1/24
Vocabulary flashcards covering psychometric concepts, reliability measures, and test scoring formulas from lecture notes.
Spearman Brown Prophecy Formula
A formula used to calculate how much longer a test needs to be to achieve a desired reliability, or to estimate reliability after splitting or doubling a test.
Split-half reliability
A reliability method that divides a test into two halves, which underestimates overall test reliability and yields different correlations depending on how the test is split.
Kuder Richardson 20
An internal consistency formula published in 1937 for tests with right or wrong answers that averages all possible split-half test combinations.
Cronbach's coefficient alpha
An internal consistency measure for tests without right or wrong answers that averages all possible split-half combinations and incorporates the Spearman-Brown correction.
Factor analysis
A statistical technique used to analyze relationships between a series of items to identify underlying scales.
Z score
A standard score with a mean of 0 and a standard deviation of 1.
T score
A capital T standard score with a mean of 50 and a standard deviation of 10.
Research reliability threshold
A reliability coefficient of 0.7 or above, which is considered acceptable for research purposes.
Decision-making reliability threshold
A reliability coefficient of 0.9 or above, required when test scores are used for hiring or high-stakes decision-making.
Internal consistency
The most commonly used reliability measure for psychological tests, requiring only a single test administration.
Test-retest method
A reliability measurement method that requires administering the exact same test to the same group of participants at a later time.
Alternate forms
A reliability assessment method that requires creating a completely new set of test items with different wording.
Inter-rater reliability
A reliability approach (also called interscore reliability) used when assessments require human judgment, calculated as the correlation between scores given by different raters.
Interval scale
A measurement scale in behavioral science where equal gaps are assumed between choices and there is no true or absolute zero.
Observed score
The actual score recorded from a test, defined mathematically as true score plus error (Observed Score=True Score+Error).
True score
An individual's actual underlying score, estimated by calculating measurement error and subtracting it from the observed score.
Standard Error of Measurement
A value calculated from the test standard deviation and reliability coefficient that measures how much scores tend to vary due to error.
95% Confidence Interval
A range calculated as the observed score plus or minus 1.96 times the standard error of measurement, within which the true score is 95% likely to fall.
1.96
The specific z-score multiplier representing the exact point in a normal distribution where 95% of scores fall around the mean.
Cut score
A fixed score threshold (such as 110) on high-stakes testing used to determine pass or fail status.
1937
The year Kuder and Richardson published their paper introducing Formula 20 for internal consistency.
0.8166667
The Cronbach's coefficient alpha value obtained for the first scale (reduced physical activity behaviors) in the climate and health study.
0.56
The Cronbach's coefficient alpha value for the third scale in the climate study, indicating the items were too unreliable to function as a unified scale.
Shorter tests
Tests that are less reliable than longer tests because they sample fewer items from the infinite pool of possible questions.
Spearman-Brown test multiplication example
An example showing that increasing test reliability from 0.5 to 0.8 requires multiplying the number of items by 4.