I.I. QUALITY MATRIX (from August 2012)

   I realized an importance of item quality and so all my future tests will be carefully prepared through IVIQ competitions.

   NEW:
   I also plan to calculate correlations of my tests with Logima Strictica 36 by Robert Lato. Logima Strictica 36 (LS36) is strong and approachable test with many testees,
   and I believe in the quality of such correlations. If you have ever taken any of my tests, please send your LS36 score to iviqsub@gmail.com.

   After 20, 40, 80 and 120 testees, together with new norm reports, I'll publish detailed statistical information about quality of each test.
   First of all, I give some notes and explain my procedure.

   S=number of points achieved by all testees on a particular item
   R=number of points achieved by the best S testees

   For each solved item I'll calculate the ratio R/S.
   This ratio would ideally be equal to 1, and otherwise less than 1.

   If this ratio will be lower than 0.7, an item will be qualified as average, and that item can measure only specific intelligence quotient (SIQ).
   If this ratio will be lower than 0.5, an item will be qualified as suspicious, and that item can measure only knowledge-dependent intelligence quotient (DIQ).
   If this ratio will be lower than 0.3, an item will be qualified as bad, and that item can measure only mystical intelligence quotient (MIQ).

   An overall quality of each test will be calculated as average ratio of all solved items. Quality in the certain range (130-150, 140-160 or 150-170) will be calculated too.
   Top three tests in each range are highlighted. Read more...

   Click here for the INTERPRETATION OF QUALITY PERCENTAGE.

IQ test
the best score
good items
average items
suspicious items
bad items*
unsolved items
overall quality
IQ 130-150 quality
IQ 140-160 quality
IQ 150-170 quality
Lipt
10
9
1
0
0
0
85,3%
83,9%
78,2%
-
WIC 2014
36
29
5
2
0
0
83,4%
82,1%
78,1%
68,5%
Numerus Light (FINAL)
30
23
4
2
1
0
81,9%
79,4%
68,4%
53,5%
Numerus Light 2
35
32
3
0
1
0
81,3%
82,8%
76,6%
66,7%
Einplex (FINAL)
28
22
5
1
2
0
78%
75,4%
67,1%
58,4%
Simtollect
40
27
9
3
1
0
75%
83,3%
73,2%
61,2%
Numerus Classic
36
21
11
4
0
0
71,8%
75,2%
62%
54%
FREE Fall
25
16
11
2
1
0
69,3%
74,1%
66,6%
56,4%
Triplex Light
29.5
20
6
5
2
0
67,4%
75,1%
64,4%
51,2%
SPEED
25
14
11
3
2
0
67%
82,7%
72,7%
62,5%
HI-Qlass
31
19
10
3
3
1
66,7%
76,8%
62,2%
53,7%
LSHR Light (FINAL)
25
16
4
4
5
1
66,3%
72,9%
58,9%
49,1%
Elementary
19
10
8
3
1
3
64,8%
68%
59,2%
48,8%
LSHR Classic
27
14
13
8
1
0
63,5%
68,5%
58,1%
55,3%
Daster
20
10
10
4
1
0
63,2%
64,1%
55,1%
51,7%
Triplex
24
14
6
4
5
1
62,3%
70,2%
61,4%
44,7%
Numerus (FINAL)
29
12
8
9
1
0
61,4%
70,9%
60,9%
58%
LSHR
22.5
8
11
6
3
2
57,3%
61,7%
55,9%
55,4%
Pert
2
2
0
0
2
1
50%
-
-
45,5%

* Hard items are easily qualified as bad on a low sample of testees. Accurate estimates are expected for final norm reports.

Second attempt average improvement: 7% (based on more than 200 submissions)