chapter 5 Flashcards
(24 cards)
reliability vs validity
reliability = same every time
- if you were to do the study twice it would come out the same
validity = whether it measures what its supposed to measure
- accurately measures what you want it to
measurements
self-report
observational
physiological
self-report
asking directly about the concept, your own answers
scales & survey
ex- interviews
observations
behaviors, increase external validity
ex- a researcher could operationalize happiness by observing how many times a person smiles
physiological
recording biological data such as brain activity, hormone levels, or heart rate
ex- using an MRI machine to see different parts of the body
types of data
Nominal, ordinal, interval, ratio
quantitative vs categorical
nominal
is a type of data that is used to label variables without providing any quantitative value
ex- labels or tags
ordinal
when the variable represent a ranked order
ex- bookstores website ranks the top 10 best-selling books
interval
needs to meet two conditions
1. equal intervals between levels
2. no “true zero”
ex- IQ test, there is equal value between 104-106 & 33-35. also 0 doesn’t mean that the person has no intelligence
ratio
- equal intervals
- 0 really means 0, truly means “nothing”
ex- normal test score
categorical vs quantitative
categorical = categories; nominal qualitative = meaningful numbers; ordinal, interval, ratio
construct validity - what is in it
indication of how well a variable was measured in a study
face validity
content validity
criterion validity
face validity
on the face (surface) does it appear to be measuring the right thing
content validity
does it measure all aspects of the subject
ex- looking at a study of love & you are testing whether or not how many times a couple kisses, does that measure all aspects of love?
criterion validity
comparing your measure to behavior that could be associated with that concept
establishing validity
convergent & discriminant
both are still apart of construct validity
convergent validity
is it correlating with other scales? things that are supposed to be similar should correlate. The scale we have created will be compared with a scale that has ALREADY BEEN VALIDATED & see if they correlate
discriminant validity
when a scale correlates with another scale it shouldn’t be
ex- a IQ test correlating with a love test
reliability
the consistency of the results of a measure
reliability
test-restest reliability
inter-rate reliability
internal reliability
test-retest reliability
test them & then resets to see if you get the same results
w/o experiment
inner-rate reliability
only for observations
internal reliability
only scales, does each item measure the concepts as well as the others
each item should correlate
cronbach’s alpha
test the internal reliability of a scale
