validity 395 test3
Terms
undefined, object
copy deck
- An experiment has internal validity if...
- ...a causal relationship can be established
- 3 parts of a causal relationship
- temporal precedence, relationship between the two variables, no alternative explanations
- fixing selection bias threat to internal validity
- random assignment
- fixing history threat to internal validity
- including control group
- fixing history threat to external validity
- hard to do
- fixing participant mortality threat to internal validity
- hard to do
- threat to longitudinal designs
- external: history
- fixing instrumentation threat to internal validity
- counterbalancing/randomizing order of experimental conditions
- selection by maturation
- differences between individuals in treatment groups that produce changes in groups at different rates
- fixing selection by maturation threat to internal validity
- use random assignment/control groups
- counterbalanced design controls for:
- internal: contamination and carryover
- experimental design validity threat:
- external
- experimental design controls for:
- internal validity threats
- Randomized two-group design controls for:
- Internal: Selection, maturation, history, instrumentation threats
- Pretest-posttest two-group design controls for:
- Internal: Selection, maturation, history, instrumentation threats
- Pretest-posttest two-group design validity threats:
- When the pretest affects both conditions equally, it’s a threat to external validity. When it affects conditions differentially, it’s a threat to internal validity.
- Solomon Four-Group Design controls for:
- Selection, maturation, history, instrumentation threats to internal validity
- Static Group Comparison validity threats:
- Selection, temporal precedence (internal validity)
- Pretest-Posttest Nonequivalent Control Group Design validity threats:
- Selection, but pretest helps
- One Group Pretest-Posttest Design controls for:
- Temporal precedence threats to validity; sometimes history, maturation, testing, instrumentation (internal)
- Replicated Interrupted Time-Series Design controls for:
- Threats to internal validity; especially sometimes history, maturation, testing, instrumentation (internal); alternate explanations