Reliability/Validity Terms Flashcards
Reliability definition
The consistency of a measure
3 types of reliability
internal consistency, stability, equivalence
Internal consistency
correlation of single item to the total scale/measure of how well a set of items on a test or survey are measuring the same thing/construct / a measure of internal consistency
Split half reliability
divide questions in half into 2 groups (odd numbered/even numbered), calculate score for each half, and compare both halves to determine correlation
Cronbach’s Alpha
Average of al possible split half reliabilities / how closely related a set of items are as a group by comparing the amount of shared variance among the items to the amount of overall variance
three or more response options
Kuder Richardson Coefficient
Average of all possible split-half reliabilities where each question only has 2 answers
Test-Retest Reliability
part of stability – giving identical measure to the same group of participants after time has passed
Interrater Reliability
Part of equivalence: Do multiple raters agree on ratings?
Alternate forms of reliability
Part of equivalence: administer more than one form of a test to the same people - check to see if the forms are equivalent
Validity
does the scale measure what it is supposed to measure?
content validity
does the instrument assess all aspects of the construct and exclude irrevelant issues? e.g. an awe scale w/ good content validity would cover both aspects and exclude irrelevant concepts
face validity
A type of content validity: does the test measure what it appears to measure?
criterion validity
how well a measure correlates with an established measure of comparison (criterion)
what are 2 ways to assess criterion validity?
concurrent and predictive validity
Concurrent validity
does the measure relate to other measures that assess the same thing (does a new measure of gratitude relate to other measures of gratitude)
Predictive validity
does the measure predict future behavior?
construct validity
how well does the scale measure the underlying concept?
What’s the first way to assess construct validity?
Convergent/divergent validity
2nd way to assess construct validity?
factorial validity
3rd way to assess construct validity?
Discriminant validity
convergent/divergent validity
is the measure related (convergent) or unrelated (divergent) to other constructs as one would expect? e.g. A new measure of gratitude is positively related to positive affect and unrelated to intelligence
Factorial validity
Factor analysis shows whether items measure construct in a coherent way – identifying underlying factors/patterns that influence observed variables - group together highly correlated variables - identifying cluster of items tend to be answered the same way
Discriminant validity
Discriminate between groups of people as expected
e.g. A measure of depression distinguishes between those diagnosed with depression and those who have not
Snowball Sample
Current participants help to recruit new ones from acquaintainces/Useful when population members are hard to locate/Hard to make inferences about population
Random Sample
Every member of the population has an equal chance of being selected
Stratified random sample
Individuals selected randomly within each strata (subgroup) of importance
e.g. E.g. divide Skidmore students into first year, sophomore, junior, and senior and randomly select within each subgroup
Convenience Sample
Selecting your sample based upon convenience – participants are readily accessible/ but Sample may not be representative of the population!!
Who launched positive psychology?
Martin Seligman (APA president from 1997-1998)