Document 1 - Purdue
Document 1 - Purdue
Document 1 - Purdue
Preliminary Manual
Purpose
The PURDUE NON-LANGUAGE TEST is a "culture-fair" test designed to measure mental ability. Since it
consists entirely of geometric forms, there is evidence to show that it can be used effectively with
persons having markely different cultural or educational backgrounds.
After the test booklets have been distributed, the participants/testtakers should be instructed as follows:
1. Print your name on the test booklet or answer sheet (if answers sheets are provided).
2. Do not open the test booklet until told to do so.
3. On the front page, look at the sample of five (5) drawings. In each set of five (5) drawings, one
drawing is different from the others. Locate the drawing that is different and cross it out with an X as
in Sample A.
4. In Sample B, one of the 5 drawings is different from the other 4. Cross out this different drawing
with a big X.
5. Look at Sample C. In this sample, also cross out the drawing that is different from the other 4 with a
great big X.
6. In taking the test, you will find sets of five (5) drawings opposite each number. In each set of five
drawings, cross out the different one with a good big X.
7. You may not be able to identify the different drawing in all of the sets, but do the best you can. It
will not pay you to spend a great deal of time on any one set. You will have 25 minutes, which is plenty
of time if you do not spend too much time on some of the sets.
8. Do not skip around. Take the sets in order.
9. Are there any questions? If none, open your test booklet and go ahead.
10. At the end of exactly 25 minutes, say: STOP. Please pass in your test booklet.
Reliability
The reliability of the Purdue Non-Language Test has been estimated by: (1) coefficients of equivalence
between the two forms, and (2) internal consistency coefficients using Horst's modification of KR-20 and
of versus even item splits stepped-up by the Spearman-Brown formula. The obtained reliabilities are as
follows:
Validity
No situational validity studies of the test are yet available. However, the test correlates (r = .65) with
scores on the Adaptability Test for 184 industrial salaried employees, and (r = .55) with the Adaptability
Test for 99 laborers. These studies clearly show that the Purdue Non-Language Test is tapping a
significant portion of the variance covered by a standard mental test.
Construction
One hundred and twenty (120) geometric-form items were constructed as the original item pool. These
120 items were divided randomly into two forms (referred to as Form Z-1 and Form Z-2) because the
subjects on whom the item statistics were to be obtained could not be held long enough for all of them
to finish the 120 items. These subjects were the 184 students from Southern Indiana high schools, 37
being from a special division for retarded students. One hundred and thirty (130) of the students were
available for two class periods and they took both forms Z-1 and Z-2 of the test.
The tests were administered to groups of students during regular school hours. In the analysis, all tests
which were marked with multiple answers on five or more items, and all tests having ten or more
omissions (items not marked) were discarded. By this procedure, eight (8) papers were rejected. The
remaining papers (159 on Form Z-1 and 147 on Form Z-2) were scored with unit weight being assigned
for each correct answer.
The 159 papers on Form Z-1 were divided into high and low criterion groups using total score on Form Z-
1 as the criterion (an internal criterion). As the nhumber of papers was relatively small for item analysis
purposes, the total 159 papers (81 and 79 papers in the high and low criterion groups, respectively) were
used for determining item statistics. The proportions of people in the total group answering each item
on Form Z-1 correctly was determined, and this proportion, expressed as percent, was used as an index
of the level of difficulty of each item. A discrimination value (internal D-value) was also determined for
each item using Lawshe's monograph for estimating the validity of each item.
The procedure was repeated for the 147 papers on Form Z-2 (73 and 74 papers in the high and low
criterion groups, respectively).
On the basis of the item statistics, data thus obtained, items were selected to construct the two matched
final forms --- Forms A and B. The criteria for selecting an item to be included in either final form were
(a) an item had to have a D-value of .5 or higher, and (b) the level of difficulty of an item had to be
greater than could be accounted for by chance (i.e., 20%). The items selected for the two final forms
were matched on the basis of D-value and difficulty level; hence, an item selected for inclusion in one
form was matched on the other form by an item similar as possible in both D-value and difficulty. The
items in each form were arranged in order of difficulty.
Scoring
Form A
____________________________________________________________________________
Item Number Correct Answer Item Number Correct Answer
____________________________________________________________________________
1 5/E 25 2/B
2 4/D 26 5/E
3 4/D 27 4/D
4 2/B 28 3/C
5 4/D 29 1/A
6 4/D 30 4/D
7 5/E 31 5/E
8 5/E 32 3/C
9 2/B 33 5/E
10 3/C 34 5/E
11 3/C 35 4/D
12 4/D 36 5/E
13 3/C 37 2/B
14 2/B 38 1/A
15 4/D 39 3/C
16 4/D 40 4/D
17 5/E 41 2/B
18 2/B 42 3/C
19 4/D 43 4/D
20 1/A 44 4/D
21 5/E 45 4/D
22 3/C 46 3/C
23 3/C 47 4/D
24 4/D 48 4/D
____________________________________________________________________________
Form B
____________________________________________________________________________
Item Number Correct Answer Item Number Correct Answer
____________________________________________________________________________
1 4/D 25 5/E
2 1/A 26 5/E
3 4/D 27 5/E
4 3/C 28 4/D
5 3/C 29 5/E
6 1/A 30 4/D
7 3/C 31 1/A
8 5/E 32 2/B
9 3/C 33 4/D
10 1/A 34 4/D
11 1/A 35 4/D
12 2/B 36 5/E
13 3/C 37 1/A
14 5/E 38 1/A
15 3/C 39 1/A
16 3/C 40 3/C
17 5/E 41 5/E
18 5/E 42 3/C
19 4/D 43 4/D
20 3/C 44 4/D
21 5/E 45 4/D
22 3/C 46 4/D
23 3/C 47 2/B
24 3/C 48 5/E
____________________________________________________________________________
*Use Form B
Norms
Four sets of norms are presently available. These norms are summarized below.
____________________________________________________________________________
Percentile US High Venezuelan Industrial Civil Service
School High School Supervisory Laborers
N = 77 N = 89 Personnel N = 99
N = 184
____________________________________________________________________________
99 47 43 47 45
95 46 41 46 42
90 45 40 45 40
80 43 38 43 37
70 41 36 42 34
60 39 34 40 32
50 38 32 38 29
40 37 30 36 27
30 36 29 32 25
20 34 27 29 22
10 28 24 27 17
5 26 21 23 12
____________________________________________________________________________
Lawshe, C.H. (1942). A monograph for estimating the validity of test items. Journal of
/nen/6/7/21