Slides 3 Flashcards
construct validity
the extent to which a measure can assess the construct of interest
how to test construct validity
correlations between related constructs
what does construct validity mean in an experimental context
what is actually driving the effects we are seeing in the experiment (assuming we have internal validity)
major threats to construct validity
placebo and participant expectancy - especially if the participant believes the treatment will be good
what is an active placebo
a pacebo that produces some of the somatic effects, mimicking the side effects of the actual treatment, but with no activity ingredient to remedy anything
single-operational/narrow stimulus sampling
- what is it
- what is it a threat to
it is when there is another factor that participants may be responding to other than the treatment. For example a therapist if the therapist is there for all of the experimental participants it could be that the therapist is really good rather than if the type of therapy worked.
- threat to construct validity
what are demand characteristics
cues of the situation associated with the study that seem incidental but may account for the results
- often happens during the informed consent process
statistical conclusion validity:
the extent to which the analysis preformed enables one to draw correct inferences about the phenomena of interest
state some threats to statistical conclusion validity
low power (small N, small effect size)
variability in procedures –> increases error in measurments
subject heterogeneity
unreliable measures
multiple comparisons
definition of concept: alpha
the probability of rejecting a hypothesis when that hypothesis is true
also called type 1 error
definition of concept: beta
the probability of accepting a hypothesis when it is false
also referred to as type II error
definition of concept: power
the likelihood of finding differences between conditions when in fact, the conditions are truly difference
also defined as 1-beta
effect size
a way of expressing the magnitude of the difference between conditions in terms of a common metric across measures and studies
what happens to the effect size when there are more methological problems in the study
the effect size gets smaller (even if in nature, the effect size is large)
effect size can be impact in two general ways:
1) increasing the difference between means
2) reducing the standard deviation by controlling for methological factors that increase variance