reported communication.
- A lot of people are familiar with the form/method/steps of experiments.
Disadvantages of experiments
- An experiment may be sterile. You may miss some important variable that impacts/explains the phenomenon you are looking at. If you use a
coding schema, there is no room for other behaviors/variables/features that may emerge. - Experiments typically do NOT deal with the complexities of human communication. The subtleties of communication are often lost or overlooked when
applying a coding schema. - Experiments are usually NOT helpful in analyzing the impact of the context/setting on the communication you are looking at.
- The study may NOT be ecologically valid or like the “real world.” The lab setting or the treatment may not be realistic.
- Experiments are often costly and tedious to use (e.g., large sample size, collecting large amounts of data, analyzing large amounts of data, etc.).
- Experiments often do NOT get at the underlying, deeper structures of meaning in communication.