Testing Intelligence

download report

Transcript Testing Intelligence

Testing Intelligence
ARE YOU OVER-TESTED?
 Your
age group is the most tested group in
the history of the United States.
Types of Tests
 Aptitude
Test - predicts your ability to
learn a new skill. (future performance)
 Achievement Test - reflects what you
have learned. (current performance)
Wechsler Adult Intelligence
Scale (WAIS)
 The
most widely used intelligence test
created by David Wechsler.
 Contains verbal and performance subtests
 Look at page 446 in test to see sample
questions.
 If scores on verbal and performance are
drastically different, it could indicate a
learning disability.
Test Criteria
A
PSYCHOLOGICAL TESTS SHOULD
BE



Standardized
Reliable
Valid
Standardization
Standardization – defining meaningful scores
by comparison with the performance of a
standardized group. Norms (normal scores) are
established through field test questions.
 For a test to be standardized, the condition of
taking the test need to be as similar as possible
 Results should form a normal curve (bell
shaped) with most scores falling in the average
and less score falling in the extremes.
 A score is given based on how the test-taker’s
score varies from the mean.

The Normal (Bell) Curve
 Standardized
test results usually form a
bell-shaped curve (normal curve – most
scores fall near the average and few
scores lie at the extremes). REFER TO
PAGE 447
The Normal (Bell) Curve and
Standard Deviation
 The
Standard Deviation (z score) for the
normal curve on the WAIS and other
standardized Intelligence tests is always
15. The S.D. is how much scores vary
from the mean.
Normal (Bell) Curve and
Standard Deviation
 What
this means is about 68% of the test
takers will score between 15 points lower
than the mean and 15 points higher than
the mean because 68% is ALWAYS 1z
(z=standard deviation). Q What are these
scores on the WAIS?. The other 32% that
take a test score outside of that.
Standard Deviation
 About
95% of test takers will score
between 30 pts. lower than the mean and
30 pts. higher than the mean because
about 95% is ALWAYS 2z (standard
deviation). Q What are these scores on
the WAIS?
The Bell Curve
If the mean on a test is 80 and the standard
deviation (1z score) is 5
1. 68% of the test takers scored between a
___ and ____?
2. What percent of the test takers scored
between a 70 and a 90?
3. If I scored 1 standard deviation above the
mean what was my score?
4. If I scored 3 standard deviations below
the mean what was my score?
If the mean on a test is 80 and the standard
deviation ((1)z score) is 5
 What
percent of test takers scored above
a 90?
 If I scored a 70 what is my z score?
 On the Normal curve for a standardized
intelligence test (ex.- the WAIS) what
percent of the population scores above an
85?
 What percent scored above a 115?
Test to Measure Optimism and
Pessimism in Personality
 1.
Prefer A. Hamburgers B. Hot Dogs
 2. What kinds of grades do you tend to
get? A. Good B. Bad
 3. Which do you prefer? A. Talking on
telephone B. Going to Movies
 4. Which color do you like best? A Red B
Blue
Validity
– the extent to which a test
measures or predicts what it is supposed
to, the accuracy of the test.
 Validity
 Was
the Optimism test valid?
 Can you make it valid?
 (glass, thoughts on rainy days)
Types of Validity

Content validity – How well does the test
measure the entire range of material it is
supposed to test?


Ex. of low content validity – spending 5 minutes on
depression in class and asking 30 questions on
depression for the Ch. 16 test.
Predictive validity – the success of the test in
predicting the behavior it is supposed to predict.


Ex. SAT is said to predict your performance in
college. It actually only predicts freshman year
success. High school grades much better predictor.
Criterion – the behavior a test is supposed to
measure.
Reliability
– the extent to which a test
yields consistent results, consistency of
the test.
 Reliability
Measuring Reliability
Equivalent –form reliability – correlation
between performance on different forms of the
test. Q State check for on EOC’s?
 Test-retest reliability – correlation between
one’s score of the first administration of the test
and the score of the second administration of the
test. (Ex. SAT’s)
 Split-half reliability – randomly dividing the
tests into two parts and correlating one’s
performance on the two parts.
 The closer the correlation is to +1 … the more
reliable the test is


Stanford-Binet and the WAIS have a reliability of +.9
Testing Mental Retardation

Those whose IQ score falls below 70 are labeled
as having mental retardation. (pg. 452)
 Criteria for diagnosis:



Low test score
Difficulty adapting to the normal demands for
independent living
Mental retardation affects 1% of the population.
 Mental retardation affects males more than
females