Bhatia’s Battery of Intelligence Tests

Name: Ishwari Sinha

Enrollment Number: A1506920422

Introduction
The Bhatia's performance Test of Intelligence is a well standardized and reliable tool for the
assessment of intelligence in the Indian context. The Battery is applicable to the illiterate as well
as the literate groups with separate norms provided for each group. While the test was originally
developed for, the 11-16 years age group, the use of this test on adults beyond 16 years of age is
based on the assumption that intelligence does not increase beyond 16 years of age which was set
as the upper limit of the test. The full battery requires 45 to 60 minutes for its administration
(Bhatia C. M., 1955). Murthy, (1966) and Pershad et al., (1988) have proposed the use of
abbreviated forms of the Bhatia Scale, while Verma et al., (1988), have utilized a scoring
procedure which is different from the method proposed by Bhatia, methods proposed by Murthy
and Verma have been compared.
The aim of the investigation conducted by Bhatia, (1951) was “to formulate a test of intelligence
which may reach the majority of the Indian population and not merely touch its fringes - the
educated middle class”, to hopefully, in turn, open a field which had been closed off to the Indian
educational psychologist.
For this purpose, the type of intelligence tests used were performance tests of Intelligence (or
more specifically, those tests of intelligence which do not require reading and writing on the part
of the subject). Thus, verbal tests were ruled out, however, a test that could be orally
administered and scored by means of oral responses could be considered (for instance,
Immediate Memory Test).
Since the testing would require actual manipulation of varying objects, along with oral tests, only
individual testing can be taken up.
This battery of tests was used to test the illiterate population of children, as well as the school-
going population equally in order to standardize it for the literate and illiterate group separately.
The age-range selected for the investigation was 11 to 16 years and only hoys were included as
subjects, girls being excluded altogether. Girls were excluded not only because of the discovery
of sex differences in (at the time) recent investigations of the Performance Tests and the Space
factor (Thomson, G. H., 1940; Emmett, W. G.), but also because there were social difficulties in
securing girls as subjects.
Defining Intelligence
The definition of intelligence for this test, has derived heavily from the views of S. C. Kohs,
(1923) who devised his "Block Design Test" which has proved to be one of the most effective
tests of intelligence up to that time; Kohs was quite emphatic about his views on intelligence and
even went as far as to try to justify them on philosophical grounds. Be that as it may, his
emphasis on the "analytic-synthetic" activity of the mind as the most characteristic feature of
intelligence is very pertinent to the discussion extended by Bhatia.
Kohs description of the synthetic and analytic activities was in the following words-
1. “By “synthesis” we mean, on the one hand, the intensity of that fundamental force or
condition innate in nervous protoplasm which binds neurons into complex systems, end,
on the other, the capacity of a living organism to construct out of mental elements and
fragmentary experiences, concepts and notions of a higher order”.
2. “By “analysis” we mean the capacity for observing or discovering parts or differences in
objects or qualities which for themselves seem unitary”.
3. “Although one can only speculate on these matters, it seems reasonable that analysis and
synthesis are but the head and tail of a single function-tendency”.
4. “This “analytic-synthetic” activity may be regarded a fundamental property characteristic
of all irritable tissue, and more markedly 30 of nervous tissue. All forms and degrees of
this function-tendency seem possible, from the simplest to the most complex.”
5. “It seems evident that if one is born with good mental endowment his brain will
“synthesize” to a degree and in a manner impossible to one whose mental endowment is
poor.”
Thus, conceptualizing a suitable intelligence test as one that “should test the fundamental
analytic-synthetic activity of the cognitive mind, the power to "grasp relations", or the "capacity
for abstraction" under appropriate circumstances, together with and through the medium of such
other cognitive mental activities as are natural to the social and cultural environment of those for
whom the test is being framed”.
A decision was made to make the tests as comprehensive as possible by the introduction of
appropriate abilities while recognizing the supreme importance of a "g" ability (general ability;
Spearman, 1927) in an intelligence test.
Use of Performance Tests for Intelligence Testing
There was at the time, a lot of criticism with regards to the use of performance tests due to their
failure to evaluate successfully, or at least as successfully as it is desirable, or as an intelligence
test ought to evaluate.
For instance, Cattel, (1936), in his book “Guide to Mental Testing” said:
“Unfortunately, the great majority of performance tests have quite low and even
negligible correlations with intelligence. So great is the attraction of performance tests,
however, alike to the subject and examiner (for even the psychologist is not immune to
the sense of in¬ creased prestige which important looking apparatus gives him) that
performance tests are widely used and depended upon, frequently in situations when, in
feet, they are misleading and a waste of time.”
And, further, in the 1948 edition of the same book, that:
“The only homage current practice pays to research findings is that one shall not calculate
intelligence quotients from performance tests - since the briefest experience shows that
such intelligence quotients are anything but constant. The score is therefore left as a
mental age. In the case of those few performance tests that are highly valid tests of
intelligence, no attention should be paid to this convention, and the usual I.Q. is best
used.”
Bhatia countered this argument by asserting that it is not performance tests as such which here at
fault, as the particular performance tests which had often been, so far, put forward, were for
different age-groups. If a performance test of intelligence is applied to an age-group where it is
obviously not suitable (usually performance tests are too easy for the higher age ranges), it
certainly cannot provide any valid assessment of intelligence, and believed that that was the case
with many of the earlier performance tests, such as those given by Pintner and Peterson (1917),
which cease to be effective after the age of 10 or 11 years, since the graphs of the median scores
for all these tests become stationary at that age having reached their ceiling. These tests,
therefore, could not be valid for the higher ages.
However, if the performance tests were devised such that they are appropriate in nature and
difficulty to the age-range for which they are meant, there seems to be no fundamental reason
why a performance test should not be as successful in measuring intelligence, as, say, the verbal
tests, except perhaps for the very highest mental level, he added.
Furthermore, in the construction of Performance Tests of intelligence, especially for the higher
age ranges, there is need for greater attention being paid to suggestions such as that put forward
by Termen (1923) in his introduction to Kohs' book on Intelligence Measurements
"In the upper ranges of intelligence especially, most performance tests have but little
differentiating value, simply because they do not draw heavily enough upon the higher
mental processes."
Adding on, the Performance Tests are often in danger of degenerating into tests for assessing the
manipulative capacity (in a physical sense) of the child in connection with miscellaneous
concrete material. Many such tests requiring no rational approach of the mind for their successful
solution easily assume the nature of a puzzle depending upon chance for the solution.
Finally, in the construction of Performance tests there is perhaps need for attention being paid to
the general suitability and appeal of the Test. A test such as a Picture Completion Test in which
the correct missing pieces have to be put in, in the cut¬ out holes in the picture, becomes
perplexing and uninviting to the subject when the size of the cut-outs become so small that
details are difficult to put in. Such a test, although suitable otherwise, does not enable the subject
to exercise his powers of intelligence suitably.
Principles Guiding Selection and Construction of Test
In conclusion, the principles guiding the selection and construction of tests constituting the
battery are:
1. Powers of analysis and synthesis of the subject have been sought to be drawn out as
heavily as possible.
2. The above has necessitated the collection together of a limited number of graded tests
rather than a miscellaneous collection of a large number of tests.
3. Within the above limitation, the tests have been presented in as varied forms as possible.
Actual concrete materials requiring discrimination of patterns (Kohs' Test), and requiring
discrimination in their movement (Passalong Test); as also a Picture Construction Test
(devised by the author) requiring discrimination in terms of pictures, and a Pattern
Drawing Test, involving the use of lines (devised by the author) have thus been included.
4. Since a broad basis for the assessment of intelligence under the given circumstances was
aimed at, an Immediate Memory Test (suitably adapted by the author for Illiterates) was
also included, as this was possible to administer.
5. General suitability of the tests needs to be carefully kept in view.

The Selection and Construction of Sub-Tests and their Description

The selection and construction of the sub-tests started in 1942 with Kohs’ Block Design Test:
a. Using all the seventeen designs proposed by Kohs, Bhatia applied the test to children of local
schools, where it appeared to work quite satisfactorily.
b. The children felt interested and attempted the tests with interest, however, not all of them
were equally successful at their solution, and even a cursory scrutiny of the results of a
number of children of known intelligence made it evident that the test gave good
discrimination, especially for children of ages 11 years and above.
c. The test appeared well worth giving a detailed and scientific trial for the purpose of
standardization on Indian children. The performance on the test also seemed to be correlated
positively with chronological age in the age range 11 to 16 years.
d. No formal paper was worked out at this stage on these results.
e. This led to the suggestion of trying out the test on village boys, particularly illiterate, of the
same age range because it appeared there was nothing in the test itself which precluded its
application to such children.
f. Only the problem and difficulty of the actual administration of the test to these children had
to be resolved. For it was known that the problems of testing under village conditions were
peculiar and often difficult to handle:
 i. A stranger would not obtain that response and ready co-operation from the village,
which alone would make any testing valid.
ii. The tester would also have to shed off many of his modern notions about the
formalities of test administration if he was to be successful at his task
iii. For example, a desk and stools for the use of the tester and the subject would usually
be difficult to obtain, and would not be very desirable in any case.
iv. The tester should find it convenient to sit with the subject on the ground in the
familiar Indian style, with his test materials arranged accordingly on a big flat
wooden board. It might often be desirable to administer the tests in the open, under
the shade of a tree, rather than inside a closed room, at least to start with, to alley the
curiosity of the common villager and to obtain his understanding and co-operation.
Of course, many of these conditions are fast changing; and was hoped it may disappear
altogether in due course of time, but at the time when the work started, they were a serious
consideration, and they often were getting mixed up with the political turmoil of the time.
g. The author received spontaneous co-operation those days from many of his students of the
Experimental Psychology class of Government Training College, Allahabad, who were
themselves residents of remote Indian villages. One of 24 of them, Mr. Karri Surat Lall,
applied the Kohs Block Design test under the guidance of the author, in his own and
neighboring villages and a report of this work was published in a local Journal of Education.
h. The success of this attempt under typical Indian village conditions convinced the author that
a Battery of Performance Tests for testing Intelligence could be successfully assembled
which would give valid results under Indian conditions.
Having received support in this from Professor Godfrey Thomson of the University of
Edinburgh, the author completed the assembly of the present Battery of Tests, after numerous
initial try-outs, by 1945, The tests finally included were:
1. Kohs' Block Design Test
o Kohs' Block Design Test although, to commence with, all the seventeen designs as
originally given by Kohs were tried, it was found that they took more time in their
administration than could be assigned to a single test of the Battery. Also, it was found
that this length of time was not justified on the basis of a greater validity of its results.
For, although the Kohs' test was good in itself, it needed to be combined with other tests
to give a satisfactory estimate of the subject's intelligence. Only ten designs were
therefore used, as suggested by Collins and Drever, (1946).
o The designs included in Bhatia’s test are Kohs' original designs numbers 1, 2, 4, 5, 7 10,
11, 15, 16 end 17, differing from Collins' and Drever's in the case of the last three.
o It was found to be necessary to include Kohs' design number 17, in order to obtain a
higher ceiling.
o The time limit for the first five designs was 2 minutes and for the last five designs, 5
minutes.
2. Alexander's Passalong Test
o Alexander's Passalong Test practically as given by him. Instead of his 9 subtests,
however, we used only 8, omitting his third sub-test. The time limit for the first four sub-
tests was 2 minutes and for the last four, 3 minutes.
3. Pattern Drawing Test (devised by the author)
o This is a test devised by the author specially for this Battery.
o The idea of this test grew out of the observation of some of the common pastimes which
boys in India, both in the urban as well as rural areas, are often engaged in. One such is to
draw different shapes on the ground or sand with the help of a stick, or on paper with the
help of a pencil.
o Often the drawing is done under various imposed conditions. Indeed, quite a number of
the less "active" or "indoor" games depend upon drawing of diagrams of this type.
o The condition of drawing which they utilized for the framing of the test was that "When
once you have started drawing, your pencil should not be lifted and no line should be
repeated".
o This condition was taken up with zest and almost a "challenge" by the subjects. There
never arose an occasion when the condition was not understood by the subjects or when
there was a lack of motivation. It was clearly understood by the subjects that it was not
the fineness of the drawing which was in demand as the successful completion of the
diagrams under the given conditions.
o We finally decided upon 8 diagrams which form the 8 sub-tests of this test.
1. The first sub-test was just a plain square figure;
2. the second was a square with one of the diagonals inserted;
3. the third was the second figure with a square of the size of the first super-added to it
on the left, so that the whole could also be viewed as a rectangle of length: breadth as
2: 1, divided into two squares, a diagonal in one of which having been drawn, and
4. the fourth was just the same as the third, but with two diagonals (parallel to each
other) shown one in each square.
The fifth, sixth, seventh and eighth sub-tests became more and more complex.
5. The fifth was a bigger square containing four squares of the size of the first shown in
it with two parallel diagonals in two opposite squares drawn in;
6. the sixth was the same 8s the fifth with two smaller squares, with an appropriate
diagonal, super-added on one side, so that the whole became a rectangle 3 by 2, with
three of the diagonals shown.
7. The seventh was the same as the sixth but with four diagonals instead of three.
8. Finally, the eighth and the last one became a square, 3 by 3, ' with all the smaller
squares shown and five of the diagonals drawn. This eighth could easily be viewed as
an extension of the seventh, and obtained by extending its breadth.
o This test, although coming under the category of Performance Tests, is really very much
akin to some of the "pure" tests of "g" which have been couched in terms of lines and
shapes. It is nevertheless in a form which can be more interestingly administered than
some of the more formal “g” tests. It may also obviously contain a K- or spatial factor or
component.
o In the actual administration of the test, various interesting features are noted which may
have relation with the individual's imagery, K-factor or other abilities.
For example, to some the mere "Complexity" of the diagram (i.e., perhaps the number of
lines in it) is more perplexing than the analytic difficulty involved in discovering the
solution.
According to Bhatia, “some seem almost to "hit" the solution at first sight while others
arrive at it by a laborious and perhaps "atomistic" effort”.
o The time limit for the first four sub-tests is 2 minutes, and for the last four, 3 minutes, the
subject being free to make as many trials on paper as they would like within the time
limit.
4. Immediate Memory Test for Digits (with an alternative form suitable for Illiterates,
devised by the author)
o Immediate Memory Test for Digits is practically the same as used in Binet tests and as
given by Terman and Merrill in their New Revised Stanford-Binet Tests of Intelligence
(1937).
o It consists of two parts:
a. Immediate Memory Direct
b. Immediate Memory Reversed.
The test was kept in two separate parts. That is to say, we begin first with the Direct part
and exhaust it by reaching a number of digits the subject is no longer able to repeat
correctly. Then the Reversed part is taken up and similarly exhaust it by again reaching a
number of digits the subject is no longer able to reverse correctly.
o When this test was administered on the illiterate boys through this medium of digits in
the initial stages of our test construction, it was found that these boys had unusual
difficulty in repeating the digits or in reversing them.
• The reason of this difficulty of theirs was that these boys are unfamiliar with the
digits, and thus even if they know the numbers, they are certainly not familiar with
them in the way the school boys are.
• The sound of a number such as "seven" or "five" constitutes a unit perceptual
experience for a literate boy, it does not do so for an illiterate boy.
• Thus, the sound "seven", for example, almost ranks as a set of nonsense syllables for
him, and constitutes for him, for the purposes of understanding and memorization, as
many units as there are unit sounds in the word.
 • It became evident, therefore, that the names of digits would have to be replaced by
something more unitary in the experience of the illiterate boys to achieve any sort of
equivalence of the test for the literate and the illiterate groups.
o The sounds of the Hindi consonants were selected for this purpose. The Hindi consonants
(Hindi is the mother tongue of the people in Northern India) are the same as the Sanskrit
consonants, and, as is commonly known, are built upon the scientific principle of a
consonant for each unit sound, and the converse also, that of one sound only for each
consonant, so that the alphabet is also perfectly phonetic.
o These consonants are like क (kā), च (chā), ल (lā), प (pā), etc. The Immediate Memory
Test, both the Direct part as well as the Reversed, was therefore set to the Illiterates
through the medium of these unit sounds and not the digits, but otherwise with no other
alteration. The results in this way were found to be definitely more satisfactory.
5. Picture Construction Test (devised by the author)
o A Picture test was used because it constituted a medium entirely different from that
utilized by any of our other tests.
o This Picture Test was different from some of the Picture tests in use at the time:
1. Firstly, in that it was taken care of that no parts of a picture become so small that
details are difficult or impracticable to put in so that the subjects do not take to the
test in zest, and
2. Secondly, that the test draws upon the subject's capacity to grasp relations as heavily
ss possible.
o The method of preparation of the test, was to cut suitable pictures of ordinary size into a
number of parts - 2, 4, 6, etc. - according to the sub-test desired.
o The cut pieces were the materials of the successive sub-tests in an order of ascending
difficulty. These parts of a picture are presented to the subject who is simply given the
instruction to put the pieces together to form the picture of which these pieces are the
parts. No further Information is given.
o The pictures have been cut along horizontal and vertical straight lines, so that in the case
of every picture (and therefore of every sub-test) the parts are all rectangular in shapes,
but generally not squares.
o It is the cutting of a picture into parts, is also the selection of a suitable picture for a
particular sub-test, that required careful thought on our part. For the easier sub-tests, the
pictures had to contain bolder but fewer details, and human figures had to be prominent.
For the later and more difficult sub-tests, landscapes could be used with many, but still
not too many, details. The details are still mostly centered around some human activity or
figure.
o Also, the pictures had to be suitable for the Indian children and of fairly good aesthetic
taste. Of the five pictures made use of in this test, four were selected from those which
were already available in children's magazines and journals. The fifth picture, however,
which constitutes the fifth sub-test had to be specially drawn to put in details, etc., which
were desired.
o Further, the cutting up of the pictures had to be done in such a manner that essential
relationships could be discovered on the basis of intelligent understanding. For example,
for the easier sub-tests, parts of the human body were separated which obviously formed
one whole. In the difficult sub-tests, the landscape was divided in o similar manner, end it
was very interesting to watch in the actual performance of the subjects how the brighter
ones could get at the cues and the duller ones just missed them.
o Generally, one of the cut pieces of the picture was such as to form the central theme of
the picture.
o The pieces of a picture were presented to the subject in a pile (or in the case of the fifth
picture in two piles) so that this part containing the central theme (or in the case of the
fifth picture, two parts) was at top and was alone seen by the subject in the first instance.
The other parts of the picture in the pile were arranged in a set order so as to make the
presentation uniform from the point of view of difficulty and chance.
o The time limit was 2 minutes for the first three sub-tests and 3 minutes for the fourth and
the fifth.
Some General Considerations
1. In the construction of the battery of tests, although some of the tests had to be adapted to the
Indian environment, it was not desired to have only such tests which had not been used
previously anywhere else. It was desired to make use of at least some tests which had already
been used in other countries and therefore about which a considerable amount of knowledge
already existed.
2. This was not very difficult in this case as most of the performance tests can be universally
used.
3. It is for this reason that Kohs' and Passalong Tests were included in forms commonly used.
4. However, they had to introduce a completely new Picture Test, as an identical Picture test, in
any case, would not work under Indian conditions as in the West.
5. The Memory Test had to be modified for the Illiterate group.
6. The Pattern Test because it seemed to be suitable for Indian children, although such a form
has not been tried elsewhere. The advantage of having a mixed Battery, composed of old and
new tests, was that they could stand respective comparison with each other in a final analysis.
Scoring
The scoring standards for the various tests are as given under. Both time and success have been
taken into account. The scoring standards have been kept as uniform and simple as possible.
1. Kohs’ Block Design Test:
For the first five designs, and for each design,
• 2 marks for success within a minute
• 1 mark for success between 1 minute and 2 minutes
• 0 mark for a failure, or success after the time limit.

For designs sixth to tenth, and for each design,

• 3 marks for success within a minute
• 2 marks for success between 1 minute and 2 minutes (but excluding 2 minutes)
• 1 mark for success between 2 and 3 minutes
• 0 marks for a failure, or success after the time limit
Maximum possible score: 25
2. Alexander’s Passalong Test:
For the first four sub-tests, and for each sub-test,
• 2 marks for success within a minute
• 1 mark for success between 1 minute and 2 minutes
• 0 marks for a failure, or success after the time limit.

For the last four sub-tests, and for each sub-test,

• 3 marks for success within a minute
• 2 marks for success between 1 minute and 2 minutes (but excluding 2 minutes)
• 1 mark for success between 2 and 3 minutes
• 0 marks for a failure, or success after the time limit.
Maximum possible score: 20.
3. Pattern Drawing Test:
Exactly the same as for the Passalong Test.
For the first four sub-tests, and for each sub-test,
• 2 marks for success within a minute
• 1 mark for success between 1 minute and 2 minutes
• 0 marks for a failure, or success after the time limit.

For the last four sub-tests, and for each sub-test,

• 3 marks for success within a minute
• 2 marks for success between 1 minute and 2 minutes (but excluding 2 minutes)
• 1 mark for success between 2 and 3 minutes
• 0 marks for a failure, or success after the time limit.
 Maximum possible score: 20.
4. Immediate Memory Test:
a. Direct: One mark each for the number of digits or sounds in the maximum correct
reproduction.
Maximum possible score provided for: 9.
b. Reverse: One mark each for the number of digits or sounds in the maximum correct
reversed reproduction.
Maximum possible score provided for: 6.
5. Picture Construction Test:
Pictures 1 to 3, and for each of them,
• 2 marks for success within a minute
• 1 mark for success between 1 minute and 2 minutes
• 0 mark for a failure, or success after the time limit.

Pictures 4 and 5, and for each of them,

• 3 marks for success within a minute
• 2 marks for success between 1 minute and 2 minutes (but excluding 2 minutes)
• 1 mark for success between 2 and 3 minutes
• 0 mark for a failure, or success after the time limit.
For Pictures 4 and 5, however, credit in addition to that earned according to the above
schedule was to be given as under:
1. For Picture 4: 1 mark, provided that at least 6 of the 8 parts had been correctly put within
the time limit.
2. For Picture 5: 2 marks, provided that at least 9 of the 12 parts had been correctly put
together, arid 1 mark, provided that at least 6 of the 12 parts had been correctly put
together, both within the time limit.
Maximum possible score: 15.
Maximum possible score for the whole Battery: 95
The total time taken in the administration of the Battery to a single individual is roughly
something less than an hour (45 to 60 minutes).

The Testing
The testing early resolved itself into two separate categories, i.e., for the urban children, mostly
literate, and the other for the rural children, mostly illiterate. The testing of the urban children did
not present any special difficulties, while that for the rural children did. Testing being individual,
both took considerable time.
The testing spread itself from 1945 to 1949, i.e., roughly over a period of four years. Some 1,400
subjects were tested in all during this period, of which 1,154 have been utilized for the purpose
of standardization and analysis. Most of those excluded belonged to the earlier period of testing
and in many cases were incomplete in some respect or other. Of the total (n=1,154), 642 are
literates and 512 are illiterates. They are in the age range 11 to 16 years.
The testing was done by Bhatia himself and a select number of his students of the Experimental
Psychology class of Government Training College, Allahabad, who passed out as teachers. They
were all post-graduate students mostly with distinguished record both in the theory and practice
of Teaching and had, in addition, undergone a full year's course both in the theoretical as well as
experimental parts of Educational Psychology under the author.
In the experimental part they had administered the present Battery and other Intelligence tests to
a number of local school children under the writer's supervision, so that when they went out and
administered these tests independently, they were well familiar with all the points of the
technique of testing. They worked on an entirely voluntary basis, as no financial aid was
forthcoming from any external quarter at that time.
About 350 subjects were tested by the present writer and the rest by the other testers, about 200
falling to the lot of each, with literates and illiterates in equal proportion. The other testers tested
the boys of their own village or town and its neighborhood, (complete groups, as falling within
our category, being tested in each area as far as possible) while the present writer visited a
number of scattered villages and towns specially for this purpose with the help of local friends.
A. Standardization for Illiterate Group (Sampling):
The Illiterate Group Geographical Distribution:
The geographical distribution of the population on which standardization is based is as
follows (please see also map):
District (with symbol used to denote it) Number Per cent of the whole
Mathura Rural (W 1) 44 8.59%
Agra Rural (W 2) 23 4.49%
Dehra Dun Rural (N1) 41 8.01%
Pilibhit Rural (N2) 47 9.18%
Sitapur Rural (N3) 41 8.01%
Allahabad Urban (C1) 42 8.20%
Faizabad Rural (C2) 73 14.26%
Kanpur Rural (C3) 25 4.88%
Deoria Rural (E1) 60 11.72%
Sultanpur Rural (E2) 76 14.84%
Jaunpur Rural (E3) 40 7.61%
Total 512 100.00%
Fairly equal numbers from various districts lying in the different regions of Uttar Pradesh
were tested.
• W, standing for the Western region;
• N, for the Northern;
• C, for the Central; and
• E, for the Eastern.
In the northern districts was also included one - namely Dehra Dun, Rural (N 1) (present day
Uttarakhand) which lies in the ranges of the Himalayan mountains. (Also, the method of
securing the sample in a particular district was to select a village or a group of neighboring
villages and then to test the complete sample, as coming within our category, provided by
this group of villages. That is to say, all the illiterate boys between the ages of 11 to 16 years
were tested without an exception. It was hoped that in this way the sample would be fairly
representative).
All the areas were rural, except one, namely Allahabad Urban (C 1) 8.20%. This gives us a
proportion of rural to urban as 91.80: 8.20, or roughly 10: 1. This is also roughly the
proportion of illiterates as distributed between rural and urban areas.
Upon combining the regions:
Region (with symbol used to denote it) Number Percent of the whole
Western (W) 67 13.09%
Northern (N) 129 25.20%
Central (C) 140 27.34%
Eastern (E) 176 34.38%
Total 512 100%
This indicates a fairly even distribution as amongst the various regions, except that the
Western region might have provided some more subjects.
The Illiterate Group Occupational Distribution:
The distribution of the occupation of the families from which the subjects came is as shown
under:
Occupation Number Percent of the Whole
Farmers (including all farmhands) 320
Shopkeepers 40 62.50%
Artisans and Craftsmen 58 7.81%
Laborers (Hired) 29 11.33%
Domestic Servants 43 8.40%
Not recorded (including unemployed) 22 4.29%
Total 512 100.00%
It was seen that farmers constituted the major portion of the sample, followed by
shopkeepers, petty businessmen, artisans, seasonal labor, and domestic servants, which
correctly reflects the proportions of the occupations as found in the general population.
The Illiterate Group Community-Wise Distribution
The distribution community-wise, as recorded, is as:
Community Number Percent of those Recorded
Brahmins 62 14.62%
Kshatriyas 42 9.70%
Vaishyas 23 5.42%
Muslims 29 6.84%
Total Non-Backward Communities 155 36.56%
Backward Communities 269 63.44%
Total recorded 424 100%
The backward communities included the following: Ahire, Nais, Dobis, Barhai, Kumhars,
Kurmi, Chamar, Pasi and Bhengi. Some of them were more backward than others and the
lowest had often been called "Harijans" in those times, after the name given to them by
Mahatma Gandhi.
The proportion of these communities as represented in our sample is practically the same as
in the general illiterate population.
 Although the "backward" communities do constitute a larger proportion among illiterates,
illiteracy is not a feature of "backward" communities alone. About 40% of the illiterates
come from the Non-Backward communities; illiteracy, where it exists, is quite general and
not confined to any groups on any basis.
On the basis of the evidences stated above, it was concluded that the sample was fairly
representative of the total illiterate population on geographical, occupational, and communal
basis.
B. Standardization for Literate Group (Sampling):
The sampling for the literate group was primarily on the basis of representative schools and
geographical areas
The Literate Group Geographical Distribution:
The geographical distribution of the sample was:
Area Number Percent of the whole
Boys’ School, Allahabad 100 15.56%
Basic Middle School, 86 13.40%
Allahabad
Gorakpur, Urban 35 5.45%
Garhwal, Urban 46 7.17%
Jhansi, Urban 66 10.26%
Sultanpur, Rural 64 9.97%
Govt. Inter College, 125 19.47%
Allahabad
Bareilly, Urban 120 18.69%
Total 642 100.00%
The proportion of the urban to rural subjects in the sample is 90.03 to 9.97, or, roughly, 9 to
1, which was also the proportion between the urban and rural literates in the general
population.
The Literate Group Occupational Distribution:
The occupational distribution of the sample was:
Occupation Number Percent of the whole group
Higher Professions 135 21.03%
Middle Class Service 227 35.36%
Lower Class Service 47 7.32%
Business 83 12.93%
Agriculture 92 14.33%
Not Recorded 58 9.03%
Total 642 100.00%
• The Higher Professions included Lawyers, Doctors, Engineers, Teachers and High
Government servants and their income would be above Rs 200 per month.
• The Middle-Class Service included clerks and other office workers in Government and
other concerns, Railway employees, Mechanics and other employees in Engineering
 firms, and employees of other public bodies such as the Municipalities, etc. Their range
of income is roughly between Rs 100 to Rs 200 per month.
• The Business group includes all who carry on business on their own. In ordinary sized
Indian cities, they form a part of the middle class.
• In the agriculture group are included those who live upon their income from the land
(landlords). It also includes farmers of the countryside.
• In the "Not-Recorded" group were orphans and those whose parents were unemployed.
The percentages show that the sample represented a fair cross-section of the Literate
population.
The Literate Group Community-Wise Distribution:
The community-wise distribution of the sample was:
Community Number Percent of the whole
Brahmins 159 21.65%
Kshatriyas (including 62 9.66%
Khatris)
Kayasthas 122 19.00%
Vaishyas 42 6.54%
Muslims 120 18.69%
Christians and Anglo- 65 10.12%
Indians
Backward Communities 31 4.83%
Others 61 9.50%
Total 642 100.00%
“Others” included those whose communities could not be ascertained, as also a few
miscellaneous such as Parsi, Sikh and Jain.
The percentages indicate that the sample is fairly representative of the general Literate
population on the basis of communities.
On the basis of the evidences stated above, it was concluded that the sample was fairly
representative of the total illiterate population on geographical, occupational, and communal
basis.

Factorial Analysis of Data

Sampling
For the purpose of factorial analysis of the Battery of tests, a group of 100 Literate subjects
(students of Boys' School, Allahabad) were also given the Terman-Merrill New Revised
Stanford-Binet Tests of Intelligence (Form L). This was considered to provide a very sound
background for the evaluation and analysis of the Battery and would be much better than
dependence upon teacher's opinion alone, although the latter has also been examined. The Binet
tests were given the same day as the Battery, usually after it, after giving the subject a rest pause
of a quarter to half an hour in between. One boy was tested every day with the Battery and the
Binet tests, usually in the mornings.
The possibility of using the Stanford-Binet Tests arises from the circumstance that in most big
Indian cities there are what were known as "European schools". The medium of instruction in
these schools is English right from the infant stage and they follow the same curriculum as
schools in Britain. The boys in these schools sit, or at least used to sit until recently, for the
Cambridge examinations, and had their papers etc. all set in Britain.
The Distributions of Data Collected
A. IQ:
IQ Class Interval Frequency
145- 1
140-144 4
135-139 2
130-134 3
125-129 6
120-124 6
115-119 6
110-114 7
105-109 13
100-104 14
95-99 19
90-94 11
85-89 5
80-84 1
75-79 1
70-74 0
Below 70 1
Total 100
Mean I.Q. = 107.0
S.D. = 15-7
The group was therefore a well representative "but rather above the average”.
B. Raw Scores of Sub-Tests:
a. Kohs’ Block Design Test:
Scores Frequency
16- 8
14-15 13
12-13 15
10-11 15
8-9 17
6-7 17
4-5 8
2-3 6
0-1 1
Total 100
b. Alexander’s Passalong Test:
Scores Frequency
16- 3
14-15 15
12-13 8
10-11 22
8-9 29
6-7 18
4-5 5
2-3 0
Total 100
c. Pattern Drawing Test:
Scores Frequency
16- 2
14-15 7
12-13 15
10-11 33
8-9 22
6-7 17
4-5 3
2-3 1
Total 100
d. Immediate Memory Test:
Scores Frequency
15 5
14 9
13 18
12 28
11 20
10 15
9 3
8 2
Total 100
e. Picture Construction Test:
Scores Frequency
15 5
14 12
13 11
12 17
11 20
10 8
9 13
8 8
7 4
6 2
Total 100
Correlations
It was observed that all of the distributions (as stated above) were normal distributions, thus
making the calculation of the inter correlation between the scores of the different tests valid.
Pearson’s Product-moment Coefficient of Correlation by Grouping and Diagonal Adding was
computed in each case. Appropriate checks were applied at all stages of computation, and, here
particularly that of diagonal adding in the other direction (Thomson, G. H., 1947) which was:
𝐶 + 𝐶 = 2(𝐴 + 𝐵)
The calculations were carried out retaining five places of decimals throughout but only 4 places,
and sometimes, only 3 places were reported.
As the group consisted of boys of differing chronological ages, and as accurate chronological
ages of this group were available, it was possible to find out the effect of age on the
intercorrelations of the tests. The intercorrelations between all the sub-tests and the chronological
age were therefore calculated by the method of Grouping and Diagonal Adding.
The Groups’ Chronological Age Distribution:
Chronological Age in Years and Months Frequency
17.6 8
17.0 6
16.6 9
16.0 7
15.6 8
15.0 8
14.6 12
14.0 7
13.6 8
13.0 9
12.6 4
12.0 2
11.6 3
11.0 4
10.6 5
Total 100

The Full correlations between chronological age (called, age), Binet M.A., Kohs' Block Design
Test, Alexanders Passalong Test, Pattern Drawing Test, Immediate Memory Test and Picture
Construction Tests, hereinafter referred to as: a, 1, 2, 3, 4, 5, 6 respectively are given in Table
below. The standard error of the correlation coefficient, r = 0, is 0.10.
From the above, partial correlations for constant age were calculated by the formula (Garrett, H.
E., 1937):
𝑟12 − 𝑟1𝑎 𝑟2𝑎
𝑟12𝑎 =
2 2
√1 − 𝑟1𝑎 √1 − 𝑟2𝑎

This was done by one pivotal condensation with age in the top row and first column, followed by
normalization (Thomson, G.H., 1940). The partial correlations for constant age between the
variates 1 (Binet M.A.), 2 (Kohs’), 3 (Passalong), 4 (Patterns), 5 (Memory), and 6 (Pictures), are
given in Table:
Factor Analysis
This correlation-matrix was analyzed factorially into orthogonal factors by Thurstone's
"Centroid" method" with guessed communalities (Thurstone, L. L., 1947; Thomson, G. H.,
1950). The communalities inserted initially for each test were the highest correlations in each
column or row.
Factor 1:
The first factor loadings were obtained by the usual method.
Test Loadings for Factor 1
Binet 1 .6802
Kohs 2 .6792
Passalong 3 .5961
Patterns 4 .6263
Memory 5 .3969
Pictures 6 .5326
The first residual matrix was then obtained, and the residuals were tested for significance, by
1
McNemar's formula: factors may be taken out until the quantity 𝜎1 reaches or falls below
√𝑁
(where N is the size of the sample), the quantity 𝜎1 being defined as:
𝜎𝑠
𝜎1 =
1 − 𝑀ℎ2

Where 𝜎𝑠 = 𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑒𝑠𝑖𝑑𝑢𝑎𝑙𝑠 𝑎𝑓𝑡𝑒𝑟 𝑆 𝑓𝑎𝑐𝑡𝑜𝑟𝑠

𝑀ℎ2 = 𝑀𝑒𝑎𝑛 𝐶𝑜𝑚𝑚𝑢𝑛𝑎𝑙𝑖𝑡𝑦 𝑓𝑜𝑟 𝑆 𝑓𝑎𝑐𝑡𝑜𝑟𝑠
In the case of this test, 𝜎𝑠 = .0701
𝑁 = 100
And 𝑀ℎ2 = .3525
𝜎 .0701 1
So that, 1−𝑀𝑠 = = .1083 >
ℎ2 .6474 √𝑁

Thus, a second factor could be extracted, but no further.

Factor 2:
The signs of the first and fifth rows and columns of the residual matrix, and again put in guessed
communalities, namely the highest residuals irrespective of sign in each column or row. The
second factor loadings were then obtained, and no further extraction of factors was done. The
loadings for Factor I and Factor II (with proper signs inserted), together with communalities
were:
Loadings Factor I Factor II h2 (communality)
Test
Binet 1 .6802 -.2563 .5284
Kohs 2 .6792 .1904 .4976
Passalong 3 .5961 .1781 .3871
Patterns 4 .6283 .2607 .4628
Memory 5 .3969 .4349 .3467
Pictures 6 .5326 +.1377 .3026

Iterations
A comparison of the commonalities for each test made it evident that the analysis was not yet
stable and that the guessed communalities had been far from the true values in many cases.
Iterations therefore had to be carried out to approximate to the true values of the communalities
as far as possible.
This was especially necessary because the number of tests included in the analysis was small,
and wrong communalities were therefore liable to make appreciable differences in the loadings.
The usual practice in iteration is to use that value of the communality (for a test) which has
resulted from the previous iteration, and thus to continue the iterations till the values do not show
further marked change, thus indicating a near approximation to the true values.
This often entails numerous iterations. In order to reduce the number of iterations, if possible, the
device*of putting in a value for the communality which a comparison of the value as put in an
iteration, and as it was obtained from it, suggested to be a near approximation to the true value
was followed.
In this way, by means of three iterations (including the first original one) approximate true values
were reached. However, a last fourth iteration was carried out using the values as obtained from
the third iteration, in order to get the final factor loadings of the tests. The results were:
Iterations Communality 1 Communality 2 Communality 3 Communality 4
Tests Put in Obtained Put in Obtained Put in Obtained Put in Obtained
Binet 1 .4685 .5284 .60 .6089 .60 .6206 .6206 .6311
Kohs 2 .4987 .4976 .50 .5179 .51 .5004 .5004 .4970
Passalong 3 .4046 .3871 .36 .3964 .40 .3880 .3880 .3836
Patterns 4 .4987 .4628 .40 .4425 .42 .4195 .4195 .4211
Memory 5 .3926 .3467 .30 .2636 .24 .2606 .2606 .2711
Pictures 6 .3682 .3026 .20 .2707 .28 .2771 .2771 .2770
The centroid factors are only mathematical in nature and have no clear psychological reference
until rotated into suitable positions. The factors were plotted and rotated so as to make the nature
of the factors clear.

Standardization of the Battery

The Battery was standardized for three separate groups, namely:
A. For the Literate Group based on the performance of the Literate group alone,
B. For the Illiterate Group based on the performance of the Illiterate Group alone, and
C. For the Literate group based on the performance of the Literate but to give the best estimate
of the general factor I1, as obtained in the analysis in the previous chapter.
Recording of Age
Since the age-group for this test is from 11 to 16 years, there is quite a wide variation of age
which had to be taken into account in the standardization. Moreover, the recording of the correct
age has been a troublesome matter in all psychological and educational investigations in Indie.
This stands true both for the Literates as well as the Illiterates. In the case of the Literate children
the trouble arises because the ages recorded in the school registers officially are often not the
correct ages, parents very often get the ages of their children recorded one or even two years (or
more) different from what these really are. The correct ages in such cases are generally in excess
of the officially recorded ages, however, there is no uniformity in this matter.
Indeed, in quite a number of cases the official recorded age may be the correct age, the parent not
having got a fictitious age recorded in the school. It was suspected since there is a vigorous
movement in various quarters to have these records maintained as accurately as possible, this
difficulty was likely to disappear at a later date.
At that time, however, it was present, and the investigations conducted till that time had been
made on various assumptions. One of them was to base the standardization on the officially
recorded ages as such. Now, although such a procedure may be justified from the point of view
of the group as a whole it was likely to do unusual injustice and create great error in individual
cases.
This situation was handled in two ways:
1. Firstly, the records of age were based correct to the nearest year only (not months).
2. Secondly, there were two columns for the record of age in the Record forms for recording the
details of the subject and his performance on the Battery:
i. Age from official records, and
ii. Estimated age of the subject.
The testers were instructed to record in the second entry for age whenever they felt any obvious
discrepancy between the official age and the age the subject could possibly have. For example, if
the official age of a subject reading in Class IX (i.e., one standard below the Matriculation) was
recorded as 12 years which is at least 3 years below what such a subject would usually have, an
attempt was made to investigate his age further which was often not difficult as the boy himself
would give an indication of it on crossexamination and also by his general appearance. Gross
inaccuracies of age were thus (hopefully), eliminated; and the standardization based on the
nearest year, makes the age allowance, it is hoped, fairly equitable
The difficulty in the case of Illiterates arises from their general lack of a time sense. No accurate
records are usually available. The age correct to the nearest year can however be easily
ascertained and will not be far wrong, especially if the tester has some contact with the village,
without which testing itself would not be feasible. So, the problem of age in the case of
Illiterates, was assumed, can be fairly well met.
Thus, they have worked with ages to the nearest year in the case of both Literates and Illiterates.
In the final tables, however, I.Q. scores corresponding to the intervening half years of age also by
means of interpolation. At this point, it was hoped that as more and more accurate record of age
may be possible in the future, these additional figures may prove useful. Of course, ultimately,
the revision of norms will be necessary when accurate records of age become fully available.
Method of Standardization
The standardization has been done on the basis of the percentile ranks of individuals within their
age groups, but the final results have been expressed in terms of ratios called as I.Q.s. This, we
have done to conform to the usual practice of expressing mental ratios in terms of I.Q.s, or at
least calling them so.
Mental Age
𝐼. 𝑄. = × 100
Chronological Age
The I.Q.s obtained according to the procedure were not exactly the same as the ratio of Mental
Age to Chronological Age, but are very nearly comparable with them. An important point about
the I.Q. tables given for the Literate and Illiterate groups must be carefully noted. The
standardization for these groups was done separately on the basis of the performance of their
own groups alone.
This was done as it was found that the performance of the two groups on the Battery were
significantly “different from each other", i.e., the same I.Q. (for example, 100) does not denote
the same level of performance with respect to the Battery of tests in both cases, the performance
of the Illiterate group being consistently lower.
The standardization of the Battery for the Illiterate group on the basis of its own results was,
however, later found to be further justified when we compared our I.Q.’s with the general
opinion about the intelligence of these boys held in the village.
The actual method of standardization followed is one given by Thomson, (1950), because of a
large variation of age range and of the uncertainty of the nature of growth curve. The method
thus, was to calculate the 5th, 16th, 50th, 84th and 95th percentile scores for the distributions of
each age group, 11 years, 12 years, 13 years, 14 years, 15 years and 16 years, separately, and
then to plot these percentile scores and obtain the various percentile lines.
Then the percentile lines were examined to see as to what type of curve should be fitted to them.
In the case of the Illiterate group, we found it best to fit straight lines throughout for all
percentile lines, but in the case of the Literate group we fitted straight lines from 11 to 15 years,
and thereafter to continue straight line by means of a curve which was tangential to it but had a
tendency to become horizontal.
From these lines of best fit, the correspondence between raw scores and I.Q.s was established in
every case by means of the familiar procedure of equating the 5th percentile score to 75 I.Q.,
16th percentile score to 85 I.Q., 50th percentile score to 100 I.Q., 84th percentile score to 115
I.Q., and 95th percentile score to 125 I.Q., thus working on a standard deviation of 15 I.Q.
points.
From these values, further sub-divisions of I.Q. values were obtained by the method of
interpolation end extrapolation. The tables of norms in each case were calculated to one decimal
point of the raw scores, and from these tables of norms, conversion tables were obtained from
which I.Q.s for 8ny given raw score for any age can be read off directly. The conversion tables
are in terms of whole numbers only and are meant to be put into the hands of those who make
use of the Buttery for the purpose of evaluating a child's I.Q.
Standardization Data
The standardization data was presented under separate categories:
a. For the Literate Group (with unweighted raw scores
b. For the Illiterate Group (with unweighted raw scores)

c. For the Literate Group (with weighted raw scores)

Reliability and Validity of the Battery
The problem of reliability and particularly validity, is a difficult one for tests of a new type, and
specially for groups amongst whom no tests have been used before. The difficulty in regard to
reliability had arisen due to the type of tests which had been used. In such tests, the repetition of
the scale on the same group after an interval of time is the only practicable method of
establishing reliability (Alexander, 1946), although the method is considered to be unsatisfactory
by many; for example, Kuder and Richardson, (1937) said:
"The retest coefficient on the same form gives, in general, estimates that are too high, because of
material remembered on the second application of the test. This memory factor cannot be
eliminated by increasing the length of the time between two applications because of variable
growth in the function tested within the population of individuals. These difficulties are so
serious that the method is rarely used.”
The difficulty in validation lies in securing a suitable criterion against which to validate the test.
As is usually recognized, according to Mcleish, (1950), validation often involves argument in a
circle; besides, of course, the difficulty of securing a suitable external criterion remains. Various
investigators have therefore sought to employ various devices to establish the validity of their
tests. Mcleish sought to employ the device of factorial analysis for this purpose.
Wechsler, (1944) author of Bellevue Intelligence Scales, had justified his Scales on the basis of
the correlations of the Scales with a number of such criteria, such as teachers' rating,
Psychiatrist's recommendations, Binet tests, etc., but had not demanded a high degree of
correlation with them. He made two assumptions in this connection. One, that:
"The Bellevue Scales were devised because of the belief that the Binet Scales were not
sufficiently "good" measures of intelligence for adults. Otherwise, indeed, we should not have
gone into the trouble of devising our tests."
And, two, that:
"No new test can be markedly out of line with established measures of intelligence and still
claim to be "good" measures of it, because that would be tantamount to saying that all other tests
were not reliable measures of it. But the degree to which any new test correlates with established
tests (e.g., the Binet) cannot in and of itself be accepted as a basic 40 proof of the new test's
validity."
Thus, it would appear that the validity of a test would be established only after its use for some
time by others besides the author himself, and on the basis of the accumulated opinion of these
people as to how "well" the test works. At that time, all that the constructor of the test could
therefore have done was accumulate all the evidence, both within the test as well as in relation to
extremal criteria, which has bearing on this problem, and to establish that all such evidence leads
one to expect that the test will fulfil its role as well as any of its kind can be expected to do,
which was the approach with this test.
Reliability
Data about the Retest reliability of the Battery could not be presented since there was not a
sufficient number of retest cases recorded at suitable intervals of time, i.e., not immediately after
(i.e., a day or two after) the original, for various practical difficulties. However, data on the basis
of the split-half method of testing reliability was presented, with the full realization that this too
had limitations in the case of the battery.
To obtain two equivalent parts of the "test", the scores on the odd and even sub-tests within each
test of the Battery were equated. Since the tests were graded in point of difficulty, the odd items
were necessarily easier than the even ones. To balance this, the scores on the odd sub-tests of the
first ' three tests (i.e., Kohs’ Block Design Test, Alexanders’ Passalong Test and Pattern Drawing
Test) were combined with the even sub-tests of the last two (i.e., Immediate Memory Test:
Direct and Reverse, and Picture Construction Test;). In the case of both the Literate as well as
Illiterate, a number of random scripts were selected for this purpose - every fifth subject being
taken up for this purpose.
Literate Group: Pearson correlation coefficient r = 0.851.
Illiterate Group: Pearson correlation coefficient r = 0.841.
Validity
A. For Literates
The test was validated against teachers’ opinions obtained on a five-point scale of:
1. Very Superior,
2. Superior,
3. Average,
4. Inferior, and
5. Very Inferior,
The corresponding I.Q. intervals being Above 130, 110-189, 90-109, 70-89 and Below 70.
The data was available for 423 cases and gave a Pearson Coefficient of Correlation:
r = +0.705
B. For Illiterates
The test was validated against the general opinion held in the village about the intelligence of
the boy. This was obtained on a five-point scale of:
1. Very Superior (i.e., Very Bright),
2. Superior (i.e., Bright),
3. Average (i.e., Ordinary),
4. Inferior (i.e., Dull), and
5. Very Inferior (i.e., Very Dull)
It was not difficult to obtain such evaluation on the part of the village-folk. The data was
available for 502 cases and gave a Pearson Coefficient of Correlation: r = +0.717
Use of Differential Scoring Methodology
Barnabas and Rao (1994), conducted a study to compare the relative efficacy of these two short
form approaches in their ability to predict the Full-Scale Intelligence quotient (FSIQ) in
psychiatric patients, since in certain clinical situations the use of a short scale may be necessary
particularly for gross screening purposes even though under normal circumstances a shortened
scale is never as reliable as the original (Cronbach L. J., 1984).
In order to increase the cost effectiveness of the test in terms of time and effort, Murthy (1966)
(Murthy Short Scale: MSS) and Pershad et al. (1988) proposed the use of abbreviated forms of
the Bhatia Scale, and Verma et al., (1988) (Verma Short Scale: VSS) have utilized a scoring
procedure which is different from the method proposed by Bhatia.
On comparison, it was found that MSS performed favorably, when compared with VSS on all of
the comparisons:
1. The FSIQ has the lowest mean and S.D., VSS had highest mean and S.D. scores. and the
MSS fall in between the two scales.
2. A study of intercorrelations revealed that Pearson's correlation coefficient between FSIQ
and MSS was 0.79 (highest), between FSIQ and VSS it was 0.57 (lowest) and between
MSS and VSS the correlation was 0.72.
All the three correlations were highly significant (p<0.001, df=249).
3. In order to examine which of the two short scales predict the FSIQ best, a stepwise
multiple regression analysis was carried out which indicated that MSS was the best
predictor (p<0.001).
The regression coefficient for VSS was not significant.
4. The efficacy of these two methods in classifying patients was then estimated. The cut-off
points were: 90 as normal and 89 as subnormal based on the classification scheme as
outlined by Wechsler.
The MSS when considered along with FSIQ classified 224 subjects (89%) correctly. 27
subjects (10.8%) were misclassified between the two.
On the other hand, VSS with FSIQ together classified only 192 subjects (76%) correctly,
while 59 subjects (24%) were wrongly classified.
Hence, even in classification of subjects MSS proved to be more accurate in classifying
individuals correctly when compared with VSS.
One might conclude that MSS scoring procedure is a good approximation of the FSIQ and more
accurate in classifying individuals correctly as compared to the VSS.
It is to be kept in mind that for both forms of scoring procedures, there is no procedure of
conversion of raw scores, leading to problems of accurate prorating to obtain the I.Q. Hence, the
use of short scales to estimate IQ is to be kept to a minimum. If, however their use becomes
unavoidable for various reasons then the MSS method may be preferred to the VSS method.
References
Alexander, W. P. (1946): Instruction Book for a Performance Scale, Thomas Nelson, p. 6.
Barnabas, I. P., & Rao, S. (1994). Comparison of two Short Forms of the Bhatia's test of
Intelligence. NIMHANS Journal.
Bhatia, C. M. (1955). Performance tests of intelligence under Indian conditions.
Cattell, R. B. (1936): A Guide to Mental Testing, London, University of London Press, pp. 17
and 27.
Cattell, R. B. (1948): A Guide to Mental Testing, London, University of London Press, p. 34.
Collins and Drever: Performance Tests of Intelligence, Edinburgh, Oliver and Boyd, Third
Edition (1946), p. 23.
Cronbach L J, (1984) Essential of Psychological Testing (4th edn) New York, Haper & Row
Emmett, W. G. Evidence of a Space Factor at 11'' and Earlier. Brit. J. Psych. Stat. Sect., II, pp. 3-
16.
Garrett, H.E. (1937): Statistics in Education and Psychology, New York, Longmans, page 414.
Gonsalvez C, (1986) Observations on the use of the Bhatia's Short Scale (Unpublished notes,
NIMHANS)
Kohs, S.C. (1923). Intelligence Measurement. New York, Macmillan, Pp. 7-9.19
Kuder, G. F., end Richardson, M.W. (1937): The Theory of the Estimation of Test Reliability.
Psychometrika, 2, 151-160.
Mcleish, J. (1950): The Validation of Seashore's Measures of Musical Talent by Factorial
Methods. Brit. J. Psychol., Stat. Sec., Ill, 129-140.
Murthy, H. N. (1966). Short Scale of the Bhatia’s Performance Tests. Indian Psychological
Review, 2(2), 133-134.
Pershad D, Arunima Mahajan & Verma S K, (1988) Bhatia's short scale of intelligence - A
suggested modification for adults Personality Study & Group Behaviour Page: 8 (1): 43-48
Pintner and Peterson (1917): A Scale of Performance Tests. See the standardisation data, pp.100-
137.
Spearman, C. E. (1927). The abilities of man (Vol. 89). New York: Macmillan.
Terman, L. M. (1923): Preface to Kohs' Intelligence Measurement, p. vil.
Terman, L. M. and Merrill, M. A. (1937): Measuring Intelligence, London, Harrap
Thomson, G. H. (1940) An Analysis of Performance Test Scores of a Representative Group of
Scottish Children. London: University of London Press.
Thomson, G.H. (1947): How to Calculate Correlations, Revised Edition, London, Harrap, p.15.
Thomson, G. K. (1950): The Factorial Analysis of Human Ability, Fourth Edition, Chapters II
and X. London, University of London Press
Thurstone, L.L. (1947); Multiple-Factor Analysis, Chapter VIII. Chicago, University of Chicago
Press.
Verma S K, Pershad D, Malhotra A & Arunima M, (1988) The Revised Bhatia's Short Battery of
performance Tests of Intelligence for Adults - A Hand Book. National psychological
Corporation, Agra
Wechsler, D. (1944): The Measurement of Adult Intelligence. P. 129. Baltimore, Williams and
Wilkins.