Psychology2016

Cognition: Thinking, Intelligence, and Language 285

distribution (Eysenck, 1994): IQ is assumed to be normally distributed with a mean IQ
of 100 and a typical standard deviation of about 15 (the standard deviation can vary
according to the particular test). An IQ of 130, for example, would be two standard devi-
ations above the mean, whereas an IQ of 70 would be two standard deviations below
the mean, and in each case the person’s score is being compared to the population’s
average score.
With respect to validity and reliability, the professor’s test fares poorly. If the results
of the professor ’s test were compared with other established intelligence tests, there
would probably be no relationship at all. Golf scores have nothing to do with intelli-
gence, so the test is not a valid, or true, measure of intelligence.
On the other hand, his test might work well for some people and poorly for oth-
ers on the question of reliability. Some people who are good and regular golfers tend to
score about the same for each game that they play, so for them, the golf score IQ would
be fairly reliable. But others, especially those who do not play golf or play infrequently,
would have widely varying scores from game to game. For those people, the test would
be very unreliable, and if a test is unreliable for some, it’s not a good test.
A test can fail in validity but still be reliable. If for some reason Professor Stump-
water chose to use height as a measure of intelligence, an adult’s score on Stumpwater ’s
“test” would always be the same, as height does not change by very much after the late
teens. But the opposite is not true. If a test is unreliable, how can it accurately measure
what it is supposed to measure? For example, adult intelligence remains fairly constant.
If a test meant to measure that intelligence gave different scores at different times, it’s
obviously not a valid measure of intelligence.

Just because an IQ test gives the same score every time a person

takes it doesn’t mean that the score is actually measuring real

intelligence, right?

Figure 7.5 The Normal Curve
The percentages under each section of the normal curve represent the percentage of scores falling
within that section for each standard deviation (SD) from the mean. Scores on intelligence tests are typically
represented by the normal curve. The dotted vertical lines each represent one standard deviation from the
mean, which is always set at 100. For example, an IQ of 115 on the Wechsler represents one standard
deviation above the mean, and the area under the curve indicates that 34.13 percent of the population falls
between 100 and 115 on this test. to Learning Objectives A.2, A.3, A.4, and 1.8. Note: The figure
shows the mean and standard deviation for the Stanford-Binet Fourth Edition (Stanford-Binet 4). The Stan-
ford-Binet Fifth Edition was published in 2003 and now has a mean of 100 and a standard deviation of 15 for
composite scores.

-3

55

52

0.135

-4

40

36

0.003

-2

70

68

2.275

-1

85

84

15.856

0

100

100

50.00

1

115

116

84.134

2

130

132

97.725

3

145

148

99.865

4

160

164

99.997

Standard Deviations

Wechsler IQ

Stanford-Binet 4 IQ

Cumulative %

34.13%