T h e F i n a l T e s t
2nd preliminary norming
March, 1998
Paul Cooijmans
De Wolwever 39
5737 AD Lieshout
Netherlands
The Final Test was published in Noesis 124 and In-Genius 126, and can
also be found on http://www.eskimo.com/~miyaguch . A paper copy can be obtained
from the above address for one dollar.
The test originally consisted of 61 verbal analogies, but was trimmed to
52 when the first preliminary norming was performed early 1997. This norming was
represented by the formula "IQ = raw score x 1.01 + 128.22". The removed items
were, according to the numbering on the 62-item version, 4, 6, 10, 11, 12, 13,
21, 33 and 42.
A second norming study is undertaken to make full use of the now availa-
ble data and encourage potential testees to take the test. The scores so far are
2, 2, 4, 4.5, 5, 11.5, 15.5, 17.5, 19.5, 22.5, 33, 35, 36, 44, 45.5 and 47 right
out of 52. The below table shows correlations with prior test scores for tests
for which more than 3 scores were reported. To make maximum use of the available
prior scores, the remaining scores are combined in (apparently) coherent groups
to make computation of aggregate correlations possible.
Tests # r
Association/Number (subtests of long TFG) 4 .99
Titan (Hoeflin) 5 .98
Mega (Hoeflin) 5 .97
Test for Genius (short) 14 .96
Test for Genius (long, Total score) 7 .96
Nemesis Test/Test To End All Tests/Chimera Test (latter
by Bill Bultas) 10 .96
WAIS/Wechsler-Bellevue/GRE/MAT/LAIT/LSFIT 7 .93
Chimera High Ability Riddle Test (Bultas) 9 .8
Cattell Culture Fair/Drenth number series/NUTE number
series 6 .78
Ultra/Power (Hoeflin) 5 .73
W-87/Quest/Möbius 5 .37
Used in the norming are all except W-87, Quest and Möbius. In case of
the Nemesis Test, Test To End All Tests, Chimera Test and Power Test this means
to use early preliminary normings; the present author however does not at all
consider that a problem. The 72 used score pairs are arranged in two columns in
numerical order, the highest raw score facing the highest prior score, etc. All
prior scores are, where needed, converted to IQ's on a scale with 16 points per
standard deviation. Basically the median of the prior IQ's facing each raw score
is used to norm that score, and missing scores are normed by interpolation.
Scores below 2 are given as "lower than" and scores above 47 are normed on face
value. The following table results ("G" is the number of analogies answered
correctly):
G IQ %ile 15 144 99.7 34 164 99.997
16 145 99.75 35 165 99.998
17 146 99.79 36 166 99.9983
18 147 99.83 37 167 99.9987
0 <127 19 148 99.87 38 168 99.999
1 <127 20 149 99.89 39 169 99.9994
2 127 95.4 21 149 99.89 40 170 99.9995
3 129 96.5 22 150 99.91 41 171 99.9996
4 132 97.7 23 151 99.93 42 172 99.9997
5 133 98 24 152 99.94 43 173 99.99976
6 134 98.3 25 154 99.96 44 174 99.9998
7 135 98.5 26 155 99.97 45 175 99.99985
8 136 98.7 27 156 99.98 46 176 99.9999
9 137 98.9 28 158 99.986 47 177 99.99992
10 138 99.1 29 159 99.989 48 181 99.99998
11 140 99.3 30 160 99.992 49 186 99.999997
12 141 99.4 31 161 99.993 50 192 99.9999994
13 142 99.5 32 162 99.994 51 199 99.99999996
14 143 99.6 33 163 99.996 52 >207 99.999999999
Two items are unsolved to date. Item analysis will be done as soon as
the harder problems are solved by sufficient numbers of testees. The test was
free in the late 1996/early 1997 period, but currently the testing fee is US $5
or 10 Dutch guilders.
This test is accepted for admission to The Glia Society (22 right) and
The Giga Society (50.5 right), and other societies may also accept it. The Glia
Society's norm on this test used to be IQ 160, as that society's policy is to
demand a higher score on one-sided tests than on wide-range tests, but in view of
the Final Test's apparent very high correlation with wide-range tests the
Administrator of TGS felt forced to lower the norm to IQ 150, just as it is on
wide-range tests.
Return to the The Final Test
Return to the Uncommonly
Difficult I.Q. Tests page.