205 Flashcards
Dependent Variable
Outcome variable that is NOT manipulated (changed by independent variable)
Independent Variable
Variable that is manipulated (changes dependent variable)
Antecedent Variable
Occurs before the study and can affect the dependent variable
Extraneous Variable
Outside variable that may interfere with the results
How to achieve the most control
Randomization and elimination of bias
Internal Validity
Asks if it was the independent variable that changed the dependent
External Validity
Extent to which results of the study can be extended/generalized to other populations/studies
Threats to Internal Validity
Selection bias
Mortality
Maturation effects
Instrumentation effects
Testing effects
History threats
Selection Bias
The way participants are chosen/grouped
Mortality
Loss of attrition of participants
Instrumentation threats
Changes in how variables are observed
Testing Effects
Repeated testing may change their response
Maturation Effects
Changes in an individual over time
History Threats
Another event that affects the dependent variable
Threats to External Validity
Selection effects
Measurement effects
Reactive Effects
Selection Effects
Ideal sample can’t be attained (unbalanced population)
Measurement Effects
Using a pre-test lets the people think about responses for follow-up tests
Reactive Effects
Change in behaviour due to being studied
Control Group
Untreated
Experimental Group
Group being tested/manipulated
Homogeniety
Level of uniformity among sampling units
Pilot Study
Assess feasibility of an approach to be used in a larger scale
Bias
Not accurate, but precise
Random Error
Accurate, but not precise
Experimental
Randomization, highest level of control
Quasi-experimental
No randomization, some control
Non-experimental
no randomization, no control
True/Classic Experiment
experimental and control, pre-tests and post-tests (normal study)
Solomon four-group design
true/classic x2, needs double the amount of people (difficult)
After-only Design
No pre-test
Non-equivalent control group design
Same as classic, no randomization
After-only nonequivalent control group design
No randomization, no pre-test
One group (pretest-posttest)
Only experimental group (no control) and no randomization
Time Series design
2 pretests and 2 posttest
Single-Blind study
only participants are blinded
Double-blind study
Participants and investigators are blinded
Triple-blind study
Participants, investigators, and data managers are blinded
Anonymity
Being unknown to most people
Assent
Expression of approval/agreement
Benefits
Positive impact from research to party/participant
Beneficence
Acts of kindness/charity that go beyond duty - DO NOT HARM
Confidentiality
Protecting identity of research subject