statistics p2 Flashcards
norm referencing
individual test score compared to average score of a group of individuals (norm group) (ex: GRE, SAT, ACT, career inventories)
criterion referencing
compares test scores to a predetermined standard/value/set criterion (ex: BDI, drivers license, MAST, college exams against a standard)
derived scores
conversions of raw scores that allow us to make further comparisons of individual score to norm score
percentiles
percentage of people falling at or below a score
-range= 1-99, mean=50
standard score
raw score–> new score with a new meaning and standard deviation
Z score
standard score with a mean of 0 and SD of 1
z= x-m/sd
conversion formula
score= z(SD)+ MEAN
T score
standard score with a mean of 50 and SD of 10
-generally used with personality tests
DIQ= deviation IQ scores
generally used with intelligence testing
mean= 100
sd=15
stanines
only whole numbers mean= 5 sd=2 range 1-9 used for achievement testing
stens
only whole numbers mean=5.5 sd=2 range 1-10 used for personality inventories and questionnaires
NCE
aka normal curve equivalent: used in education community
mean= 50
sd=21.06
range 1-99 in equal units along the curve
SAT
mean= 500 SD= 100
ACT
mean= 21
sd=5
age comparisons
comparison of an individual score to average score of others who are the same age
purpose: see how your performance compares to performance of others at your age or other ages
grade equivalents
compare individual score to average score of children AT SAME GRADE LEVel
**DO NOT COMPARE SCORES AT OTHER AGES: while a student may have a higher GE, doesn’t mean student has actually mastered content at that grade level
standard error of measurement
range/band where individual true score lies–where would they score in test was taken over and over
true score obtained if test were perfectly reliable
CONFIDENCE BAND AROUND OBTAINED SCORE
-relationship with reliability–as reliability decreases, SEM increases
standard error of estimate
confidence interval around PREDICTED score
-based off scores from one interval–allows us to predict where new score will fall (a range) on another variable
nominal scale
numbers arbitrarily assigned to categories
ordinal
numbers assigned to categories with magnitude/rank order, no true “distance”
interval
true distance between measurements but no real zero
ratio
true zero and equal distance between measures