Chronic cannabis users: Further analysis of objective test results

Sections

Abstract
Introduction
Procedure
Findings and discussion
DISTANCE ESTIMATION
TIME ESTIMATION10
DIGIT SPAN
BENDER GESTALT VISUAL MOTOR TEST
Summary and conclusion

Details

Author: M. I. SOUEIF
Pages: 1 to 26
Creation Date: 1975/01/01

Chronic cannabis users: Further analysis of objective test results

Ph.D. M. I. SOUEIF Professor of Psychology, Faculty of Arts, Cairo University, Egypt

Abstract

Twelve objective tests, generating 16 psychological test variables were administered to 850 chronic cannabis takers and 839 non-takers, all Egyptian males, ranging in age between 15 and 50 years, and representing various positions on continua of literacy-illiteracy and urbanism-ruralism. The tests were intended to assess different modalities of 5 broad psychological functions: speed of psychomotor performance, estimation of moderate distances, estimation of a few minutes time interval, immediate memory for digits and designs and visual-motor co-ordination. Analysis of variance based on 3 x 3 x 2 factorial design was carried out and revealed statistically significant differences between the two groups of subjects on most of the test variables. Detailed analyses showed that controls displayed better performance than drug users on the majority of the variables. Significant second order interaction effects emerged in a big number of cases denoting that, under certain conditions relating to literacy-illiteracy and urbanism-ruralism, the superiority of controls over cannabis users became much more impressive.

Our findings were discussed in the light of investigations which were conducted on the psychological correlates and/or effects of long term and short term cannabis consumption. The clinical psychological literature was cited where slowness of psychomotor performance, defective visual-motor co-ordination and impaired memory for designs were reported to be associated with psychiatric disorder. It is suggested that our results point to meaningful associations between chronic cannabis consumption and manifestations of psychopathology.

Introduction

In a series of publications we have reported on various stages of the study of chronic cannabis consumption Egyptian males (Committee 1960, 1964; Soueif, 1967, 1971, 1973). The study included, among other parts, the administration of 12 objective psychological tests generating 16 test variables to 850 cannabis takers and 839 non-takers. In one of our previous publications we gave a brief account of the tests, together with the results of a preliminary analysis of the test-scores obtained by the subjects (1971). This paper is intended to present rather a thorough analysis of our objective test findings and a discussion of their significance within the context of the clinical literature on the cited tests and selected information recently published on behavioural correlates of cannabis consumption. Occasionally we will refer to subjects' responses to a standardized interview in the hope that such responses may render their test performance more meaningful (Soueif 1967).

Procedure

A description of the takers and controls was given elsewhere (Soueif 1971). It should be noted that our group of users included practically all the male population convicted expressly for hashish use and were incarcerated in Egyptian prisons from June 1967 to March 1968. Their ages ranged from 15 to slightly over 50 years with an average of 39 years (± 10.5). 460 users were detained in prisons situated in big cities and intended for urban offenders, while 300 subjects were villagers kept in rural prisons. Sixty per cent were illiterate and the rest were distributed among various levels of education with only 6 subjects holding high school certificates. The majority of our users (89 per cent) had been taking hashish on regular basis for long periods before they were arrested, the periods ranging from slightly over 5 years to approximately 30 years. About two thirds of the whole group of users had been taking the drug more than once daily. The least amount of hashish reported to have been used by the single user per time was 0.21 gm; this amount was taken by no more than 3.1 per cent of ours subjects. On the other hand 3.3 per cent of the group went for something around 2.50 gm. The majority, however, (67.3 per cent) took an amount ranging between 0.54 and 1.08 gm. Smoking, mainly by means of a water-pipe, but sometimes in cigarettes, was the prevailing route of administration (by 89.4 per cent of the group). In 1971 we arranged for the analysis of seven samples of hashish seized on the illicit market, to be carried out at the Laboratories of the National Centre for Social and Criminological Research in Cairo, to determine percentage of Δ 9-THC in the stuff usually used by Egyptian takers. Estimation of Δ 9-THC ranged from 1.9 per cent to 3.6 per cent with an average of 3.04 per cent by weight.

Though there were rumours to the effect that hashish and other narcotic substances were smuggled into the prisons and clandestinely traded among inmates, to the best of our information our testees were not under the effects of any such substance when they were examined.

Controls were defined as those subjects who had never taken hashish nor any other narcotic substance as specified in Act 182 passed in June 1960 by the UAR Government as to be prohibition and/or regulation of the use of narcotic substances (Soueif 1967). They were selected, according to criterion, from among the inmates of the same prisons where our users were detained. A serious attempt was made at equating controls to cannabis takers regarding main parameters of age, literacy and residence (viz. position on a hypothetical continuum of urbanism-ruralism). We finally came up with a closely comparable control group with the following main characteristics: Age ranged from 15 to slightly over 50 years. Mean age was 33 years (± 9.75). 454 subjects were selected from urban prisons and 385 from rural ones. About 55 per cent were illiterate, and the rest were semi-literates and literates, with 44 subjects holding high school certificates and 9 university graduates. Thus our controls were almost equated to the takers regarding age range and variance. The two groups were also equated with respect to residence. But the average age among controls was slightly lower than takers. As to literacy controls were, also, slightly better than users. However, it should be noted that the two groups included an almost identical percentage of skilled labourers (about 26 per cent), and the takers had a higher average monthly income than the non-takers.

To ensure a genuinely co-operative attitude among our testees the following steps were undertaken:

  1. Detainees were invited in groups of 30 to 40 subjects to meetings held in one of the big halls of the prison. The aims of the study and the conditions under which it should be carried out were then explained in simple and straightforward words. The help they were expected to offer us, within this framework, was then stated.

  2. It was made clear that their co-operation was solicited on a completely voluntary basis.

  3. It was emphasized that the testing sessions would take place in complete privacy, where nobody except the examinee and the tester (a qualified psychologist) would be present. The information obtained, it was also pointed out, would remain absolutely confidential. 1

Findings and discussion

Analysis of variance based on 3 x 3 x 2 factorial design is the main statistical method utilized in this paper. [ 2] Table 1 presents the number of subjects filling in 18 cells (for each test variable) based on the adopted three way classification. It is conceivable that the numbers would vary slightly from test to test owing to a multiplicity of factors which were beyond the investigator's control. Only in the case of the Trail Making test were required to draw lines between digits enclosed in small circles scattered all over the test sheet, the number of testees was markedly reduced since illiterates had to be exempted from being tested.

Tables 2, 3 and 4 show mean scores and standard deviations obtained by takers and controls from urban, semi-rural and rural areas respectively. Those tables are provided to serve a number of objectives:

  1. To supply the reader with a clear idea as to the basic ingredients on which the analysis of variance is founded.

  2. To make possible detailed comparisons with data collected in the future on the same (or on correlated) test variables. Such comparisons, hopefully, would lead to meaningful and more basic generalizations (WHO technical report No. 516, p. 33).

  3. To make available data, usually needed (to be utilized as norms) by clinical psychologists, social phychologists, educationalists etc., working in the area of drug dependence or in other related fields.

TABLE 1

Number of subjects examined along 16 test variables

 

Cannabis takers

Non-takers

 
 

Urban

Semi-rural

Rural

Urban

Semi-rural

Rural

 

Objective tests

Illit.

Semi-lit.

Lit.

Illit.

Semi-lit.

Lit.

Illit.

Semi-lit.

Lit.

Illit.

Semi-lit.

Lit.

Illit.

Semi-lit.

Lit.

Illit.

Semi-lit.

Lit.

Total

1.Tool matching
274 129 43 165 93 22 55 30 8 240 107 98 141 65 70 70 14 10 1 634
2.H marking
242 123 41 149 75 18 56 29 14 223 93 91 135 54 60 73 16 25 1 517
3.Speed
244 125 41 142 65 19 57 30 8 212 93 95 125 57 63 72 24 11 1 483
4.Mark making
242 124 39 143 68 19 54 30 8 205 93 92 131 57 65 66 14 10 1 460
5.Trail making (Part A)
138 122 36 56 86 18 13 28 8 111 101 95 43 62 67 12 14 11 1 021
6.Initial reaction time
281 132 43 171 92 22 60 30 8 244 106 99 143 66 70 73 14 11 1 665
7.Distance overestimation
276 132 43 173 91 22 58 30 8 242 106 98 144 65 70 69 14 10 1 651
8.Distance underestimation
277 132 43 174 93 22 60 30 8 244 107 99 146 66 71 69 14 10 1 665
9.Distance estimation: discrepancy
277 132 43 174 91 22 58 30 8 242 106 98 145 65 70 69 14 9 1 653
10.Time estimation
278 132 43 173 92 20 59 30 8 243 106 99 143 65 69 72 12 11 1 655
11.Time estimation: discrepancy
259 121 40 158 84 18 46 19 7 219 95 79 103 61 64 54 11 8 1 446
12.Digits forward
278 132 42 175 93 22 60 30 8 242 107 99 143 66 69 73 14 11 1 664
13.Digits backward
279 132 42 176 93 22 60 30 8 243 107 99 138 66 67 73 14 11 1 660
14.Digits forward and backward .
278 132 43 176 93 22 60 30 8 243 107 99 137 66 67 73 14 11 1 659
15.Bender Gestalt: copy
264 130 41 161 92 22 56 30 8 231 106 100 144 66 71 72 14 11 1 619
16.Bender Gestalt: recall
249 129 41 157 91 22 54 30 8 221 105 99 139 66 71 70 14 11 1 577

TABLE 2

Means (M) and standard deviations (S.D.) earned by urban takers and controls on 16 test variables

 

Urban takers

Urban controls

 

Illiterate

Semi-literate

Literate

Illiterate

Semi-literate

Literate

Tests

 

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

1.
Tool matching
11.0 4.34 12.5 5.00 15.1 8.73 10.6 4.10 13.5 4.49 16.6 5.62
2.
H marking
12.7 4.77 14.7 4.99 14.4 4.79 13.8 4.76 15.9 4.34 19.6 4.10
3.
Speed
17.1 5.42 21.6 5.55 23.6 5.08 17.2 5.00 23.0 5.45 27.7 6.88
4.
Mark making
12.4 5.32 17.7 5.34 18.2 6.94 13.4 6.22 18.9 5.57 23.1 7.59
5.
Trail making (Part A)
103.5 62.30 91.5 49.47 86.2 36.84 93.5 53.45 89.3 48.69 65.9 41.00
6.
Initial reaction time
9.3 5.50 8.9 5.10 8.5 4.75 8.2 4.17 7.6 5.41 6.6 3.32
7.
Distance overestimation
8.2 12.34 12.0 14.83 5.9 9.83 8.3 15.48 7.5 13.96 5.7 8.88
8.
Distance underestimation
14.3 13.16 12.2 11.58 13.5 15.8 12.08 14.07 11.5 9.41 7.4 7.16
9.
Distance estimation: discrepancy
21.2 22.60 17.9 12.73 17.6 11.89 22.2 15.89 17.1 13.53 11.1 9.52
10.
Time estimation
5.4 3.39 4.7 2.91 4.9 4.57 6.2 4.22 5.5 3.34 2.72  
11.
Time estimation: discrepancy
2.2 3.00 1.8 2.88 1.2 1.67 1.8 3.97 2.3 2.90 1.4 2.31
12.
Digits forward
4.3 0.87 4.9 1.00 4.8 1.00 4.1 0.89 4.8 0.94 5.5 1.13
13.
Digits backward
2.5 1.20 2.9 0.92 3.3 0.90 2.6 1.07 3.1 1.05 3.5 1.13
14.
Digits forward and backward
7.2 1.65 8.1 1.98 8.1 1.57 6.7 1.52 7.7 1.70 9.0 1.67
15.
Bender Gestalt: copy
13.2 2.72 15.7 2.72 16.3 3.38 13.2 2.98 16.0 3.35 17.8 2.99
16.
Bender Gestalt: recall
6.1 3.16 8.5 4.17 10.0 4.66 6.6 2.49 8.6 3.54 11.7 3.87

TABLE 3

Means (M) and standard deviations (S.D.) earned by semi-rural takers and controls (from lower Egypt) on 16 test variables

 

Semi-rural takers

Semi-rural controls

Illiterate

Semi-literate

Literate

Illiterate

Semi-literate

Literate

Illiterate

Tests

 

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

1.
Tool matching
9.6 4.09 11.8 4.83 13.4 3.35 10.9 4.47 12.5 4.46 16.3 5.16
2.
H marking
11.7 4.27 15.2 4.51 16.9 4.14 12.9 4.23 16.2 3.91 21.6 5.35
3.
Speed
16.4 5.68 19.7 5.95 24.7 5.70 18.7 6.30 22.2 4.90 27.1 6.25
4.
Mark making
12.0 5.30 17.7 6.86 20.3 7.19 12.4 5.96 17.3 5.60 23.3 7.29
5.
Trail making (Part A)
108.6 56.57 89.8 52.92 77.5 23.00 123.5 55.35 83.7 39.67 68.3 36.65
6.
Initial reaction time
8.4 5.05 8.3 5.97 6.9 3.73 7.8 4.57 7.9 5.13 7.3 4.21
7.
Distance overestimation
7.3 10.84 8.2 15.50 7.7 11.67 7.2 13.13 5.6 6.78 5.0 6.32
8.
Distance underestimation
13.2 12.99 10.2 9.70 9.1 9.86 14.5 14.26 12.6 12.94 9.0 9.68
9.
Distance estimation: discrepancy
19.5 14.74 18.5 16.16 14.2 12.73 21.9 17.25 17.6 11.92 12.5 8.96
10.
Time estimation
6.0 4.56 4.9 2.82 5.9 4.31 5.4 4.42 5.5 3.10 5.8 3.50
11.
Time estimation: discrepancy
2.0 3.80 1.9 2.67 2.0 3.83 2.8 3.68 2.1 3.23 2.4 3.08
12.
Digits forward
4.1 0.92 4.6 1.05 4.8 0.93 4.4 0.93 4.8 1.12 5.3 0.93
13.
Digits backward
2.3 1.25 3.0 1.03 3.3 1.15 2.5 1.33 3.0 0.97 3.7 1.33
14.
Digits forward and backward .
6.4 1.73 7.5 1.73 8.2 1.77 6.3 1.72 7.8 1.74 8.3 2.15
15.
Bender Gestalt: copy
12.6 2.94 15.2 2.72 16.6 2.41 12.9 2.73 14.4 2.40 17.0 3.30
16.
Bender Gestalt: recall
6.1 2.99 7.9 3.42 8.6 4.34 5.9 3.05 8.7 3.62 10.4 3.33

TABLE 4

Means (M) and standard deviations (S.D.) earned by semi-rural takers and controls (from Upper Egypt) on 16 test variables

 

Rural takers

Rural controls

Illiterate

Semi-literate

Literate

Illiterate

Semi-literate

Literate

Illiterate

   

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

M

S.D.

1.
Tool matching
8.5 4.61 8.8 2.45 15.2 2.32 8.1 3.35 11.4 2.21 12.1 4.00
2.
H marking
11.9 4.75 12.9 4.64 16.5 3.13 11.3 4.34 13.6 1.66 15.0 3.37
3.
Speed
16.1 5.10 20.2 4.79 28.0 4.77 14.1 3.38 21.2 4.21 24.5 3.30
4.
Mark making
12.6 6.19 16.1 5.06 19.7 6.65 11.3 4.73 18.9 5.15 14.5 4.56
5.
Trail making (Part A)
128.9 55.46 114.3 56.03 87.3 25.54 108.3 47.07 112.0 50.37 76.3 32.39
6.
Initial reaction time
7.8 3.47 6.4 2.43 5.8 3.70 7.6 3.74 7.9 5.03 6.1 2.88
7.
Distance overestimation
10.1 18.25 10.9 9.47 12.3 14.63 6.2 10.07 5.5 5.97 12.7 13.54
8.
Distance underestimation
17.1 15.69 8.0 10.26 5.1 4.92 21.5 14.75 13.4 13.15 8.5 6.34
9.
Distance estimation: discrepancy
17.7 13.04 23.7 11.14 16.0 13.16 24.9 13.82 23.5 12.57 16.5 7.47
10.
Time estimation
7.1 5.90 4.3 2.59 4.0 2.45 6.3 4.74 5.83 5.0 3.10  
11.
Time estimation: discrepancy
3.8 4.35 1.9 2.14 1.5 1.31 3.9 3.78 3.7 4.50 2.2 1.91
12.
Digits forward
4.1 0.82 4.6 0.80 4.9 0.33 3.9 0.78 4.4 0.59 4.8 1.02
13.
Digits backward
2.0 1.27 2.6 1.03 3.1 0.60 2.1 1.20 3.2 0.57 3.5 0.54
14.
Digits forward and backward .
6.0 1.79 7.2 1.67 8.1 0.71 6.3 2.14 7.6 0.75 8.6 1.47
15.
Bender Gestalt: copy
12.8 2.94 14.9 3.16 15.8 2.53 12.6 2.63 15.5 3.76 17.4 3.12
16.
Bender Gestalt: recall
4.9 2.47 8.1 3.12 7.8 3.13 4.3 2.09 8.2 3.29 9.2 3.05

It should be noted that all our computations were based on raw scores. [ 3]

Table 5 presents the outcome of analysis of variance comparing takers with controls, irrespective of literacy and residence, on 16 test variables.

It will be readily seen that cannabis takers are significantly differentiated from non-takers on most of our tests. On six test variables the differentiation reached a very high level of confidence.

For the rest of this article we will be discussing at some length, the results obtained on each test variable.

TABLE 5

F ratios between, takers and controls irrespective of literacy and residence

Tests

 

F

P

1.
Tool matching
13.23 0.001
2.
H marking
76.26 0.001
3.
Speed
30.40 0.001
4.
Mark making
2.71
n.s.
5.
Trail making (Part A)
7.13 0.01
6.
Initial reaction time
2.36
n.s.
7.
Distance overestimation
4.75 0.05
8.
Distance underestimation a
5.54 0.05
9.
Distance estimation: discrepancy b
0.73
n.s.
10.
Time estimation c
0.56
n.s.
11.
Time estimation: discrepancy a, b
5.78 0.05
12.
Digits forward
3.70
n.s.
13.
Digits backward
27.70 0.001
14.
Digits forward and backward
3.04
n.s.
15.
Bender Gestalt: copy
11.51 0.001
16.
Bender Gestalt: recall
9.34 0.001

a Controls made more mistakes than users.

b Analysis of discrepancy irrespective of direction.

c For one technical reason we could not carry out two analysis, one for overestimation and one for underestimation.

(1) Tool Matching: Part A of the General Aptitude Test Battery (GATB)

The reader will remember that this test was constructed for the assessment of form perception (Soueif 1971). The testee is scored for both speed and accuracy of performance lumped together. Drug takers, compared with controls, proved to be very poor on this test. The importance of this, and similar results on other tests derived from GATB to be discussed presently, is bi-faceted. First, it implies that cannabis takers are consistently inferior to non-takers on a certain psychological function. [ 4] Secondly, one can make the prediction that, in all probability, chronic cannabis takers, as a whole, will prove inferior to non-takers when both are compared as regards their performance on certain jobs. The interested reader can find an ever increasing body of information on the validation of all the tests that comprise the GATB against a very large variety of occupations (Cronbach, 1960, p. 280; Anastasi, 1969, p. 345; Bemis, 1968). [ 5] A number of correlational studies on the knowledge as to the sort of behaviour one may expect from cannabis takers in a variety of test situations. For example correlations of 0.64 and 0.58 have been reported between Tool Matching on one hand and Name Comparison and Form Matching respectively on the other. [ 6] We have here the main components of an empirical evidence to the effect that the behaviour displayed by our subjects on the Tool Matching Test can be generalized. This means that it is not only on jobs requiring matching tools that cannabis takers will prove inferior to non-takers, but also on all jobs requiring comparison between visual percepts as such. [ 7]

There is a significant second order interaction effect between drug taking, literacy and residential area (F = 4; d.f. 4/1599; p 0.01). Hence, it is to be concluded that the discrepancy obtained between takers and non-takers is related to the subjects' level of literacy and the position of their residential area on the continuum of "urbanism-ruralism". Inspection of tables 2 to 4 may help clarification. For instance the differences between users and controls become rather marked with the literate and semi-literate groups in the urban and semi-rural areas (c.f. takers v. controls in semi-rural areas; literate takers v. controls in urban areas), compared to illiterates in rural districts.

(2) H Marking (Part C of GATB)

This test is designed to assess precision of aiming. Again speed of performance is emphasized in the instructions. Controls obtained better scores than users and the discrepancy is very highly significant. Judging by the size of F this test proved to be the best differentiator between the two groups in our battery. The reader is invited to think of the large number of skilled jobs where precision of aiming is part of their required profile of abilities. On such jobs the performance of cannabis takers would be poorer than the performance of non-takers would be poorer than the performance of non-takers. One can also think of jobs requiring other abilities which were shown to be significantly correlated with precision of aiming. It has been shown in the original standardization group of the GATB (N = 4000) that H Marking has a correlation of 0.50 with Tool Matching and with Name Comparison, plus a correlation of 0.47 with Form Matching. Obviously, therefore, the poor performance expected from cannabis takers on jobs requiring comparison between visual percepts is more emphasized by the sort of behaviour demonstrated on the H Marking Test.

Second order interaction effect came out to be highly significant (F = 5.2; d.f. 4/1482; p 0.001). This shows a certain degree of dependency between drug taking effect, literacy and urbanism-ruralism. Examining tables 2 to 4 the reader can detect a trend which amplifies this result. By comparing the corresponding cells in table 2 of takers v. controls the reader can detect a trend towards more discrepancy in more literate groups. The same trend can be traced in table 3. But in table 4, where the subjects are definitely rural, the trend is greatly blunted.

It should be noted, however, that the over-all difference between users and non-users on H Marking is so big that it can be shown to persist regardless of literacy or urbanism-ruralism. To demonstrate that, a small statistical experiment was carried out as follows: We may consider residential areas from which our subjects were drawn a random sample from a much bigger population (that could be defined as the whole country). In this case residential areas, as utilized in our investigation, turn into replications of the experimental design in which the main independent variables are drug taking and literacy. Under this condition we may take the second order interaction (drug X literacy X residence) as the error term (instead of the within group variance) against which to test the significance of the first order interactions. When we did this all the first order interactions proved non-significant. Now, it was quite legitimate to pool all the interactions together, use the pooled estimate of variance as an error term, against which to test the significance of drug-taking effect. This done we obtained an F equal to 17.56 which is very highly significant still (p almost = 0.001; d.f. 1/12) (A. Edwards, 1956, p. 247).

(3) Speed (Part G of GATB)

This test was designed to assess motor speed in a more or less pure form. The subject is required to put three dots inside each rectangle on the test sheet and to fill in as many rectangles as he could within 30 seconds. [ 8]

This test discriminates at a very high level of confidence between cannabis takers and controls (F = 30.40; p 0.001), the former are definitely slower than the latter. One can therefore conclude that in so far as the Speed Test has an occupational validity, cannabis takers are expected to give poor performance on jobs requiring the ability measured by the test. It should be noted that motor speed is one of the main components of a very big number of jobs in modern industry.

Second order interaction effect in highly significant (F = 16.18; d.f. 4/1466; p 0.001), denoting mutual dependency between drug use, literacy and residence. Thus in table 2 (of the urban groups) the discrepancy between takers and controls grows bigger the higher is the level of literacy. In table 3 (of examinees from lower Egypt or semi-rural areas) the same relationship holds to a great extent. With more interference of ruralism (in table 4) the relationship is not all that clear.

(4) Mark making (Part K of GATB)

This test is constructed to assess aiming and motor speed. This is shown by the fact the test has highly significant correlations of 0.65 and 0.64 with H Marking (which measures "aiming") and Speed (which gauges motor speed) respectively (N = 4000). (Mimeographed Handbook, Institute of Psychiatry, London, 1957.) It seems, moreover, that the test measures some kind of orientation in space.

The test failed to discriminate between takers and controls. In order to get some illumination as to how this failure came to take place the reader may look into table 2. Comparing the corresponding cells of users and non-users, the reader can see for himself that the superiority of the latter is obvious and that it reaches its highest value between literate takers and non-takers. In tables 3 and 4 the expected superiority disappears almost completely, and in the case of some comparisons it even gets reversed. However, this need not disturb us. "It is inherent in the nature of experiments that odd-looking results will appear from time to time." (Moroney, 1960, p. 416.)

The fact that "Mark Making" did not differentiate between takers as a whole and non-takers, though it is significantly correlated with "H Marking" and "Speed" (and indeed with "Tool Matching" too) both of which did discriminate significantly between the two groups means that something specific about "Mark Making" (not what is common between it and the other tests) triggered the subjects' responses. Such specific component seems to be identifiable as "orientation in space" (mentioned in this article). We do not know of any existing study which could extract (by factor analysis) this component, hence we have no way to determine its weight (as a loading on an identifiable factor). However, it might be interesting to know that "Mark Making" is significantly correlated with two tests for spatial ability (F and H of the same GATB). The correlations are 0.29 and 0.24 respectively (N = 4000).

(5) Trail Making (Part A)

This test is scored for speed of performance. Yet it is not as simple as the Speed test of the GATB. The task required is rather complex, involving shifts in organization, recall and recognition as well as motor performance (Partington and Leiter, 1949).

The Trail Making test did differentiate significantly between cannabis takers and controls (F = 7.13, p 0.01), takers being definitely slower than non-takers. In a group of 100 Egyptian male University students this test was shown to be significantly correlated with a number of tests gauging speed of psychomotor performance (A. El-Sheikh, 1971). Following are the correlations and the correlates:

Manual Dexterity (Part M of GATB): 0.41;

T.X.O.I. Cancellation Sheet: 0.32;

Following the contours of a star: 0.32.

In the same study the Trail Making test came out to be positively loaded (0.33) on a general factor of speed of psychomotor performance. Integrating relevant pieces of information as to the correlates of performance on this test the conclusion can be drawn that the results obtained on both our takers and controls on the test may be generalized to other tests and/or other situations in so far as those require speed of psychomotor performance. This conclusion adds more emphasis and further illumination to the findings already obtained an the "H Marking" and the "Speed" tests of the GATB.

Second order interaction effect on this variable is moderately significant (F = 2.83, d.f. 4/986; p 0.05), denoting mutual though rather weak dependency between drug taking, literacy and residence. In other words the main effect of drug taking is almost independent of literacy and/or residence. [ 9]

A number of points have to be emphasized: ( a) That the Trail Making test has been tried on several occasions on Egyptian subjects (including psychiatric patients belonging to different diagnostic groups as well as normals) prior to our work with cannabis takers. In all such occasions it proved adequate to measuring speed of psychomotor performance with a reasonable test-retest reliability. ( b) That it could differentiate between various psychiatric groups of patients, allocating each to a well defined position on the parameter of speed of performance, showing, among other things, normals to score fastest, organics with psychosis to score slowest and schizophrenics in between. ( c) That, with improvement of psychiatric condition, patients tend to perform faster. ( d) That within this context the scores obtained by cannabis takers (showing them to be slower than comparable non-takers) may be considered as pointers towards behaviour pathology.

(6) Initial reaction time

Score is time in seconds elapsing between the subject being exposed to a specified visual stimulus (a Rorschach card) and his giving any verbal response. Four such scores were given (in reaction to four Rorschach cards, I, IV, V and VI) to permit obtaining a rather reliable average. This test did not differentiate to any acceptable level of significance between users and controls. Inspection of tables 2 and 3, however, shows that takers in urban and semi-rural areas tend to show a longer reaction time than controls.

Initial Reaction Time is the last in our series of tests which require speed of performance. Other tests to follow have nothing to do with this parameter. At this juncture, a comment is warranted as to the consistency of the results obtained so far, and their meaning in the light of related findings reported by other research workers.

* * *

By and large cannabis takers proved to be significantly slower than controls on a variety of speed tasks. This finding is in line with the self description our subjects gave when they were interviewed (Soueif, 1971). Asked about the usual quantity of their output when at work our takers gave the following answers: 12.8 per cent said they produced very little, 27.7 per cent said they produced much, and 59.5 per cent said a moderate output of whatever they produced.. The respective percentages, according to the non-users' responses, were 7.1, 47.7 and 45.2. The discrepancy between the two sets of percentages was highly significant (Δ 2= 40.82, 2 d.f., p > 0.001). Most of the difference between the two groups may be accounted for by the category of "producing much" which favours the controls. It should be noted that we got a similar pattern of answers with our non-prisoner subjects (in Cairo) (Soueif, 1967). Of the users 14.53 per cent stated that they produced very little, 31.74 per cent said they produced much and 53.80 per cent opted for a moderate quantity. The respective control data were as follows: 5.22 per cent, 40.00 per cent and 54.78 per cent. The disparity between the two sets of responses was statistically significant (Δ 2= 7.018; 2 d.f., p > 0.05).

The effects of long-term chronic use of cannabis on psychomotor functioning has not been adequately explored (Report II of the Canadian Commission of Inquiry, 1972, p. 38). This is all the more true when it comes to the study of speed (rather than accuracy) of psychomotor performance. A few relevant studies, however, may be mentioned. Rubin and Comitas conducted an ethnopsychopharmacological study in Jamaica (starting in 1970, through December 1971). In the clinical part of the work they administered 19 psychological tests to 30 chronic cannabis smokers and 30 controls. The battery included one test (Finger Tapping) for the assessment of speed of motor performance. The investigators reported no significant differences between takers and controls on the test (Rubin and Comitas, 1972). It should be noted however that their "controls were defined as confirmed non-smokers who may have had some limited experience" (p. 226), with smoking ganja in the past. Moreover, the authors state it clearly that, " Ganja is taken by both smokers and controls in tea and tonics * for medicinal purposes and occasionally in food" (p. 228). In other words controls might have been smoking cannabis in the past, but they have stopped sometime before they volunteered to be studied. Yet they continued to take the drug in other ways ! According to the same authors, "Paradoxically, non-smokers who are regular tea drinkers may consume as much or more ganja as does a confirmed smoker" (p. 131). Under these conditions the definition of the control group sets a serious methodological limitation to the comparability of Rubin and Comitas' results with ours. In actual fact we contend that this clinical part of the Jamaican study might have (for the sake of clarity) been entitled "A comparative study of the psychological correlates of cannabis use by oral v. oral plus smoking routes of administration". Such a headline would have made it clear that the independent variable in this study was not drug taking but the route of administration.

* Our italics.

Rubin and Comitas, however, in the anthropological part of the work ended with another type of conclusion: "Data on agricultural pursuits and other activities among ten heavy users, eight moderate users and twelve light users corroborate (other) data summarized (before). Total space covered or amount accomplished in, or number of plants reaped, is usually reduced per unit of time after smoking". (Rubin and Comitas, 1973, p. 211.) The reader should be aware that this conclusion was reached as a result of studying chronic cannabis takers while under the immediate effect of ganja smoking. This underlines an important difference between the anthropological part of the Jamaican study and our work, where we interviewed and tested chronic cannabis takers while (according to their own statement) they were not under the influence of the drug. Nevertheless, the cited conclusion of Rubin and Comitas is in line with our main interview and test findings.

Clark and Nakashima (1968) gave marihuana extract to a group of educated naive subjects, and used them as their own controls. Those investigators found that both simple and complex hand and foot reaction times increased (showing slowness) under drug effect. This is in agreement with our data regarding the performance of urban and semi-rural subjects on the Initial Reaction Time test.

Boflman and Pihl used (among other tools) the Halstead Tapping test for fine psychomotor speed, the Pins Test-Lincoln-Oseretsky for fine psychomotor speed with accuracy, together with a test for gauging reaction time, in two field studies conducted in Jamaica (Bowman and Pihl, 1973). In the first study they administered the tests to 16 users and 10 controls. In the second study, the investigators gave the tests to 14 users and 14 controls. The results of the two studies showed no significant differences between takers and controls. It should be noted that in the two studies ruralism and illiteracy were quite high among testees (Third Annual Report to the U.S. Congress 1973, p. 141). In the light of the interaction effects, already demonstrated, between cannabis taking, literacy and residence, the following question is raised as to Bowman and Pihl's findings: To what extent could illiteracy and ruralism have blunted the differences between users and controls?

The Canadian Commission of Inquiry (1972) carried out a number of experiments in which they administered placebo, marihuana and synthetic THC to some 14 subjects, mostly university students who were already experienced with the drug. The Commission reported no effect of (smoked) cannabis on maximum tapping speed ( op. cit., p. 58). This finding, warrants some comment in explanation, in view of the fact that our controls showed a highly significant superiority over cannabis takers on the Speed test (Part G of the GATB) which might look similar to the tapping test used by the Commission. Closer examination, however, shows that the two tests are not all that similar. At least one main difference can be underlined: That the Speed test is more structured than the tapping test. Describing a test as more structured here means that the task specified by this test, has more differentiated details (or steps) that go into it. Such details are explicity defined in the instructions and/or in the test items, thus triggering for their accomplishment a rather broad variety of mental processes. In the Speed test, the reader may recall, that examinees were required to put: ( a) 3 dots, ( b) inside each rectangle, ( c) filling as many rectangles as they could in a limited span of time. In other words our subjects were instructed to execute a rather complex action implying three different dimensions. Moreover, they had to maintain a reasonable degree of alertness to keep the count of the dots within each rectangle right, and to follow the distribution of the rectangles on the test sheet. In contrast with that, the Tapping-Speed test was certainly less complex. Subjects were simply required to tap, as often as they could, with a hand held pencil-like stylus. The score was the total number of taps made within one minute ( op. cit., p. 133). This difference in the complexity of structure of the two tests should cause a marked difference in the dynamics of performance on each. On the Speed test, our subjects had to follow a pattern of rhythmically alternating between putting dots and getting a visual-perceptual feedback to be followed by a change of locus of the dots to come next. But on the Tapping test no such pattern was imposed. In terms of Hull's learning theory, performance on the Tapping test would, therefore, be expected to generate reactive inhibition (experienced as muscle strain, fatigue and boredom) at a higher rate than the Speed test (Hull, 1952). The more repetitive and homogeneous the task is the more reactive inhibition it is bound to generate in a specified time unit. It is also obvious that the longer the time allotted for continuous performance on such a test, the more confounded the results would be, because of more interference of reactive inhibition. In the Commission's work the time allotted for performance on the Tapping test was one minute (compared with 30 seconds for performance on the Speed test in our study). It is possible that the end result of the accumulation of inhibition all through one minute of continuous tapping would level out significant differences that would have emerged between placebo and drug effect. It is therefore suggested that with tests like Tapping speed, time allowed for the test should be rather short, say 20 seconds, if the aim is to get at a rather unconfounded measurement of motor speed. This does not imply that Tapping speed should have been impaired by cannabis use. It might be the case that this parameter is not sensitive to the drug effect. What is actually measured however, whether it is Tapping-speed or any other variable, should be given, experimentally and conceptually, optimum clarity. Obviously, the same comment is also relevant to Bowman's (Bowman and Pihl, 1973) and Rubin's (Rubin and Comitas, 1972) findings already discussed.

Apart from the remotely related four works we have mentioned, speed of psychomotor functioning in chronic cannabis takers was seldom investigated. Yet this parameter has been repeatedly demonstrated by a whole series of clinicians to be significantly correlated with psychiatric disorder. Babcock (1930, 1933) first demonstrated that performance on error-free psychomotor speed tests was abnormally impaired by schizophrenics (A. Yates, 1973). Shapiro and associates (1955) concluded that psychiatric patients were slowest in tasks which did not involve problem solving and which emphasized speed in the instructions. A few years later Payne and Hewlett (1960) confirmed those findings in a well designed large scale study (A. Yates, 1973). In a series of studies carried out in Egypt, Soueif and Metwally (1961), Soueif, Abd-el-Naby and Helmy (1964) and Ashour, Soueif and Hassan (1971) cross-validated a number of the findings reported by the previously cited researchers, on various groups of functional and organic (hospitalized and non-hospitalized) psychiatric patients. It is interesting to note that in all the Egyptian studies tests scored for speed of psychomotor functioning (including the Trail Making and the Manual Dexterity of the GATB) were the best differentiators among our clinical groups.

Noteworthy, here, is the fact that this part of the clinical literature was among the sources that inspired us to enquire into the possibility of hashish takers showing any aspect of psychomotor retardation. Not that we imply that the mechanisms operating to bring out psychomotor slowness should be the same in cannabis takers as in co-operative psychotics. Indeed we do not claim to have a theory to account for the slowness shown by our chronic takers, and we would rather keep our report at a descriptive level. Yet the research worker cannot, and should not, ignore some analogies. The history of scientific research knows many examples where perceiving certain analogies provided a passage towards an adequate theory.

* * *

DISTANCE ESTIMATION

This test consisted of three sticks of varied length, made of plain wood. The sticks were presented to the testee one at a time with a request to give verbal estimates of their lengths in centimetres (Soueif, 1971). The following three test variables were derived therefrom:

(7) Distance overestimation

Discrepancies (between estimated and objective lengths) showing a tendency towards overestimation were analysed. The average size of overestimations made by cannabis takers was found to exceed significantly what was made by controls (F = 4.75, p > 0.05). This result is in agreement with the verbal responses given by our convicted takers. To the question "when you are high, do you find that distances seem longer, shorter or as usual?", the answers were as follows: 38.9 per cent said distances seemed longer, 8.6 per cent shorter and 52.5 per cent as usual. When our takers were asked, "what about the morning after, longer, shorter or as usual?", 4.5 per cent said longer, 5.8 per cent shorter and 89.7 per cent as usual. We obtained the same pattern of answers (though more sharpened) with out urban non-convicted interviewees. Thus under the immediate effect of the drug respective percentages were 81.86, 8.82 and 9.41 per cent; and the morning after 9.57, 6.91 and 87.52 per cent.

Second order interaction effect was found to be significant (F = 6.37, d.f. 4/1613, p .> 0.01). One could therefore expect some combinations between · cannabis taking, literacy and residence,to be more favourable than others to showing the superiority of controls over takers on this test variable.

(8) Distance underestimation

On this variable we analysed those discrepancies between estimated and objective lengths which show a tendency among our subjects towards underestimation. The outcome of analysis of variance disclosed that the average magnitude of underestimation made by cannabis takers was significantly less than the average made by controls (p > 0.05). In other words controls were more than takers in the direction of underestimating distances of moderate lengths.

Second order interaction was significant (F = 8,35; d.f. 4/1628; p > 0.01). Inspection of tables 2, 3 and 4 shows that the rural groups give the clearest picture of the discrepancy cited above. In the urban groups the biggest difference is only shown between the literates.

(9) Distance estimation: discrepancy

The target data for analysis was the size of the differences between estimated and objective lengths of the sticks, irrespective of the direction of such differences. Since both test variables No. 7 and No. 8 disclosed significant differences at almost equal levels of statistical significance but in opposite directions, test variable No. 9 would naturally show no significant discrepancies. This illustrates how analysis of global disparities between drug takers and controls on some variables might conceal facts which would radically modify what was revealed.

The experimental literature on distance estimation as related to cannabis consumption is really meagre.. The LaGuardia group of researchers reported that marihuana produced no significant changes in perception of length of lines on paper (Clark and Nakashima, 1968). However, we cannot judge the value of this finding since, "Placebos were not generally used in this section (of the study) and the details of the statistical analysis were hot presented" (Interim Report, 1970, p. 87).

A distantly relevant study has recently been reported by Beach and others (1973). The investigators tried to shed light on what they called the "objective" and the "subjective" estimation of distance within the context of a comparative study of cannabis and alcohol effects under simulated driving conditions. By "objective estimation" the investigators meant that their subjects were instructed to use intellectual clues. The intellectual clue, here, was the word " think" as embedded in the following question: "How far do you really think you have been driving?" The emotional clue designed to induce subjects to make subjective estimations was the word " feel" replacing the word "think' in the same question. The investigators found that objective as well as subjective estimates increased under cannabis (Bech et al., 1973).

TIME ESTIMATION10

Under this rubric two variables were gauged:

(10) The ability to estimate verbally the duration of a certain period (3 minutes clock-time)

When comparison was made between average estimates made by takers and controls, differences were found to be of no statistical significance. In fact the variance estimate due to individual differenees other than drug-taking behaviour was almost twice as large as the variance estimate due to drug use.

(11) Time estimation discrepancy

Analysis was carried out on the differences between estimated and clock-time duration of the target period, irrespective of the direction of such differences. Discrepancies between takers and controls on this .variable were found to be statistically significant (p > 0.05). Contrary to popular notion, however, the average size of error made by cannabis takers is smaller than that made by non-takers (c.f. tables 2 to 4).

Second order interaction effect between drug taking, literacy and residence was highly significant (F = 4.74; d.f. 4/1411, p > 0.001), again underlining the fact that some combinations between those variables would give more prominence to the discrepancy cited above.

The experimental literature on the effect of cannabis consumption on time perception has been oscillating between conflicting results (Brombers, 1941; Dornbush et al., 1971; Clark et al., 1970). But recently more studies have come out with almost one basic finding, i.e. significant change in time estimation is effected by the drug.

Weil and colleagues reported a tendency among subjects who had no previous experience with the drug (N = 9) to overestimate a period of 5 minutes, when under the effect of smoked marihuana (Weil et al., 1968). Tinklenberg and associates gave orally social doses of marihuana extract calibrated for THC content to 15 subjects who were all moderate cannabis takers. Placebo was also given within the frame-work of a double blind design. Time perception was assessed by the technique of time production. The investigators found that marihuana induced under-production (implying overestimation of clock-time) on this test compared with placebo (Tinklenberg et al., 1972). In another experiment by Hollister and Tinklenberg the same test of time production was utilized. But this time there were no significant differences between time-production under marihuana and that under placebo (Hollister and Tinklenberg, 1973). Bech and associates demonstrated that cannabis had more marked effect than alcohol on time estimation and that this was more particularly so on subjective rather than on objective estimation (Bech et al., 1973). The Canadian Commission of Inquiry reporting on one of their experiments on "A comparison of ΔE9 THC and marihuana effects in humans'', maintained that "estimates of the time spent in the visual imagery, speech, and painting tasks were longer with cannabis (natural and synthetic) than with no drug; the cannabis resulted in over-estimates of clock-time, while under-estimates of clock-time were usually observed in the placebo condition. The drugs did not result in a change in 15-second time interval production". (Canadian Commission, 1972, p. 134.)

Collating the results cited above, it seems rcasonable to conclude that there is growing evidence that the drug induces some kind of change in the ability to estimate time intervals of moderate duration. To that extent the cited results are in agreement with ours. There is some consistency, moreover, in the direction of the acutely induced change, moving towards overestimation of clock-time. It is interesting to note that such important points of convergence emerged in spite of the fact that the investigators used widely divergent methods of assessment of time estimation under experimental conditions that.varied from one study to the other. We know from the literature on time perception studies in normals and abnormals, that the estimation of a certain period is very sensitive to the techniques utilized in testing, to the duration of the intervals to be estimated, and to other experimental conditions (Bindra and Waksberg, 1956; Orme, 1962; Wallace and Rabin, 1960). To attain more clarity, however, it might be advisable to conduct some research in the near future, on the various techniques of time estimation to help determine their relative sensitivities to drug effect. Investigators can then proceed to use one method, the most sensitive, with a strictly standardized procedure of administration, while systematically varying other independent variables. This approach would, hopefully, provide optimum conditions for comparability of findings across experiments and eventually across experimenters.

* * *

DIGIT SPAN

Three test variables will be presented 'in this section: "Digits Forward", '"Digits backward" and "Digits forward and backward'. They are derived from the Wechsler intelligence scale (Guertin et al., 1966; Wechsler, 1954).

(12) Digits Forward

In this test the subject is required to repeat sets of digits of increasing number right after the tester. Score is highest number of digits repeated without error. The over-all difference between takers and non-takers on this test was hot statistically significant.

(13) Digits backward

In this test the examinee is instructed to repeat, in reverse order, sets of digits of increasing number right after the tester. Score is the highest number of digits correctly reversed. This test differentiates at a very high level of signifieance between takers and non-takers (p > 0.001). Inspection of tables 2 to 4 shows that controls are definitely superior to cannabis takers on this variable.

Second order interaction between drug taking, literacy and residential area was found to be moderately significant (F = 3; d.f. 4/1625; p > 0.05). Because the F ratio of the main effect (of drug taking) was so large, compared with the F for the interaction effect we were tempted to undertake the same statistical exercise we did with the "H Marking" test. Taking the variance estimate for the second order interaction, as the error term against which to check the significance of the mean squares for the first order interactions, and finding that they were not significant, we pooled them all together with the variance estimate of the second order inter-action. Now we used the pooled variance estimate as an error variance, against which to check the significance of the drug taking effect. This was highly significant (F = 17.93; d.f. 1/12; p > slightly below 0.001). The implication is that if we consider residential areas as random sample representing the whole country, drug taking effect, then, is so prominent that the difference between takers and non-takers does not depend on how urban or rural our subjects are. Since the first order interaction between cannabis use and literacy is, also, hot statistically significant, the second conclusion to be drawn is that takers are inferior to non-takers on the ''Digits Backward" irrespective of literacy.

(14) Digits /forward and backward

Analysis of variance of the total score on Digits forward and blackward showed no significant differences between takers and non-takers. Here we have another case where scores on a global test variable conceal meaningful differences.

'Theresults obtained on the ''Digit Span' tests invite some comments on the corresponding interview data. There is some agreement between the two sets of findings. Thus, among the convicted takers 13.8 per cent said they usually had a weak memory, 66.2 per cent claimed they had a sound memory and 20 per cent stated their memory was moderate. The percentages among the controls were: 12.7 per cent, 56.5 per cent and 30.8 per cent respectively (x2 = 13.73, 2 d.f., p > 0.01). But the picture in the non-prisoner groups (of Cairo) is much clearer (though it does not present a higher level of significance of differences). Among the takers 19.79 per cent maintained that they usually had a weak memory, 50.5 per cent a sound memory, and 29.26 moderate. Contrasted with this we have among the controls: 7.83 per cent, 54.78 per cent and 37.39 per cent respectively. The difference between the two sets of percentages is highly significant (x 2= 8.17; 2 d.f., p > 0.02) and can readily be accounted for by the percentages coming under the two categories of weak and moderate.

* * *

The acute effect of cannabis on immediate memory has been investigated by a number of research Workers. But their work varied regarding the kind of tools they used, the subjects they experimented on, and the results they reported. Melges and others administered the digits forward and backward to 8 normal male graduates within the framcwork of a double blind design of experiment, using various oral doses of marihuana extract calibrated for THC content. The investigators found that, "Both forward and blackward digit spans were significantly deceased by all doses of THC" (Melgcs et al.. 1970). Waskow and colleagues gave the same two tests to 32 male criminal offendcrs who took the drug under rather similar experimental conditions. However, this group of experimenters stated that the drug had no significant effect on repeating digits forward or backward (Waskow et al., 1970). Tinklenberg and associates used another test called the "Running memory span". "In this test the subjects listened to sequences of random digits presented by audiotape at the rate of 1/sec. The scquences varied randomly in length from 10 to 20 digits and were arranged so that the last seven digits had no repetitions... The subject was instructed to write in correct sequential order as many digits as he could accurately remember from the end of the series. Four trials were given at each test period. The score consisted of the average number of correct digits in the exact sequence of presentation" (Tinklenberg et al., 1972). Obviously the procedure followed in administering the "Running memory span' is parable to the digit span technique. The experimenters administered the test to 15 college-educated men who were given, orally, marihuana extract calibrated for THC content (26 mg.), placebo, and alcohol within the context of a double blind design of experiment. Those investigators reported the following, "A three way analysis of variance (drug rime of testing subjects) indicated no significant differences in the "Running memory span performances during any of the experimental conditions". They added, however, that "The poorest mean scores were obtained during peak drug effect (... 11/2 hours for marihuana), but these differences did not reach significance." ( op. cit.) The fact that the subjects were instructed to write the digits they remembered instead of repeating them orally might have been one of the subtle reasons why the results were inconclusive compared with the results obtained previously by the same group of researchers on the usual Digit span test (c.f. Melges et al., 1970). Raeburn and Tong in a series of studies on a test closely similar to the Running memory span, except that the former used words instead of digits, found that "speed of writing" was a factor which partly accounted for the differences shown on the test between normals and abnormals (Raeburn and Tong, 1968). It is possible that this variable as a component of individual differences, unaccounted for in a systematic way, contributed to the confounding of the results reported by Melges and colleagues.

Various studies have been carried out on the related phenomena of short term memory (Abel, 1970; Canadian Commission, 1972, p. 134; Dittrich et al., 1973; Zeidenberg and others, 1973; Darley et al., 1973) and temporal disintegration (Weil and Zinberg, 1969; Dittrich et al., 1973). Short term memory "can be operationally distinguished from immediate memory span in that in the case of short-term retention, the attention is directed away from the material presented before the subject is later asked to recall it. To ensure this, subjects are usually required to carry out some interpolated task before they are asked to recall the initial material" (R. W. Payne, 1973). Temporal disintegration is a mental condition that can be detected through a specific speech difficulty which "has two principal manifestations: simple forgetting of what one is going to say next, and a strong tendency to go off on irrelevant tangents because the line of thought is lost". (Weil and Zinberg, 1969). According to the same authors, the disturbance seems to be due to some kind of"interference... with retrieval of information while it is in an immediate memory storage" (viz. the past few seconds). It is interesting to note that, the results of the studies on short term memory and temporal disintegration, in contrast with those on immediate memory, tend to agree rather than disagree, pointing out a deleterious effect of the drug. Whether the tests used in the two kinds of investigations assess various modalities of one and the same function, can only be decided on empirical grounds. Until this is verified (possibly by using factor analytical techniques) any extrapolation from one group of inquiries to the other should be exercised with caution. In actual fact, even attempting to present a theory to account for the findings within any one of the two areas of investigation, viz. immediate or short term memory, should be practised with reservation because of the multiplicity and/or complexity of the variables which could be contributing to the end result. Recent experimental literature on immediate memory span in psychiatric patients points out three such variables: ( a) their ability to concentrate on the stimulus material. ( b) their ability to organize this material, where this is possible, into larger units easier to recall, and ( c) the speed with which they can reproduce the information before it decays (Payne, 1973). Experiments on short term memory provide strong evidence as to the influence of motivational factors (Weiner, 1966).

BENDER GESTALT VISUAL MOTOR TEST

Two test variables were utilized under this caption:

(15) Bender Gestalt test (copying)

In this test subjects were required to copy six designs presented by tester, one at a time. Subjects' performance was scored for accuracy but not for speed (using the Loefving system of scoring, c.f. Shapiro et al., 1957). The test was found to differentiate at a very high level of significance between takers and non-takers (p > 0.001). Inspection of tables 2 to 4 shows that on the whole controls were superior to drug takers.

Second order interaction between cannabis taking, literacy and residence was statistically significant (F = 4.37; d.f. 4/1584; p > 0.01). Again, the reader can see in tables 2 to 4 that some combinations of the three independent variables show better than others the main drug taking effect.

(16) Bender Gestalt test (recall)

In this test subjects were instructed to draw, out of memory, those designs they have just copied in (15). Performance was scored for accuracy. Analysis of variance disclosed highly significant differences between cannabis takers and controls (p > 0.001). Again the detailed tables 2, 3 and 4 show the superiority of the latter over the former.

Though all first order interactions were not significant, the second order interaction effect came out very highly significant.

* * *

The Bender Gestalt test (copying and recall versions scored according to the Loefving system) has been used in a number of investigations on functional and organic psychiatric groups in Egypt (Soueif and Metwally, 1961; Soueif et al., 1964; Ashour et al., 1967) and in various other countries (Shapiro et al., 1956, 1957; Garron and Cheifetz, 1965). The clinical efficiency of the test in discriminating between the two groups of patients has been subject to some controversy (Savage, 1973). In our Egyptian studies the recall version of the test proved to be a good differentiator in so far as statistical tests of significance are concerned, the copy version gave conflicting results and the question of misclassification remained a source of concern for clinical investigators.

Following the logic of its scoring system (without going any deeper into the question of associating certain scores with some diagnostic labels), this test discloses superiority of controls over cannabis takers regarding ( a) proficiency of hand-eye co-ordination and ( b) memory for designs.

In the literature on the acute effects of cannabis in man, other tests for the assessment of hand-eye co-ordination (but not the Bender Gestalt) were reported on. Perhaps the pursuit rotor (and/or variations of it) were utilized by most workers. Clark and Nakashima reported no consistent drug effect on the pursuit rotor, but maintained that "performance decrement did appear in some subjects at the highest dose level" (Clark and Nakashima, 1968). Weil and colleagues, on the other hand, demonstrated significant decrements in performance on pursuit rotor of naive subjects while under both effects of low and high doses of cannabis. (Weil et al., .1968.) The results obtained by chronic cannabis users while under immediate drug effect were confused, and hence were not presented in the published report (Weil et al., 1968). Manno and associates reported that impairment on pursuit tracking performance would occur under the effect of cannabis smoking whether the smokers were naive or drug-experienced. (Manno et al., 1970; 1971.)

One may conclude that there is growing evidence to the effect that decrement in performance on unspeeded motor tasks requiring hand-eye co-ordination is demonstrated under the immediate effect of cannabis. So far, the experimental literature tends to show agreement with our findings on the copy version of the Bender Gestalt test. In the absence of empirical demonstration of significant correlation between this test and the pursuit rotor (and its derivatives) any inferences from this agreement should be considered with caution.

"Memory for designs" was seldom reported on in this area of research.

Summary and conclusion

Twelve objective tests, generating 16 test variables, were administered to 850 chronic cannabis takers and 839 controls. The tests were intended to assess various modalities of some 5 broad psychological functions: speed of psychomotor performance, distance estimation, time estimation, immediate memory and visual motor co-ordination. Most of the test variables differentiated significantly between takers and non-takers and detailed analyses showed the latter to be most of the time superior to the former. In the majority of our test variables statistically significant second order interaction effects emerged, denoting that, under certain conditions relating to the two dimensions "literacy-illiteracy" and "urbanism-ruralism", the superiority of controls over cannabis users became much more impressive.

Except for two studies carried out on few chronic takers in Jamaica, no empirical studies were found in the literature where investigators utilized standardized objective tests to give dependable descriptions of behavioural correlates of chronic cannabis consupmtion. Moreover, very few studies showed enough interest in the investigation of psychomotor performance and/or memory for designs in so far as such functions are related to cannabis taking. The experimental literature on the immediate effects of the drug, however, tackled some areas which were identical with or closely related to those we investigated. Some agreement could be detected between the findings emerging in the two areas of chronicity and immediate effects of the drug.

The clinical psychological literature was mentioned where such findings as slowness of psychomotor performance, defective visual-motor co-ordination and impaired memory for designs were reported to be associated with psychiatric illness. It was thereby suggested that our results pointed to meaningful associations between chronic cannabis consumption and manifestations of psychopathology.

A number of areas, where further research is urgently needed, were underlined.

001

1 Thanks are due to all the personnel involved in the prison administration, for the understanding and the flexibility they showed, without which it would have been impossible to carry out this work.

002

2 Three comparisons and four interactions were supplied for each test variable, as follows:

  1. Comparisons: between drug takers and controls irrespective of literacy and residence. Between levels of literacy irrespective of drug taking and residence. Between residence positions regardless of drug behaviour and literacy.

  2. Interactions: drug taking - literacy; drug taking - residence; literacy - residence; drug taking - literacy - residence.

003

3 The reader may notice that in some rows the standard deviations vary widely, thus raising questions as to the assumption of homogeneity of the variances supposed to be necessary for the applicability of the analysis of variance. Concerning this assumption, however, and the assumption of normality of distribution McNemar states the following:

"Although these assumptions are incorporated in the mathematical derivation of the F distribution, there is ample evidence that marked skewness, departures from normal kurtosis, and extreme differences in variance... do not greatly disrupt the F test as a basis for judging significance in the analysis of variance... If the investigator wishes to have some assurance that he is not risking the making of the type I error more often than his chosen level for judging significance, he may wish to adopt a somewhat more rigorous level: requiring a computed F to reach the 0.01 level provides a very safe base for claiming significance at the 0.02 level." (McNemar 1969, p. 288.) And Hays maintains that:

"A test for homogeneity of variance before the analysis of variance has rather limited practical utility, and modern opinion holds that the analysis of variance can and should be carried on without a preliminary test of variances..." (Hays 1963, p. 381.)

004

4 This is inherent in the following two facts: ( a) that the "Tool Matching" test has a high reliability (i.e. consistency of results from one testing situation to another), ( b) that this test is a good tool for measuring a psychological function (viz. form perception) the existence and the significance of which have been well established (by sophisticated statistical techniques involved in methods of factor analysis).

005

5 To be sure, the battery has been developed and validated in the USA. This, however, does not invalidate our conclusions regarding Egyptian Ss. One of the main reasons for this immunity is that we have used a control group whose scores are utilized, all through this work, as a frame of reference for evaluating the meaning of the scores obtained by cannabis takers. It should also be noted that the items comprising the four tests we opted to select from the GATB did not seem to have aroused any cultural shock among our Ss during the testing situation. Moreover, the National Centre for Social and Criminological Research in Cairo, Egypt, has sponsored and financed a project of restandardization of the battery on some three thousand Egyptian young men, the intention being to use it for occupational selection and vocational guidance (Abd-el-Quader, 1971). In the end most of its items and instructions were kept as in the original. When using well known standardized (such as those we derived from the GATB) for the description of certain features of the behaviour of drug takers, the research worker can be pretty sure that the literature is generally full of information on such tests. Such information may shed light on further problems not tackled directly by the investigator.

006

6 The sample used for this study is the GATB General Working Population Sample of 4,000, including 1,844 males and 2,166 females. (c.f. Mimeographed Handbook of Tests used at the Institute of Psychiatry, London 1957.)

007

7 As any other prediction based on statistical analysis, this one remains valid within certain limits. Among the elements which define those limits are the magnitudes of the correlations and their invariance from one sample of testees to another. The two correlations cited above are very highly significant considering the size of the sample on which they were based (N-4000). Moreover, it is a well known fact that correlations among highly reliable test variables tend to remain stable across samples (at least much more stable than norms). Consequently, we are reasonably confident that Egyptian cannabis takers will prove inferior to non-takers on tests (and jobs) which require matching any visual percepts.

008

8 Because some aiming seems to be involved in such performance this test is correlated with H Marking up to 0.56 (N-4000).

009

9 The Trail Making test has been used by a number of investigators as a screening device for the detection of brain damage (Brown et al., 1958). In Egypt it has been successfully used to differentiate between groups of male organic psychiatric patients, schizophrenics, organics without psychotic manifestations and normals (Soueif and Metwally, 1961; Ashour, Soueif and Hassan, 1967). It has also been used in this country to assess improvement of hospitalized male and female schizophrenics who were administered reserpine, and was found sensitive enough to detect progressive increase in speed of performance (Soueif et al., 1964). We could, also, establish a number of significantly negative correlations between the test and several tests of creative thinking abilities administered to 200 schizophrenic male patients. Noteworthy is the fact that all our tests of creativity used in the cited research emphasized speed of performance in the instructions. (Soueif and Farag, 1971.)

010

10 For some technical reason we could not analyse the time estimation data following the same pattern utilized in the analysis of distance estimation.

Bibliography

Abd-el-Quader, M. Standardization of the General Aptitude Test Battery in Egypt, Publications of the National Centre for Social and Criminological Research, Cairo 1971 (in Arabic).

Abel, E.L. Marijuana and memory, Nature, 227, Sept. 12, 1151-1152, 1970.

Anastasi, A. Psychological testing, London: Macmillan, 3rd ed., 1968.

Ashour, A. M., M. I. Soueif and A. H. Hassan. Clinical and electroencephalographic validation of psychometric tests of organicity in Egyptian psychiatric patients, in press.

Babcock, H. An experiment in the measurement of mental deterioration, Arch. Psychol., 18/117, 1930.

-------------------. Dementia Praecox: A psychological study, New York: Science Press, 1933.

Bech, P., L. Rafaelsen and O. J. Rafaelsen. Cannabis and alcohol: Effects on estimation of time and distance, Psychopharrnacologia (Berlin), 32, 273-381, 1973.

Bemis, S.E. Occupational validity of the General Aptitude Test Battery, J. appl. Psychol., 52, 240-244, 1968.

Bindra, D. and Waksberg, H. Methods and terminology in studies of time estimation, Psychol. Bull., 53, 155-159, 1956.

Bowman, M. and R. O. Pihl. Cannabis: Psychological effects of chronic heavy use, Psychopharmacologia (Berlin), 29, 159-169, 1973.

Bromberg, W. Marihuana intoxication, Amer. J. Psychiat., 91, 303-305, 1941.

Cannabis: A report of the Commission of Inquiry into the non-medical use of drugs, Ottawa: Information Canada, 1972.

Clark. L. D. and E. N. Nakashima. Experimental studies of marihuana, Amer. J. Psychiat., 125/3, 135-140, 1968.

------------------- . R. Huges and E. N. Nakashima. Behavioural effects of marihuana, Arch. gen. Psychiat., 23, 193-198, 1970.

Committee for the Investigation of Hashish Consumption in Egypt. Hashish consumption in Egypt: Research in progress, Report I: The interviewing schedule: Construction, reliability and validity. Cairo: Publications of the National Centre for Social and Criminological Research, 1960 (in Arabic).

-------------------- . Hashish consumption in Egypt: Research in progress, Report II: Hashish users in Cairo City: A pilot study, Cairo: Publications of the National Centre for Social and Criminological Research, 1964 (in Arabic).

Cronbach, L.J. Essentials of psychological testing, New York: Harper, 2nd, ed., 1960.

Darley, C. F., J. F. Tinklenberg, L. E. Hollister and R. C. Atkinson. Marihuana and retrival from short-term memory, Psychopharmacologia (Berlin), 29, 231-238, 1973.

Dittrich, A., K. Buttig and llka von Zeppelin. Effects of ( - )Δq-trans-Tetrahydrocannabinol (Δq-THC) on memory, attention and subjective state, Psychopharmacologia (Berlin), 33, 369-376, 1973.

Dornbush, R. L., M. Fink and A. M. Freedman. Marihuana, memory and perception, Amer. J. Psychiat., 128, 194-197, 1971.

Edwards, A. Experimental design in psychological research, New York: Rinehart, 1956.

El-Sheikh, A. Personal tempo and the metric tempo in preferred poetry: A correlational study, Unpublished M.A. thesis, Cairo University, Cairo, 1971.

Garron, D. C. and D. I. Cheifetz. Comment on "Bender Gestalt Discernment of Organic Pathology", Psychol, Bull., 63, 197-200, 1965.

Guertin, W. H., C. E. Ladd, G. H. Franck and A. I. Rabin. Research with the Wechsler intelligence scales for adults, Psychol. Bull., 66/5, 385-409, 1966.

Hays, W. L. Statistics for psychologists, New York: Holt, Rinehart and Winston, 1963.

Hollister, L. N. and J. R. Tinklenberg. Subchronic oral doses of marihuana extract, Psychopharmacologia (Berlin), 29, 247-252, 1973 (mimeographed).

Hull, C. L. A behaviour system: An introduction to behaviour theory concerning the individual organism, New Haven: Yale Univ. Press, 1952.

Interim Report of the Commission of Inquiry into the non-medical use of drugs, Ottawa: Information Canada, 1970.

Manno, J. E., G. F. Kiplinger, S. E. Haine, I. F. Bennett, and R. B. Forney. Comparative effects of smoking marihuana or placebo on human motor and mental performance, Clin. Pharmacol. Ther., 11: 808-815, 1970.

------------------------ , --------------------- . N. Scholtz and R. B. Forney. The influence of alcohol and marihuana on motor and mental performance, Clin. Pharmacol. Ther., 12: 202-211, 1971.

Marihuana and Health: Third annual report to Congress from the Secretary of Health, Education, and Welfare. National Institute on Drug Abuse, Rockville, 1973.

McNemar, Q. Psychological statistics, New York: J. Wiley, 4th ed. 1969.

Melges, F. T., J. R. Tinklenberg, L. E. Hollister and H. K. Gillespie. Marihuana and temporal disintegration, Science, 168, 1118-1120, 1970.

Mimeographed Handbook. Institute of Psychiatry, London, 1957.

Moroney, M. J. Facts from figures, London: Penguin, 3rd ed., 1960.

Orme, J. E. Time studies in normals and abnormal personalities, Acta Psychol., 20, 285-303, 1962.

Partington, J. E. and R. G. Leiter. Partington Pathways Test, Psychol. Serv. Center Bull., 1, 9-20, 1949.

Payne, R. W. and J. H. G. Hewlett. Thought disorder in psychotic patients, Experiments in Personality, vol. II, Psychodiagnostics and psychodynamics, H. J. Eysenck ed., London: Routledge and Kegan Paul, 1960, 3-104.

------------------- . Cognitive abnormalities, Handbook of abnormal psychology, H. J. Eysenck ed., London: Pitman med., 2nd ed., 1973, 420-483.

Raeburn, J. H. and J. E. Tong. Experiments on contextual constraints in schizophrenia, Brit. J. Psychiat., 114, 43-52, 1968.

Rubin, V. and L. Comitas. Effects of chronic smoking of cannabis in Jamaica. A report by the Research Institute for the Study of Man to the Center for Studies of Narcotic and Drug Abuse, National Institute of Mental Health, Contract No. HSM-42-70-97, 1973 (mimeographed).

Savage, R.D. Old age, Handbook of abnormal psychology, H. J. Eysenck ed., London: Pitman med., 2nd ed., 1973, 645-688.

Shapiro, M. B., R. Kessel and A. E. Maxwell. Speed and quality of psychomotor performance in psychiatric patients (mimeographed copy, 1955).

------------------- . J. Field and F. Post. An enquiry into the determinants of differentiation between elderly "organic" and "non-organic" psychiatric patients on the Bender Gestalt Test. J. ment. Sci., 103, 364-374, 1957.

Soueif, M. I. Hashish consumption in Egypt, with special reference to psychosocial problems, Bulletin on Narcotics, XIX: 2, 1-12, 1967.

------------------- . The use of cannabis in Egypt: A behavioural study, Bulletin on Narcotics, XXIII: 4, 17-28, 1971.

------------------- . Cannabis ideology: A study of opinions and beliefs centering around cannabis consumption, Bulletin on Narcotics, XXV: 4, 33-38, 1973.

------------------- . Some determinants of function deficits associated with chronic cannabis consumption. (In press, 1975.)

------------------- . and A. Metwally. Testing for organicity in Egyptian psychiatric patients, Acta Psychol., 18, 285-296, 1961.

------------------- . S. Abd-el-Naby and A. Q. Helmy. Objective assessment of psychiatric changes produced by reserpine in Egyptian schizophrenics, Acta Psychol., 22, 85-99, 1964.

------------------- . and S. Farag. Creative thinking aptitudes in schizophrenics: A factorial study, Sciences de l'Art. Scientific Aesthetics (Paris), 8/ 1, 51-60, 1971.

Tinklenberg, J. R., B. S. Kopel, F. T. Melges and D. N. Hollister. Marihuana and alcohol, Arch. Gen. Psychiat., 27, 812-815, 1972.

Wallace, M. and Rabin, A. I. Temporal experience, Psychol. Bull., 57/3, 213-236, 1960.

Waskow, I. E., J.-E. Olsson, C. Salzman, M. M. Katz and C. Chase. Psychological effects of tetrahydrocannabinol, Arch. Gen. Psychiat., 22, 97-107, 1970.

Wechsler, D. The measurement of adult intelligence, Baltimore: Williams and Wilkins, 1954.

Weil, A. T. and N. E. Zinberg. Acute effects of marihuana on speech, Nature, 222, May 3, 434-437, 1969.

Weiner, B. Effects of motivation on the availability and retrieval of memory traces, Psychol. Bull., 65/1, 24-37, 1966.

WHO. Technical Report No. 526, Geneva 1973.

Yates, A. Abnormalities of psychomotor functions. Handbook of abnormal psychology, H. J. Eysenck ed., London: Pitman med., 2nd ed., 261-283, 1973.

Zeidenberg, P., W. C. Clark, J. Jaffe, S. W. Anderson, S. Chin and S. Malitz. Effects of oral administration of Δ q tetrahydrocannabinol on memory, speech and perception of thermal stimulation: Results with four normal volunteer subjects. Preliminary report, Compr. Psychiat., 14/ 6, 549-556, 1973.