Experimental Research Flashcards
What is the purpose of experimental research?
To determine cause and effect (manipulating the IV to see its effect on the DV)
What are the criteria to establish cause and effect?
- The cause must come before (precede) the effect
- The cause and effect must be correlated
- The correlation cannot be explained by another variable
True or False?
Cause and effect are established by statistics.
False. Cause and effect are established by logical thinking and well-designed experiments.
What can statistics do?
- Reject the null hypothesis
- Establish effect size
What is internal validity?
Determines if the results are due to the treatment.
What do you need to show internal validity?
Control over the participants and environment.
What is external validity?
Determines if the results are generalizable.
What do you need to show external validity?
A sample that looks like (represents) our population.
What is the relationship between external and internal validity and why?
- Inverse
- Experiments have high control, the real world has little control
What are the threats to internal validity?
Give examples for them.
- History (Pandemic)
- Maturation (Puberty)
- Testing (Pre-test)
- Instrumentation (Data Error)
- Statistical Regression (Highest move towards middle)
- Selection Bias (Different groups to begin)
- Experimental Mortality (Withdrawing)
- Selection - maturation interaction (groups maturing at different rates)
- Expectancy (may treat someone better)
Which threat(s) does randomization control?
Internal Validity
Maturation, History, Statistical Regression, Selection Bias, Selection-maturation interaction
What threat(s) does calibration control?
Instrumentation
What threat(s) does interest/incentives control?
Internal Validity
Mortality
What threat(s) does study design control?
Internal Validity
Testing, expectancy
What are the threats to external validity?
Give examples.
- Testing (pre-test)
- Selection Bias (may only work for the selected population)
- Setting (Very controlled settings)
- Multiple Treatments (getting more than one treatment does not happen often in real-world)
What threat(s) does research design control?
External Validity
Testing, Multiple Treatment Interference
What threat(s) does randomization control?
External Validity
Selection Bias
What threat(s) does the researcher control?
Experimental Setting
When looking at experimental designs… what does ROT stand for?
R - Random
O - Observation
T - Treatment
Describe pre-experimental designs
- Control very few threats
- Not good designs
- None are randomized
What are the pre-experimental designs and what is wrong with them?
- One-shot study (nothing to compare and no internal validity)
- One-group Pretest-Posttest Design (nothing to compare)
- Static Group Comparison (groups are not randomized)
Which design is this:
T O
One-shot study
Which design is this:
O T O
One-group Pretest-Posttest Study
Which design is this:
T O
O
Static Group Comparison
Describe True experimental designs
- All have randomization
- Have control over IV
- Have control groups
- Can control many threats
What are the true experimental designs and what do they show?
- Randomized Groups (Randomization allows us to assume differences are due to treatment)
- Randomized Groups Pretest-Postest ((Randomization allows us to assume differences are due to treatment with a pretest)
- Solomon Four (controls testing threat)
- Quasi-Experimental (fits real world)
- Non-equivalent Control Group (Very common)
- Time Series (show change over time)
- Reversal Design (want stability over time)
- Switched Replications (No problems)
What is wrong with Solomon Four and Switched Replications?
Solomon Four = Double the subjects and which tests
Switched replications = A lot of subjects and pain to analyze
What is wrong with Quasi-experimental?
Not randomized
Which design is this:
R T O
R O
Randomized Groups
Which design is this:
O T O
O O
Non-equivalent Control Group
Which design is this:
O O O O T O O O O
Time Series
Which design is this:
R O T O
R O O
Randomized Groups Pretest - Posttest
Which design is this:
R O T O
R O O
R T O
R O
Solomon Four
Which design is this:
O O T O O T O O
Reversal Design
Which design is this:
O T O O O
O O T O O
O O O T O
O O O O T
Switched Replications