Experiments Flashcards
Causality is (and conditions) …
Relationship between two parameters, where the first one causes the second one (condition and consequence). Concomitant variation, time order, control of other factors are the conditions.
Concomittant variation (correlation) is …
How the condition and consequence vary, which is predicted by the hypothesis
Extraneous variables are …
Other variables, except for the independent, that affect the response of the test
Internal validity is …
Whether the manipulated independent variable caused the observed effects on the dependent variable - control the extraneous variable
External validity is …
Can the cause-and-effect relationship be generalized
Mortality (MO) is …
The loss of test units while the experiment is in progress
Extraneous variables can be controlled by …
Randomization, matching, statistical control, design control
Randomization is …
Random assignment of test units to the groups
Matching is …
Before assigning test units to treatment conditions, they are compared to a set of key background variables
Statistical control is …
Measuring and adjusting the extraneous variables through statistical analysis
Design control is …
Use the designed experiments to control a set extraneous variables
Pre-experiment means …
NO control of extraneous factors, NO randomization. Includes one-shot case study, one group pretest-posttest, static group
True experimental design is …
Randomization. Includes pretest-posttest control group, posttest-only control group
Quasi-experimental design is …
Not possible to get full manipulation. Includes time series, multiple time series
Statistical design is …
A series of basic experiments that allows control of external variables. Includes randomized blocks, latin square, factorial design
One-shot case study (pre-experimental)
One group exposed, one measurement taken, no randomization
One-group pretest-posttest (pre-experimental)
One group, measured before and after
Static group (pre-experimental)
Two groups, where one is exposed, but the control is not, measured only after, no randomization
Pretest-posttest control group (true experimental)
Two groups, where one is exposed, but the control is not, measured before and after, randomization
Posttest-only control group (true experimental)
Two groups, where one is exposed, but the control is not, measured only after, randomization
Time series (quasi-experimental)
One group, multiple measurements before and after, no randomization
Multiple time series (quasi-experimental)
Two groups, where the control is not exposed, multiple measurements, no randomization
Randomization block (statistical)
One major external variable (experience), the test units are grouped based on this
Latin square design (statistical)
Two non-interacting external variables
Factorial design (statistical)
2+ independent variables at different levels