Validity Flashcards
(41 cards)
what is internal validity
the ability of your research design to test the hypothesis
for what it was designed for
= change in the independent variable caused change in the dependent variable
what can threat internal validity
- extraneous variables
2. confounding variables
extraneous variables
(alternative explanations for the observed effect) = rival hypothesis
confounding variables
= if two independent variables influence the dependent at the same level and you cannot distinguish which variable has actually caused the observed effect
Threats to internal validity
- History
- Maturation
- Testing (pretesting)
- Instrumentation
- Statistical regression
- Biased selection of participants
- Experimental mortality
Threat: History
multiple observations over time
-> different events can occur in between = affecting the results
Threat: Maturation
Effect of age or fatigue
- can affect performance changes (confounding variable)
- cannot difference which variable caused the outcome
Threat: testing
making pretests before the actual treatment
- subjects get presensitized
= better performance afterwards
Threat: Instrumentation
unobserved changes in an criteria or measurement instrument.
- difficult to distinguish what caused the actual response
Statistical regression
if you have chosen participants which an extreme score,
- Tendency in the second test/measurement = regression to the population mean
Biased selection of participants
initial differences in the subjects–> leads to changes in the performance/results
- cannot separate the effect of the independent variable
Experimental Mortality
participants drop out of the study due to death, moving
external validity
study results can be generalized beyond the research setting
- applying to real world problems
What is the aim of for external validity ?
gaining inside into the underlying mechanisms of the studied behavior
Threats to external validity
- Highly controlled laboratory settings
- Reactive testing
- Interaction: selection particip. bias and IV
- reactivity of experimental arrangements
- multiple treatment inference
reactive testing
pretest influences participant’s reaction on an experimental variable.
Interaction: selection bias
study results can only be applied to studied group of subjects (e.g college students, gender )
Reactivity of the experimental arrangements
- artificial setting of the experimental situation
- participant’s knowledge that they part of the study
multiple treatment inference
- treatment in the first condition may affect the performance in the later condition (e.g. fatigue)
Problem with internal and external validity ? interaction
if one increases the other usually decreases
What kinds of research settings are available ?
- Laboratory setting
2. Field setting
What are the characteristics of the laboratory setting ?
=mostly artificial setting
+ control over extraneous variables (which could affect your dependent variable)
subcategory: simulation
What are the characteristics of the field setting ?
= studying behavior in naturally environment
+ higher generalization
- no control over extraneous variables (confounding variables)
What is a sample?
small subgroup of a larger population