Abstract
Introduction
A WORKING HYPOTHESIS et seq.
Findings and discussion
Summary
Bibliography
Author: M. I. SOUEIF
Pages: 25 to 42
Creation Date: 1976/01/01
The present study was done in order to investigate the following general hypothesis: "Other conditions being equal, the lower the non-drug level of proficiency on tests of cognitive and psychomotor performance the smaller the size of function deficit associated with drug taking". Twelve objective tests generating 16 test variables were administered to 850 chronic cannabis takers and 839 comparable non-takers, all males ranging in age between 15 and 50 years and representing various positions on "literacy-illiteracy" and "urbanism-ruralism". Six specific predictions were formulated: (1) Performance on the test is expected to be correlated with the level of literacy of the subject. (2) the lower the level of literacy the smaller the size of function deficit associated with cannabis taking. (3) Performance on the tests is expected to be correlated with the level of urbanism of the subject. (4) The lower the level of urbanism (i.e. more ruralism) the smaller the size of function deficit associated with drug taking. I (5) Performance on the tests is expected to correlate inversely with age. (6) The older the subject the smaller the amount of function deficit associated with cannabis consumption. All our predictions were confirmed. The consistency of these results raises the possibility that a basic regularity governing the relationship between cannabis consumption and psychological-function deficit is being uncovered. The suggestion was made that our major hypothesis might prove valid with regard to the area of acute effects of cannabis taking as well. Our hypothesis was shown to provide a broad framework capable of integrating a number of conflicting reports both in the area of long term and that of short term cannabis use.
One of the key findings recurring throughout the study of objective test results obtained by 850 chronic cannabis takers and 839 comparable non-takers, was the fact that certain combinations between drug-taking, level of literacy and residence (viz. position on a continuum of urbanism-ruralism) revealed, better than others, the main drug effect in the majority of our test variables (Soueif, 1975). The present paper will elaborate on this finding, suggesting that here we may have a pointer to a fundamental rule regulating the emergence of a significant association between cannabis use and function deficit. 1
1The quantitative determination of Δ9-THC in cannabis consumed in Egypt, was carried out by Dr. Z. I. El-Darawy and Z. M. Mobarak at the laboratories of the National Centre for Social and Criminological Research (in Cairo). The average percentage of the active principle in seven different samples of hashish seized on the illicit market was found to be 3.04 per cent by weight.
Our group of takers were all males, ranging in age from 15 to slightly over 50 years, with an average of 39 ± 10.5. They were all prison immates who were incarcerated on the charge of possessing cannabis for the explicit purpose of taking it. The number of users kept in urban prisons catering for urban offenders was 460, while 390 subjects were villagers detained in rural prisons. We found a fairly good agreement between such distribution and the distribution of the subjects according to locale of upbringing. About 70 per cent of the former group were born and brought up in big cities, and almost all members of the group had lived for a rather long time (before their arrest) in urban districts. On the other hand 95 per cent of the rural group reported that they have been brought up in small villages, in Lower or Upper Egypt. Sixty per cent of our takers were illiterate and the rest were distributed among various levels and kinds of formal education, with only 6 subjects attaining the high school level. The family background presented almost the same picture, at least regarding the male members of the family. Twenty-six per cent of the group were classified as skilled labourers, 17.8 per cent shopkeepers, 1 per cent civil servants, 3 per cent some odd jobs and the rest were unskilled labourers. As to physical and mental health 15.8 per cent stated that they had some physical illness before their imprisonment and 5.5 per cent said they had had some psychiatric complaint.
The controls were all males, with an age range between 15 and about 50 years, and an average of 33±9.75 years. They were selected from the same prisons where our users were located. Non-takers derived from urban prisons was 454, and the rest were from rural ones. The agreement between that distribution and the distribution of subjects according to locale of upbringing was, however, not satisfactory. Only 45 per cent of the urban prisoners maintained that they were brought up in big cities, whereas the rest spent their childhood in small towns and/or in villages. Nevertheless, 94 per cent of the rural controls admitted having a rural background of upbringing. One may, therefore, conclude that there was less urbanism in controls than there was in drug takers. The percentage of illiterate was 54.8, and the rest had various periods of schooling, starting from one year or two up to university level. We had among the controls 44 subjects who reached the high school level and 9 persons obtaining university degrees. As to the family background 63.7 per cent of the controls' fathers were illiterate. 26.4 per cent were categorized as skilled labourers, 5.7 per cent shopkeepers, 3.5 per cent civil servants, 1.5 per cent students and the rest unskilled labourers. Regarding physical and mental health 11.7 per cent said they had some physical ailment before imprisonment and 4.0 per cent reported having had some psychiatric trouble.
Obviously takers and non-takers seem comparable to each other on most of the mentioned variables. Controls are slightly better than users on literacy, but this seems counterbalanced by the fact that the latter are higher on urbanism.
1. The following is proposed as a working hypothesis: The size of psychological deficit likely to be associated with chronic cannabis taking in the individual is a function of his general level of proficiency without drug-taking (presumably ascertainable through comparable non-takers). Other conditions being equal, the lower the non-drug level of proficiency the smaller the size of the deficit.
2. Before proceeding to an empirical examination of this broad hypothesis, a number of facts relating to some "ordinary determinants" of the level of proficiency have to be mentioned. In this article three such determinants will be discussed: literacy, residence and age. Similar other determinants may be pointed out by future investigators. The expectation is, that such determinants will yield, then, patterns of results that bear ressemblance to what will be presented here
2.1 A recurring finding in the literature on psychological testing is that the majority of those tests designed to assess psychomotor and/or cognitive abilities are highly correlated with the level of literacy of the individual. Wechsler puts it clearly that, "Practically all studies show that educational attainment (as measured by the number of years of school attendance) and intelligence ratings (as measured by test scores) correlate to a relatively high degree". (Wechsler 1954, p. 103.) What is true of intelligence ratings can, to some extent, be extended to include tests of general abilities. The Wechsler intelligence scale for adults includes a number of subscales which tap various aspects of perceptual and motor functioning. It follows that the level of literacy can be taken as a predictor of the level of proficiency on tests of abilities. According to our broad hypothesis, the amount of deficit (as demonstrated on such tests) expected to be associated with cannabis taking would be a function of the individual's level of literacy.
2.2 While carrying out our work in the rural areas we were constantly struck by what may be labelled the "hypo-responsivity" of the rural subjects (prisoners as well as non-prisoners), in contrast with the "hyper-responsivity" of urban subjects. In a country like Egypt, with a wide cultural gap between town and country, such a contrasting effect could be more impressive than in most Western countries. (Berger 1962, of. chapter 3). We could detect the hypo-responsivity of our villagers in various aspects of behaviour, e.g. slowness of gross and fine motor activity such as walking and handling equipment used in various crafts; long reaction time in starting to answer questions or to comment on some idea posed by the speaker; a tendency to repeat a single word or few words every few sentences; long intervals of silence between phrases or small units of discourse, and slow utterance of single words. True, such manifestations may appear in an exaggerated form when villagers are communicating with a stranger from the city, but certainly the same characteristics can also be noticed in social situations where rural persons are interacting with their fellow villagers. Those observations are in line with various hints made by social anthropologists who studied non-urban societies (e.g. Hallowell, 1949). Another kind of research which we found congenial with our remarks was the work on identical twins reared apart. Regardless of the standpoint adopted by the student as to the controversy between environmentalists and defenders of a genetic determination of behaviour, a definite association has been noted by various investigators between low scores on various tests of abilities and being brought up in rural areas (Newman et al., 1937; Newman, 1947). Commenting on that finding, Shields, who has done a good deal of work in this and related areas, maintained the following: "Of course, if one were to study only selected pairs of twins where the environment of one was grossly 'enriched' and that of the other grossly 'restricted' it would be surprising if they were still found to be alike" (Shields, 1973). Relevant here is the famous study carried out by Klineberg in the early decades of this century. Klineberg tested, by means of the National Intelligence Test, 425 twelve-year old Negro boys in New Orleans.
Of these boys, about 165 had been born in New Orleans, the others had come to the city at various ages since birth. He separated the subjects in the following groups: those who had been in the city less than one year; those who had been in the city between one and two years; between two and three years; three and four years; four and five years; and five and six years and those who had been in the city between six and twelve years. Now the average National Intelligence Test scores of these twelve-year-old groups, when thus arranged, show an almost perfect straight line of increasing intelligence, from the score of about 40 made by those who had been in the city less than a year, to a score of about 75 made by those born in the city. (Murphy et al., 1937, p. 64.)
Guided by such considerations we would suggest that the level of urbanism-ruralism be taken as a predictor of the level of proficiency on tests of psychomotor and cognitive abilities. Thus, urban persons would be expected to earn high scores, rurals low scores and semi-rurals in between. It would also follow (according to our general hypothesis) that the magnitude of deficit expected to be associated with cannabis taking would be a function of the individual's level of urbanism.
2.3 Chronological age has always been a concern for students of human behaviour whenever administering tests of abilities to adults (Bromley, 1966). There is good evidence that psychomotor performance slows down with age. Experimental psychologists are used to identify three factors involved in speed of movement: (a) speed of initiating a movement, that is reaction time, (b) speed of repetitive movement, for instance tapping ability, (c) speed and co-ordination of complex movements (Hicks and Birren 1970). The prevailing impression is that at least factors (a) and (c) are most vulnerable to age effect. Information about factor (a) has been known since the work of Bellis (1933), confirmed recently by Simon (1967); that about factor (b) since Miles carried out his early study (1931), which was confirmed later by Pierson and Montoy (1958). Interestingly enough the association between slow performance and old age seems to be a biologically rooted phenomenon, so much so that it could be demonstrated in rats (Birren and Kay, 1958), a fact which can be creatively exploited by students interested in cannabis research. Accuracy of psychomotor functioning (partly involved in factor (c)), is also affected by age. Savage puts it clearly that, "There is considerable evidence, that perceptual motor functioning declines with age..." (Savage, 1973). This has been shown in a number of studies where examinees were required to copy, to trace or to draw out of memory, a number of geometric designs (e.g. Benton Visual Retention Test; Perceptual Maze Test) (Benton, 1963; Elithorn, 1955).
The negative correlation between the efficiency of immediate and short term memory on one hand and old age on the other has also, been recognized for a long time. Indeed this fact presented clinicians, who cared to use standardized tests of memory, with a rather complex problem that had to be solved before they ventured to diagnose psychiatric illness in old age (Hulika, 1966).
By and large, therefore, the age factor cannot be ignored when using tests of abilities (psychomotor and/or cognitive) with adult subjects. Basing ourselves on such findings and relevant extrapolations, the suggestion is made that age be taken as a predictor of proficiency on tests of psychomotor and cognitive abilities. Young adults would be expected to earn higher scores on such tests than old persons. Hence, it would follow (according to our broad hypothesis) that the size of deficit expected to be associated with cannabis intake would be a function of age.
3. As a further step towards an empirical examination of our working hypothesis six specific predictions were formulated:
3.1 We would expect performance on psychomotor and cognitive tests to be significantly correlated with the position occupied by the subject on the continuun of "literacy-illiteracy"
3.2 The lower the level of literacy the smaller the size of function defici associated with cannabis taking.
3.3 We would expect performance on psychomotor and cognitive tests to be significantly correlated with the position occupied by the subject on the continuun of "urbanism-ruralism".
3.4 The nearer the subject's position to the ruralism pole the smaller the size of function deficit associated with cannabis taking.
3.5 We would expect performance on our psychomotor and cognitive test to be significantly (but inversely) correlated with age.
3.6 The older the subject the smaller the amount of function impairment associated with cannabis consumption.
Table 1 presents the outcome of the analysis of variance of scores obtained on 16 test variables by our subjects classified as representing three levels of literacy regardless of drug use. Except for "distance over-estimation", all test variable differentiated at a very high level of significance between examinees. As to the direction of such differentiation we found that those who were high on literacy get the best scores on the tests, the illiterates the worst scores while the semi-literate tended to fall in between (Soueif, 1975, tables 2 to 4).
Test variables |
F |
|
---|---|---|
1.
|
Tool matching
|
243.00 |
2.
|
H marking
|
240.00 |
3.
|
Speed
|
1036.10 |
4.
|
Mark making
|
268.00 |
5.
|
Trail making (Part A)
|
50.92 |
6.
|
Initial reaction time
|
10.89 |
7.
|
Distance overestimationa
|
0.35 |
8.
|
Distance underestimation
|
62.14 |
9.
|
Distance estimation: discrepancy
|
37.98 |
10.
|
Time estimation
|
6.29 |
11.
|
Time estimation: discrepancy
|
11.55 |
12.
|
Digits forward
|
156.80 |
13.
|
Digits backward
|
1778.00 |
14.
|
Digits forward and backward
|
232.60 |
15.
|
Bender Gestalt: copy
|
357.00 |
16.
|
Bender Gestalt: recall
|
267.00 |
a Except for distance overestimation P was beyond 0.00l level of significance in all cases.
Table 1 confirms our first prediction and strongly supports our comment on Wechsler's contention as to the positive correlation between educational attainment and intelligence ratings.
In table 2 are presented t tests between median scores obtained by takers and non-takers within each level on the literacy continuum. We have used medians, instead of means, as measurements of central tendency, to avoid the possibility of capitalizing on any deviations of the distributions of test scores from normality. Inspection of table 2 shows clearly a trend towards less significant differences between users and non-users, the more we move from the literate to the illiterate groups.
The same finding is presented graphically in figure I. In the literate groups 10 test variables differentiated between takers and non-takers, mostly at very high levels of significance. In the semi-literates, only 5 variables differentiated reliably in the same direction. Among the illiterates, the number of reliable differentiators was reduced to 3 only and paradoxically enough the three variates discriminated between users and non-users in the unexpected direction.
One could, therefore, conclude that no function deficit was correlated with cannabis consumption in the illiterate groups. We do not have a ready made explanation for the mentioned paradoxical result concerning illiterates. So until such an explanation can be provided, we may keep concentrating on our major finding, viz. the trend towards less significant differences between users and non-users correlating with less literacy. This finding is a confirmation of our second prediction.
Test variables |
Literates |
Semi-literates |
Illiterates |
|
---|---|---|---|---|
1.
|
Tool matching
|
3.19a | 3.78a | 1.10 |
2.
|
H marking
|
4.27a | 1.97b | 1.61 |
3.
|
Speed
|
1.84c | 1.70c | 1.11 |
4.
|
Mark making
|
3.22a | 1.50 | 1.42 |
5.
|
Trail making (Part A)
|
2.80a | 1.38 | 2.60b |
6.
|
Initial reaction time
|
0.44 | 1.05 | 1.30 |
7.
|
Distance overestimation
|
0.90 | 0.40 | 0.54 |
8.
|
Distance underestimation
|
2.68a | 1.55 | 1.54 |
9.
|
Distance estimation: discrepancy
|
1.40 | 0.76 | 1.36 |
10.
|
Time estimation
|
0.53 | 1.19 | 3.00b |
11.
|
Time estimation: discrepancy
|
0.36 | 0.06 | 1.53 |
12.
|
Digits forward
|
2.71a | 0.17 | 2.41c |
13.
|
Digits backward
|
2.37b | 0.44 | 0.21 |
14.
|
Digits forward and backward
|
3.38a | 0.05 | 1.17 |
15.
|
Bender Gestalt: copy
|
1.41 | 5.81a | 0.43 |
16.
|
Bender Gestalt: recall
|
2.60a | 5.17a | 1.09 |
NOTE. One tail test. (N.B. In the illiterates P was based on a two tailed test because the difference was in the unexpected direction.)
a P > 0.001.
b P > 0.01.
c P > 0.05.
5. Table 3 shows the results of the analysis of variance of test scores obtained by our subjects classified as representing 3 positions on the continuum of urbanism irrespective of drug taking. In all cases F ratios were highly significant, either at or beyond 0.001 level of confidence. (For the direction of the differences of. Soueif, 1975.) On the whole, other conditions being equal, urbans earn high scores, rurals low scores and semi-rurals in-between. This result confirms our third prediction and adds support to similar previous findings (whether such findings were psychological or socio-anthropological).
Table 4 presents t tests between median scores obtained by users and controls within each category along the continuum of urbanism. In this table readers can readily detect a trend towards less significant disparities between cannabis takers and non-takers the more we move from the urban through the rural groups. The same fact is depicted graphically in figure II. In the urban groups takers were differentiated, mostly at high levels of statistical significance, along 8 test variables; in the semi-rurals the number of differentiators came down to 7, and in the rurals it was reduced to 3. One of the test variables differentiated between rural takers and non-takers in the unexpected direction, an odd result for which we have no explanation at present. Apart from that odd point, the principal finding, as disclosed in table 4, confirms our fourth prediction.
Test variables |
Fa |
|
---|---|---|
1.
|
Tool matching
|
73.30 |
2.
|
H marking
|
120.00 |
3.
|
Speed
|
81.90 |
4.
|
Mark making
|
193.60 |
5.
|
Trail making (Form A)
|
17.86 |
6.
|
Initial reaction time
|
16.46 |
7.
|
Distance overestimation
|
9.99 |
8.
|
Distance underestimation
|
13.91 |
9.
|
Distance estimation: discrepancy
|
18.29 |
10.
|
Time estimation
|
6.43 |
11.
|
Time estimation: discrepancy
|
18.44 |
12.
|
Digits forward
|
33.56 |
13.
|
Digits backward
|
5.20 |
14.
|
Digits forward and backward
|
24.60 |
15.
|
Bender Gestalt: copy
|
18.65 |
16.
|
Bender Gestalt: recall
|
47.60 |
a P > 0.001 in all cases.
Test variables |
Urbans |
Semi-rurals |
Rurals |
|
---|---|---|---|---|
1.
|
Tool matching
|
2.11a | 3.08b | 0.63 |
2.
|
H marking
|
4.0lb | 3.49b | 0.42 |
3.
|
Speed
|
1.73e | 2.89b | 1.98c |
4.
|
Mark making
|
4.13b | 2.59b | 0.93 |
5.
|
Trail making (Part A)
|
0.45 | 2.49a | 1.19 |
6.
|
Initial reaction time
|
2.24a | 0.87 | 0.06 |
7.
|
Distance overestimation
|
0.24 | 0.34 | 0.66 |
8.
|
Distance underestimation
|
1.46 | 0.60 | 3.99b |
9.
|
Distance estimation: discrepancy
|
2.60a | 0.36 | 0.07 |
10.
|
Time estimation
|
0.73 | 0.19 | 0.27 |
11.
|
Time estimation: discrepancy
|
0.98 | 0.03 | 0.44 |
12.
|
Digits forward
|
0.24 | 1.21 | 1.59 |
13.
|
Digits backward
|
0.70 | 0.70 | 0.57 |
14.
|
Digits forward and backward
|
1.50 | 1.00 | 1.00 |
15.
|
Bender Gestalt: copy
|
1.42 | 1.86c | 2.05a |
16.
|
Bender Gestalt: recall
|
3.46a | 1.84c | 2.33a |
NOTE. One tail test (except in the case of test No. 3 in rurals, because the observed difference was in the unexpected direction).
a P > 0.01.
b P > 0.001.
c P > 0.05.
6. To test predictions No. 5 and No. 6 we decided to limit all the relevant comparisons to two age groups: subjects below 25 years and those above 35 years. This decision was made to permit optimum clarity for the contrast effect between groups compared.
6.1 Table 5 presents tests of significance between median scores obtained on the objective tests by young and old controls. On all the variables the young subjects earned better scores than the old testees. The differences between the two groups reached a satisfactory level of statistical significance on 9 out of the 16 test variables, thus substantiating our fifth prediction. The reason why we did not do a similar comparison within takers was that it proved rather difficult to find two sizable subgroups who would differ in age but not in duration of cannabis consumption (a variable which cannot be ignored in this respect).
6.2 To test prediction No. 6 we had to accept the fact that we could not equate the compared takers (representing two age groups) regarding duration of drug taking. This problem seemed a methodological impasse; duration was, to a great extent, a function of age since the majority of cannabis users get initiated into their drug behaviour at rather an early age (Committee, 1964; Soueif, 1967, 1971).
Test variables |
t |
|
---|---|---|
1.
|
Tool matching
|
6.04 |
2.
|
H marking
|
6.03 |
3.
|
Speed
|
4.86 |
4.
|
Mark making
|
5.06 |
5.
|
Trail making (Part A)
|
2.75 |
6.
|
Initial reaction time
|
0.95 |
7.
|
Distance overestimation
|
0.37 |
8.
|
Distance underestimation
|
0.40 |
9.
|
Distance estimation: discrepancy
|
1.38 |
10.
|
Time estimation
|
0.23 |
11.
|
Time estimation: discrepancy
|
0.45 |
12.
|
Digits forward
|
3.46 |
13.
|
Digits backward
|
1.26 |
14.
|
Digits forward and backward
|
2.33 |
15.
|
Bender Gestalt: copy
|
3.49 |
16.
|
Bender Gestalt: recall
|
6.77 |
N.B. To reach 0.05 level of confidence (one tail test) t should be about 1.70 at least.
Test variables |
Takers:age - 25 years Duration: - 5 years Frequency: 30+ a mth |
Takers: age 35 - 50 years Duration: 20+ years Freuuency: 30+ a mth |
|
---|---|---|---|
vs.
Controls: age - 25 years
|
Controls: age 35 - 50 years
|
||
1.
|
Tool matching
|
2.44a | 1.16 |
2.
|
H marking
|
4.99b | 0.82 |
3.
|
Speed
|
2.83b | 2.11a |
4.
|
Mark making
|
2.54a | 0.27 |
5.
|
Trail making (Part A)
|
0.17 | 0.29 |
6.
|
Initial reaction time
|
0.21 | 0.22 |
7.
|
Distance overestimation
|
0.24 | 0.44 |
8.
|
Distance underestimation
|
4.04b | 0.27 |
9.
|
Distance estimation: discrepancy
|
4.08b | 0.05 |
10.
|
Time estimation
|
0.44 | 0.02 |
11.
|
Time estimation: discrepancy
|
0.60 | 0.29 |
12.
|
Digits forward
|
1.36 | 2.13a |
13.
|
Digits backward
|
0.62 | 0.20 |
14.
|
Digits forward and backward
|
0.99 | 2.87b |
15.
|
Bender Gestalt: copy
|
1.54 | 1.91e |
16.
|
Bender Gestalt: recall
|
2.44a | 1.23 |
NOTE. One tail test. a P > 0.01. b P > 0.001. c P > 0.05.
Since we were trying to compare young takers with old, longer duration was almost automatically contingent with being older. Our two groups of takers were, however, equated regarding frequency of drug use.
Table 6 shows t tests between median scores obtained by takers and non-takers within each one of two age groups, below 25 and above 35 years. Note-worthy is the fact that for the younger group of takers the duration of their drug experience was less than 5 years, whereas the older takers had a taking age of at least 20 years. In both groups the frequency of drug taking was above 30 times a month. Inspection of table 6 reveals a definite trend towards less significant differences between takers and non-takers in the older group of subjects; in other words less function deficit came out as associated with drug taking in the older subjects. Whereas 7 tests differentiated significantly between young takers and non-takers, the older groups were reliably differentiated on 4 tests only. The same trend is presented graphically in figure III.
Table 7 is similar to table 6 except for the frequency of drug taking, which, in this case, is below 30 times a month. Table 7 shows the same trend revealed in table 6; less significant differences between older takers and controls (on 2 tests variables only), than between younger corresponding groups (on 7 test variables).
Test variables |
Takers: age --25 years Duration: -5 years Frequency: 30 or below |
Takers: age 35--50 years Duration: 20+ years Frequency: 30 or below |
---|---|---|
vs.
Controls: age --25 years
|
vs.
Controls: age 35-50 years
|
|
1. Tool matching
|
1.91a | 2.44b |
2. H marking
|
1.95a | 0.96 |
3. Speed
|
1.71a | 0.20 |
4. Mark making
|
1.61 | 1.75a |
5. Trail making (Part A)
|
4.70c | 1.11 |
6. Initial reaction time
|
0.34 | 0.28 |
7. Distance overestimation
|
0.50 | 0.05 |
8. Distance underestimation
|
0.92 | 0.48 |
9. Distance estimation: discrepancy
|
2.13b | 0.44 |
10. Time estimation
|
0.26 | 0.04 |
11. Time estimation: discrepancy
|
0.76 | 0.03 |
12. Digits forward
|
1.04 | 1.37 |
13. Digits backward
|
0.58 | 0.72 |
14. Digits forward and backward
|
0.91 | 1.29 |
15. Bender Gestalt: copy
|
4.14c | 0.90 |
16. Bender Gestalt: recall
|
3.35c | 0.41 |
NOTE. One tail test. a P > 0.05. b P > 0.01. c P > 0.00l.
Figure IV is a graphic representation of the same fact. Obviously tables 6 and 7 substantiate our sixth prediction.
7. The general conclusion to be drawn is that all our predictions were confirmed. This gives substantial support to our initial working hypothesis, to the effect that "the lower the non-drug level of proficiency the smaller the function deficit to be associated with drug taking".
7.1. To be sure our conclusion is based on the detection of a trend that makes itself faintly (though consistently) perceptible through our analyses. Perhaps this trend might have made a more impressive appearance, if we had been willing to drop the groups of semi-literates and semi-rurals when testing our second and fourth predictions, thus restricting our comparisons to the extreme poles of the relevant continua (the same as we did in testing prediction No. 6 about age). It has been justly said, "that no single experiment can establish the absolute proof of any result, however significant the results may happen to be" (Edwards, 1956, p. 30). The gist of such dictum is that research workers should be wary of imprudently accepting as a fact what may turn out, on careful replication, to be just an illusion; a mistake which methodologists used to call "Type I error", It is not a less serious flaw, however, to make a "Type II error", namely to brush aside (as chimera) a true regularity which may reveal an important fact or a basic relationship, if we continue to examine closely the conditions of its emergence.
The differential association between the size of function deficit and chronic cannabis consumption, which we have been trying to delineate, seems to be one of those low relief regularities which may prove generously rewarding. At the present stage of its development, however, our hypothesis seems to raise more questions than answers. For one thing, it would be ill-advised to generalize such a formula to other samples of subjects (e.g. females) and other cultural settings on an a priori basis. Cross-validating studies are needed to define the limits beyond which the formula should not be extended. And, would the hypothesis stand a direct empirical examination? Supposing a prespective study was conducted on a cohort of subjects, who were followed up until they got differentiated into confirmed cannabis takers and non-takers. Would it be possible to establish the same relationship between the magnitude of psychomotor and cognitive function deficit they display, then, and their pre-drug level of proficiency? One way of carrying out such a project is to select a group of chronic drug takers who have already been tested for their basic abilities in their teens, before taking to cannabis (say on their being called up to serve their term in the military forces, or on admission to some sort of apprenticeship or some kind of school). By gauging those subjects on the same test variables which provide their pre-drug level of proficiency the investigator will be in a position to examine directly the above-mentioned hypothesis. A third question would be concerning the possibility of a transposition of the hypothesis to the domain of enquiry into the acute effects of cannabis consumption. The hypothesis should read, then, as follows: The amount of impairment that is likely to be effected by the drug is a function of the level of pre-drug proficiency: other conditions being equal, the lower the initial level the less impairment. A fourth question would be inviting an explanation of the phenomenon, a step beyond mere statistical description; why do we get measurable impairment in cannabis takers who are urban, literate and young, but not in other takers who are rural, illiterate and old? Numerous other questions are posed by the same broad hypothesis. They may not be as meaningful as those we have cited, but not necessarily less important.
7.2. One of the merits of field research, of the kind presented here, is that the investigator can cope with a large number of subjects, a fact which, more often than not, permits teasing out important, though subtle, patterns of regularity. Because human laboratory experiments are done under many constraints that are part and parcel of their nature, such experiments cannot handle big numbers of subjects. Tinklenberg, who has himself done a number of laboratory experiments puts it clearly as follows:
The scientific consideration of sampling may limit laboratory studies and necessitate field investigation. the research sample should represent all features of the phenomenon to be studied in concentrations appropriate to the total population itself. The ideal situation is seldom obtained and is particularly difficult to achieve in laboratory settings, in which subjects are rarely representative of the total population of drug users or perhaps more importantly, of the relatively small number of individuals for whom social control of drugs is mandatory. (Tinklenberg, 1974).
Field work, therefore, might do a scouting job which can provide laboratory studies with useful insights into the optimum conditions for a design of experiment to give meaningful results.
It might be a good exercise in the methodology of cannabis research (and possibly in drug research in general) to speculate on what kind of results we might have got had we been concentrating on a group of subjects who were all illiterates, rurals and/or rather old. "No significant differences between takers and non-takers" would have been, in all probability, the main finding. On the other hand, had we selected our testees from among literates, urbans and/or young adults solely, the opposite finding would have been reported without a hint of the other part of the story. It is possible that this is the case behind some empirical studies reported in the last few years.
Bowman and Pihl used a whole battery of objective tests in two field studies conducted in Jamaica. In the first study, they administered the tests to 16 users and 10 controls who were drawn from rural or semi-rural areas. In the second study, the investigators gave the tests to 14 users and 14 controls derived from the slum areas of one city. The results of the two studies showed no significant differences between takers and controls (Bowman and Pihl, 1973). It should be noted that the majority of the examinees in the two studies were either illiterate or semi-literate, and their average age was 30±5 years. In the light of our broad perspective, Bowman and Pihl were bound to end up with negative results.
It is interesting to note that some conflicting reports on the acute effects of cannabis taking can be, logically, reconciled with each other through our broad hypothesis (if we use the transposed version suggested for the area of short-term effects). The two reports, one by Waskow and colleagues reporting no significant differences (1970), and the other by Melges and associates reporting reliable differences between users and non-users (1970) provide almost a classic example for the case we are presenting. What is truly instructive about those two studies is that both used a double blind experimental approach, yet they ended up with two mutually contradictory results. Obviously the double blind approach did not automatically provide protection against all sources of bias. The main reason behind their conflicting findings lay, in all probability, in the fact that the two groups of researchers tested two different types of subjects. The Waskow group examined 32 men who were all criminal offenders with an average I.Q. of 95, and average eighth grade educational level. The Melges group tested 8 normal male graduate students. Unfortunately the latter group did not specify the average I.Q. of their subjects. Moreover, we were not told about the age distribution in both studies. However, we can, safely, assume that the graduate students belonged to the group of bright normals (according to Wechsler's classification), which means they had an I.Q. of, at least, 110. Because tests of cognitive abilities of the kind used in the two studies are known to be significantly correlated with the level of education and with the intelligence quotient too, it is conceivable that the subjects' levels of pre-drug proficiency would differ in the two Studies, being appreciably lower in the Waskow than in the Melges study. Following our hypothesis, the discrepancy between the findings reported in the two studies would, thus, be naturally expected; subjects with low pre-drug level of proficiency (in the Waskow study) displayed minimal function deficit contingent with THC taking, while those with high pre-drug ability level (in the Melges study) showed impairment of a sizable magnitude. The two studies, thus, seem to complement each other in showing the role of some organismic variables (i.e. "characteristic ways in which the particular group or organisms under observation vary", of. Edwards, 1956, p. 7) in determining the effects of cannabis ingestion in man.
Interestingly enough, Melges and his colleagues (this time working under the senior authorship of Tinklenberg) did another study in which they gave cognitive tests (more or less similar to those used in their previous work) to 15 college-educated men in their twenties to assess the immediate effects of cannabis ingestion. Again they confined themselves to the double blind design of experiment (1972). However, they ended up with no significant effects of cannabis consumption on two out of the three tests they used, a result which was in disagreement with what they themselves had reported in 1970. Those investigators, in their attempt to shed light on the disparity between their two consecutive studies stated the following:
This discrepancy might be explained by the principle of initial values--that is, the subject group that was initially more proficient might have been more prone to show impairments in cognitive operations during marihuana intoxication. The group of subjects used in this experiment were not as proficient on their initial baseline cognitive tests as the subjects we used in a previous study of marihuana intoxication. (Tinklenberg et al., 1972).
In all those investigations the small numbers of examinees (rendering impossible an exploratory breakdown of the samples to discover meaningful associations), and the homogeneity of the samples with respect to relevant organismic variables (such as literacy, urbanism and age) resulted in those narrowly focused reports that would not permit an integrated view of our main finding, viz. the differential pattern of association between cannabis taking and possible function deficit.
It is tempting to bring into focus, together with such a finding, a seemingly similar one though reported in another field of investigation. Reviewing the work done, from 1960 to 1970 on intellectual deterioration under conditions of mental illness, Payne found that he was faced with two groups of studies: group A, which concentrated on patients who had, originally, above average or average I.Q. Those patients showed significant deterioration. And group B, which were concerned with patients who were initially subnormal. Those patients displayed very little deterioration. Under group A, Payne cites three studies by Lubin et al., 1962; Kingsley and Struening, 1966; and Schwartzman and Douglas, 1962. Under group B, are cited Batman et al., 1966; Albee et al., 1963; and Griffith et al., 1962. Commening on the divergence between the two kinds of reports, Payne gave the following hint: "It is possible to speculate that the dull who had become psychotic deteriorate little intellectually" (Payne, 1973).
A more or less similar basic paradigm, thus, seems to account for possible function impairment contingent with mental illness or cannabis consumption.
Stretching the paradigm to account for the social history of the drug, would it be the case that cannabis taking was believed to be benign (or not really harmful), throughout its past, not only because rigorous scientific research was lacking then, but also because the drug was mainly taken by the illiterate and the rural (actually displaying little if any visible impairment) within the context of truly debilitating socio-cultural conditions? (One has only to think of the social conditions prevailing in countries like Egypt, India and South Africa from the 19th century backwards). This might explain why the famous Indian Hemp Drugs Commission (1894) came to the conclusion that: "In respect to the alleged mental effects of the drug, ... the moderate use of hemp drugs produces no injurious effects on the mind..." (Interim Report 1970, p. 99). The same paradigm may, also, explain the fact that, even as far back as the twelfth century, some controversies were aroused between writers as to the main effect of the drug: harmless or harmful? Two Arab writers who lived (in Egypt and Syria) around late twelfth and early thirteenth centuries, Ibn-el-Saigh and Al-Quadi-el-Fadel, represented the two poles of the debate. Perhaps Ibn-el-Saigh, who defended the drug, had his attention selectively focused on the illiterate or the under-priviledged taker, while A1-Quadi-el-Fadel was interested in the sophisticated user (Soueif, 1972; Hussein, 1957).
To summarize, a working hypothesis has been presented to the effect that, "other conditions being equal, the lower the non-drug level of proficiency on tests of cognitive and psychomotor performance the smaller the size of function deficit associated with drug taking". For the empirical examination of the hypothesis six predictions were formulated. Three predictions defined specific relationships between levels of performance on one hand, and each of three organismic variables on the other: literacy, urbanism and age. The remaining predictions delineated relationships to be expected between size of function deficit and the three organismic variables. All our predictions were confirmed showing less function impairment to be contingent with cannabis taking among the less sophisticated and/or the older subjects.
Since the version presented of our main hypothesis was formulated with reference to chronic material, a reformulation of a new version (keeping the same Determinants of psychological deficits associated with chro nic cannabis consumption 41structure), judged to be more suitable to research in the acute effects of the drug, was attempted. The suggestion was made that our major hypothesis, in either form, is capable of establishing logical reconciliation (or indeed genuine integration) between reports presenting conflicting results on possible function deficits contingent with cannabis consumption.
Bellis, C. J. Reaction time and Chronological age, Proceedings of the Society for Experimental Biology and Medicine, 30, 801-803, 1933.
Benton, A. L. The Revised Visual Retention Test, 3rd ed., Iowa City: Iowa State Univ., 1963.
Berger, M. Arab world today, New York: Doubleday, 1964.
Birren, J. E. and H. Kay. Age and sex differences in swimming speed of the albino rat, J. Gerontology, 13, 374-377, 1958.
Bowman, M. and R. O. Pihl. Cannabis: Psychological effects of chronic heavy use, Psychopharmacologia (Berlin), 29, 159-169, 1973.
Bromley, D. B. The psychology of human ageing, Harmondsworth: Penguin Books, 1966.
Committee for the Investigation of Hashish Consumption in Egypt, 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).
Edwards, A. Experimental design in psychological research, New York: Rinehart, 1956.
Elithorn, A. A preliminary report on a perceptual maze test sensitive to brain damage, J. neurol. neurosurg. Psychiat., 18, 287-292, 1955.
Hallowell, A. I. Acculturation processes and personality changes, Personality in nature, society and culture, C. Cluckhohn and H. A. Murray eds., New York: A. Knopf, 1949.
Hicks, L. H. and J. E. Birren. Ageing, Brain damage and psychomotor slowing, Psychol. Bull., 74, 377-396, 1974
Hulicka, I. M. Age differences in Wechsler Memory Scores, J. Genet. Psychol., 109, 135-145, 1966.
Hussein, M. K. Studies in poetry during the Ayyubid dynasty, Cairo: Dar-al-Fikr-al-Arabi, 1957 (in Arabic).
Interim Report of the Commission of Inquiry into the non-medical use of drugs, Ottawa: Information Canada, 1970.
Melges, F. T., J. R. Tinklenberg, L. E. Hollister and H. K. Gillespie. Marihuana and temporal disintegration, Science, 168, 1118-1120, 1970.
Miles, W. R. Correlation of reaction and coordination speed with age in adults, Amer. J. Psychol., 43, 377-391, 1931.
Murphy, G., L. B. Murphy and T. M. Newcomb. Experimental social psychology, New York: Harper, 2nd. ed. 1937.
Newman, H. H., F. N. Freeman and K. J. Holzinger.Twins: A study of heredity and environment, Chicago: Univ. Chicago Press, 1937.
Newman, H. H. How differences in environment affected separated one-egg twins, Readings in social psychology, T. M. Newcomb et al. eds., New York: Holt 1947.
Payne, R. W. Cognitive abnormalities, Handbook of abnormal Psychology, H. J. Eysenck ed., London: Pitman Med., 420-483, 1973.
Pierson, W. R. and H. J. Montoye. Movement time, reaction time and age, J. Gerontology, 13, 418-421, 1958.
Savage, R. D. Old age, Handbook of abnormal psychology, H. J. Eysenck ed., London: Pitman Med., 645-688, 1973.
Shields, J. Heredity and psychological abnormality, Handbook of abnormal psychology, H. J. Eysenck ed., London: Pitman Med., 540-603, 1973.
Simon, J. R. Ear preference in a simple reaction time task, J. Exper. Psychol., 75, 49-55, 1967.
Soueif, M. I. Hashish consumption in Egypt, with special reference to psychsocial 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.
_____ . The social psychology of cannabis consumption: myth, mystery and fact, Bulletin on Narcotics, XXIV:2, 1-10, 1972.
_____ . Chronic cannabis users: Further analysis of objective test results, Bulletin on Narcotics, XXVII: 4, 1-26, 1975.
Tinklenberg, J. R., B. S. Kopell, F. T. Melges and L. N. Hollister. Marihuana and alcohol, Arch. Gen. Psychiat., 27, 812-815, 1972.
_____ .Laboratory and field investigation, Controlling drugs, R. Blum, D. Bovet, J. Moore and others, San Francisco: Jossey-Bass, 211-224, 1974.
Wechsler, D. The measurement of adult intelligence, Baltimore: Williams and Wilkins, 1954.