Transcript SPHS 5780 Diagnostics Lecture 06c psychometrics, NEW

TYPES OF SCORES
(PP. 117-121 IN HUTCHINSON;
MERRELL & PLANTE;
SECOND PART OF HAYNES AND PINDZOLA )
TYPES OF SCORES
Item score
Raw score
Converted scores (See normal curve…)
 percentile rank
 Standard scores
z-scores
T-scores
Stanine
 age-equivalent and grade-equivalent scores
Examples
of item
scores
More
examples
of item
scores.
Items
scores
together
allow us to
calculate
raw score
TYPES OF SCORES
Item score
Raw score
Converted scores (See normal curve…)
 percentile rank
 Standard scores (based on central tendency and distribution around measure of central
tendency)
z-scores
T-scores
Stanine
 age-equivalent and grade-equivalent scores
Converted scores all
us to compare client’s
score with that of the
normative sample
(sometimes called
standardization
sample)
PERCENTILE RANK
Tells us what proportion of the norming population scored lower than the subject
99th percentile means out of 100 children 99 scored lower.
10th percentile means out of 100 children 10 scored lower.
To understand standard scores, you need to understand..
Measures of central tendency and variability……
30
Esp. mean and standard
deviation
20
Example: 4:6
year olds
10
0
0
58
204
To understand standard scores, you need to understand..
30
Standard deviation, in-class discussion
(Standard deviation as the “average” distance of all
deviations from the measure of central tendency)
20
Example: 4:6
year olds
10
0
0
58
204
In a normal (bell-shaped curve)
 68% of the population will have a mean score between
+/- 1 sd (e.g., b/w 85 and 115 on this sample IQ test)
 95% of the population will have a mean score between
+/- 2 sd (e.g., b/w 70 and 130 on this sample IQ test)
Note that standard deviation and percentile rank are related!
•Raw score transformed to standard scores…. (3 types)
30
Standard
scores
20
Example: 4:6
year olds
10
0
0
24
70
40
85
58
100
76
115
95
130
•Raw score
transformed to
standard scores….
(We’ll consider
three types of
standard scores)
Transform to a distribution of Z-scores

70
85
100
115
130
Note that standard deviation and percentile rank AND z-scores
are related!
Note that
standard
deviation and
percentile rank
are related.
Standard
scores like zscores and tscores say the
same thing,
just in a
different
language, as
do stanines
TYPES OF SCORES:
AGE/GRADE-EQUIVALENT SCORES
(“DEVELOPMENTAL SCORES” LAWRENCE, 1992)
E.g. Jack’s raw score: 59
age-equivalent score “5;2”
Jack’s 59 is average (mean or median) raw score expected for
child who is 5;2
(AGE EQUIVALENT SCORE, CONT.)
Problems? Age equivalent score doesn’t consider:
Range of performance for that age
Someone whose chronological age is higher than
their age-equivalent score may still be performing
normally
e.g. child 6;2 with age-equivalent score of 5;2
Same score for kids of different ages may come
from qualitatively different patterns of performance
(AGE EQUIVALENT SCORE, CONT.)
Useful to:
Help parents understand extreme cases
Meet agency requirements
BUT, rule of thumb:
Don’t use these scores!
AGE EQUIVALENT FUNCTION
Normreferenced tests
will give you
table with
converted
scores, and will
give you place
to record them