The aim of this study is to highlight some social processes and conditions that will help explain the university students' involvement in illegal conduct. A number of hypotheses is listed in the previous chapter. In this chapter we describe some methodological considerations, focusing on the major techniques of data collection and analysis used in this research.
What kind of information would be suitable for testing the hypotheses listed in Chapter II? What kind of data would readily be available for analysis? To what extent could we rely on and justify the use of the available official statistics on illegal conduct? What type of data analysis would promote our understanding of the selected correlates of illegal behaviour among Canadian university students? These were some of the questions that we asked ourselves before we planned to collect and analyze the data.
The paucity of information on illegal conduct among Canadian university students can hardly be exaggerated. Generally, researchers have relied on available official statistics derived from the police files, court records, etc.. This has meant that youths who have been adjudicated have been the major focus of sociological inquiry. Many sociologists have commented on the inadequacy of the official statistics for purposes of research and generalization, and for accurately reflecting the distribution of illegal conduct among the youth of various social classes. Studies using self-report devices have indicated clearly that unadjudicated youths are just as much or more involved in illegal acts as are the adjudicated ones?
Considering the limitations of the available official statistics on the illegal conduct of university students, we decided to collect primary data by using the following techniques: (a) informal semi-structured interviews, (b) quasi-participant observations of "pot parties", and (c) anonymous questionnaires for self-reporting information. Admittedly, self-report techniques and self-reported data have their own problems,2 but we tried to overcome some of them by exercising controls wherever possible.
I. COLLECTION OF DATA
Our research questionnaire was composed of seven parts covering 169 items. Each part included a set of questions designed to measure specific variables stated in our hypotheses. Before constructing the questionnaire, some exploratory work was done through (a) informal, semi-structured interviews with student volunteers from the author's courses, and (b) quasi-participant observations of "pot parties". Our aim was to get some basic information on the everyday activities of students' lives; this included the organization of their activities; their attitudes and orientations; their beliefs and personal experiences. Some items on the questionnaire were based on our participation at these parties and discussions with students. What follows is a brief description of the contents of the questionnaire.*
Part I (Items 1-15): Part I sought information on the student's registration status, present year of study at the university, faculty, major field of study, age, usual living arrangements, frequency of visiting parents, father's occupation, father's education, formal academic performance, religious affiliation, church attendance, and self-confidence.
* For the detailed contents of the questionnaire see Appendix D.
Part II (Items 16-43): Part II contains check lists of 26 illegal acts plus 5 statements of the 5 illegal possession of drugs. Involvement in illegal acts and the illegal use of drugs were measured by these check lists. Included in the first check list were major and minor violations of law on the basis of which students could be adjudicated. The selection of these illegal acts was based on the following sources: (1) self-reported criminal offenses by the author's students in Juvenile Delinquency and Social Deviance courses at the University of Waterloo and the University of Guelph, (2) consultation with a number of University of Waterloo students about the kinds of criminal offenses usually committed by students, (3) criminal offenses by university students reported in the available criminological literature, (4) official criminal statistics for the 18-24 year age group from Historical Statistics of Canada, 3 (5) consultation with a local police officer about the kinds of illegal acts usually committed by the University of Waterloo students, and (6) the writer's knowledge of deviant and criminal behaviour.
Illegal acts were defined in descriptive terms rather than in terms of legal categories. Items were used that referred to stealing articles of a particular value rather than to descriptions of property offenses, e.g., robbery and burglary. Each item was designed to tap the frequency of engaging in an illegal act and took the following form: During your last school year, how many times have you driven a car without a driver's license or permit? Never, once or twice, 3-4 times, 5-7 times, 8 times or more.
The illegal drugs included in the check list were marihuana, LSD, amphetamines, barbiturates, and opiates. Our choice of these drugs was based on the following considerations: ( 1) There are frequent references to these drugs in the literature; (2) These drugs are highly publicized through the mass media of communication; (3) A considerable number of university students are familiar with these drugs, and the names of some of these drugs are part of their common vocabulary; (4) Involvement with each of these drugs is legally prohibited. Two questions were asked to measure the frequency and duration of the student's involvement with each of these drugs.
Part III (Items 44-69): Part III contains three check lists: the first contains 12 items on attitudes towards involvement in illegal acts; the second contains 10 items on attitudes towards involvement in illegal drugs, and the third contains 7 items on attitudes towards marihuana laws. Items for the first and second check lists were selected from Part II of the questionnaire. Each item was designed to measure the perceived degree of seriousness of engaging in illegal acts. Four response categories were provided for each item: very serious offense, serious offense, not really serious offense, not a serious offense.
In Canada unauthorized possession, trafficking in, and the import and export of narcotics and cannabis, as well as the cultivation of the opium poppy or marihuana, are prohibited by the law with the following penal consequences:
(a) Unauthorized possession is punishable
(i) upon summary conviction for a first offence, by a fine of one thousand dollars or by imprisonment for six months, or by both fine and imprisonment and for a subsequent offence, by a fine of two thousand dollars or by imprisonment for one year or both by fine and imprisonment; or
(ii) upon conviction on indictment, by imprisonment for seven years.
(b) Trafficking is punishable upon conviction on indictment by imprisonment for life.
(c) Being in possession for the purpose of trafficking is punishable in the same way.
(d) Unauthorized importing or exporting is punishable upon conviction on indictment by imprisonment for life, and in any case by imprisonment for not less than seven years.
(e) Unauthorized cultivation of opium poppy or marihuana is punishable by imprisonment for seven years.4
Items on attitudes towards marihuana laws sought the following information: (a) whether marihuana should be legally available to anyone 18 years or over, and (b) what ought to be the penalties for first, second and third offenses if unauthorized possession and selling of marihuana continue to be illegal in Canada. The penalties listed ranged from "No penalty" to "execution".5
Part IV (Items 70-142): Part IV was designed to gather information on parent orientations, peer-group orientations and youth culture involvement.
The degree to which students identify and associate with peers is at least one measure of their commitment to peer group expectations. There were 3 items on parent and peer group orientations. Each of the items was designed as a hypothetical situation with two choices indicating whether the students would be willing to meet parental, or peer group expectations. The choices were: (1) hitchhiking across Canada with friends or going to work in your parents' business, (2) continue seeing your friends against parents' wishes or stop seeing your friends according to parents' wishes, (3) going for a drive with your friends or attending church with your parents.
Our perspective on youth culture suggests that the quality of the content of most of the legitimate group activities is characteristically "informal" and "social". The participants who engage in common activities place a relatively high premium on personal, informal, social encounters. Regular active participation in informal group activities consolidates the social bonds of intimate persons with similar interests, and helps to reinforce favourable attitudes towards active involvement in such activities.
Membership in Selected Groups on Campus: A list of all the groups on campus ''was obtained from the office of the Federation of Students. All groups were distributed in the following arbitrarily designed categories of groups: (1) Federation of Students Organizational Category: Student's Council, Executive Board, Board of Communications, Board of Student Activities (e.g., pubs, movies, dances, concerts, major weekends such as Orientation, Homecoming, Winter Weekend and Summer Weekend), Board of External Relations-Education, Board of Publications. (2) Social, Cultural and Creative Arts Category: International Students Association, House of Debates, Amateur Radio Club, WHIPLASH, Drama, Carol Fantasy, Concerts, Dance, Folksinging Club, Camera Club, Student Wives Association. (3) Sports Category: Motorsport Club, Tiddlywinks Club, Chess Club, Duplicate Bridge Club, Hatha Yoga, Flying Club. (4) Political Category: Committee to End the War in Vietnam, Liberal Club, Young Socialists, Progressive Conservative Club, New Democratic Party Club, Gay Liberation Movement, Women's Caucus, Women's Liberation Movement, Women's Coalition for Repeal of Abortion Laws. (5) Religious Category: Christian Perspective, University Parish, Inter-Varsity Christian Fellowship, Students International Meditation Society, Student Christian Movement, the Navigators. (6) Ethnic Category: India Canada Association, American Students Association, Ukrainian Students Club, Caribbean Students Association, Muslim Students Association, Arab Students Association, Association of Greek Students, Chinese Students Association.
Fourteen formal and informal youth culture activities were included in Part IV. They are listed as follows:
Of 14 sets of five items each, six are related to formal and eight to informal youth culture activities. Items in each set were designed to tap the frequency of participation in one type of youth culture activity. Items were designed also to establish the status of other participants in the activity, attitudes towards involvement in the activity, status identicality of participants in the activity and involvement in marihuana use (the degree to which participants in formal group and informal youth culture activities are the same persons who usually smoke marihuana with one another), and the status of other participants in marihuana use.
Fourteen items were designed to measure the frequency of a student's participation in formal and informal youth culture activities. Each item was stated in the following form: On the average, how many times do you go drinking at a tavern, pub, etc., during the week? Five or more times a week, about 3-4 times a week, about twice a week, about once a week, hardly ever or never.
Our perspective on youth culture suggests that the more one's active involvement with "significant others" the more integrated one is in the youth culture. The more integrated one is in the youth culture the more likely one is to participate actively in legitimate youth culture activities and also in illegal acts. Fourteen items were designed to explore the relative influence of differentially statused youth culture participants on the actor's involvement in marihuana use. The differentially statused participants were classified as friends, acquaintances, siblings, others (e.g., instructors). The respondents were asked to indicate with which categories of statused persons did they participate in legitimate youth culture activities.
Another set of 14 questions was included in Part IV to explore student's attitudes towards involvement in youth culture activities. Each item took the following form: How much do you like to go drinking with these persons (friends, acquaintances, siblings, others)? I like it very much, I like it, I dislike it, I dislike it very much, I don't know.
The degree to which participants in the youth culture had identical status with participants in marihuana use was defined as status identicality. Fourteen items were included to measure the degree of status identicality. Each item was stated in the following manner: Are these the same persons (friends, acquaintances, siblings, others) who usually smoke marihuana with you? Response categories were: Yes, sometimes, no, I don't smoke marihuana.
A set of 14 items supplemented items on status identicality. These items explored the status of students who smoked marihuana with others with whom they were not socially active in formal group and informal youth culture activities. Each item was stated in the following form: If no, with whom do you usually smoke marihuana? Other friends, other acquaintances, others (instructors), alone.
Part V (Items 143-153): The items in part V were designed to gather information on the degree to which selected legitimate and illegal practices were institutionalized in the youth culture. A practice is institutionalized when three conditions are met: (1) a large number of members of a social system accept the norm, (2) many of those who accept the norm take it seriously, and (3) the norm is sanctioned.6 The sanctioning behaviour includes the essential meaning of institutionalization as described by Levy:
A given normative pattern affecting human action in terms of a social system will be considered more or less well institutionalized to the degree to which conformity with the pattern is generally to be expected and to the degree to which failure to conform with the pattern is met by the moral indignation of those individuals who are involved in the system and who are aware of the failure....Differences of degree relating to the second source of indeterminacy will be referred to as differences in the sanction aspects of the institution or its institutionalization. 7
Each item was designed to measure sanctioning behaviour by the degree to which non-participation in selected, legitimate, youth culture activities, and marihuana use was likely to result in loss of popularity among "significant others". Five response categories were provided for each item. A typical item is: A student who will never go drinking with his/her friends is likely to be: popular, a little popular, no difference, a little unpopular, unpopular.8
Part VI (Items 154-158): Part VI was designed to explore students' beliefs and experiences concerning the effects of marihuana use. Five items were used to measure the degree to which students thought that marihuana use generally affects one's appreciation of art, of music, sexual enjoyment, "self-awareness", and religious communion. Another five items attempted to measure the degree to which students felt that their own use of marihuana had affected their appreciation of art, music, sexual enjoyment, "self-awareness", and religious communion.
One item was used to measure change in the number of friends "caused" by the respondent's marihuana use. Eight items were used to measure the "positive and negative" effects on psychological behaviour. Positive feelings of well-being included conviviality, elation, tranquility, and euphoria; negative feelings included depression, anxiety, hostility and delusions.
Part VII (Items 159-169): Eleven items in part VII provided information on a student's role orientations. The roles and purposes implied in each item were typical of a student's university life.9
2. Pretest of the Questionnaire
Would all the items measure what they were designed to measure? Were all the items precisely applicable to university students? How would our research subjects respond to the content of the items? How much time would it take to complete the questionnaire? These were the kinds of questions that we asked ourselves after we had constructed our research instrument.
To answer these questions 17 undergraduate students from the University of Waterloo, and 12 undergraduate students from the University of Guelph who were enrolled in my courses were asked to do the following: (1) independently fill out the questionnaire, (2) to single out any particular problems that they encountered, (3) to list any general or specific comments about the content of items, and (4) to record the time spent in filling out the questionnaire.
This exercise proved useful. The students were favourably disposed to the questionnaire. Their comments resulted in the modifications of some items. However, very few problems were listed. The recorded time ranged from 23 to 38 minutes.
To further ensure that our research instrument was measuring what it was designed to measure we conducted a formal pretest on a class of 74 University of Waterloo undergraduate students in May, 1973. The questionnaires were checked, and the data were prepared for analysis. A computer run for frequency distributions revealed that the information was satisfactory for further analysis.
3. The Sample
A. Selecting the Sample
Our aim was to test relationships controlling for several variables simultaneously. A sample of 4000-5000 students was considered large enough to meet our objective. Among various alternatives, the administration of questionnaires to students in regular class hours was considered most efficient and productive. The Fall 1973 Timetable Enrolment Reports were obtained from the scheduling department of the University of Commons.* The graduate and correspondence courses were excluded from the Reports. There were 2168 courses representing 10,355 students from which to choose our sample. These courses were numbered serially for each separate faculty. Our estimates indicated that every 12th course selected would give us the required sample size. The followng technique was adopted to complete the course selection: numbers 1-12 were printed on 12 3" x 5" plain cards. The cards were shuffled, and one card was pulled from the deck. The number on this card was the course selected from the first twelve courses listed on the timetable report. This process was repeated six times - to select the first course from each faculty. Thereafter every 12th course was included in the sample by systematic selection. From a total of 2168 courses for 10,355 undergraduate students, 182 courses representing 5364 students constituted our disproportionate, random sample of classes drawn from six faculties at the University of Commons. Table 1 summarizes the structure of the sampling units.
B. Contacting Instructors
Before we could reach students in regular class hours it was necessary to get the instructors' approval. The instructors of each of the 182 sampled classes were contacted twice, either by telephone or in person. Initial contacts were made in late August 1973, during which we explained to them the nature of our research, our procedures of gathering data, and the importance of their co-operation. The same instructors were contacted again in early September 1973 in order to establish regular class times during which we could administer the questionnaires. Instructors who refused to cooperate were substituted by other courses randomly selected from the same sets of 12 courses.
* This research was conducted in a medium-sized university located in Southwestern Ontario, Canada. University of Commons is used as a fictitious name to preserve the anonymity of our respondents.
A total of 182 instructors were contacted. One hundred and sixty-six gave their approval and assured us their full co-operation in administering questionnaires to their classes.* Sixteen refused to co-operate** and were consequently replaced by others. The number of instructors who refused from each faculty is listed below:
A number of them allowed us to administer questionnaires during lab or tutorial hours. This was relatively common among instructors from Engineering, Science, and Mathematics faculties.
These instructors gave a number of reasons for their non-cooperation. The common ones were "too tight in teaching schedule", "can't afford to lose 40 minutes", "can't cooperate on moral and legal grounds". A letter from one instructor is reproduced below:
Department of Sociology, University of YYYYY, YYYYY, Ontario.
When you contacted me as regards administering a sociology questionnaire in my XXXX class I agreed without reservations. I now find myself having to withdraw my original agreement.
When I agreed I did not realize the character of your questionnaire. In particular I did not know that the questions in part II amounted to asking the students to confess to prosecutable crimes. Granted that participation in this project is voluntary and that anonymity is promised. [nevertheless don't believe that I can subject my students to this experience. I do not question the value of your research. The small size of the groups I teach, however, leads me to believe that students could not feel sure of security. In addition, the night section has a number of extension students whose special character leads me to believe it would not be to the interest of the College to have them exposed to your questionnaire.
I wish you the best in your research. I am sorry that I could not aid you in carrying it forward.
In each case the instructor was told that the research was for a Ph. D. dissertation, and that we were interested in exploring students' attitudes and orientations, the extent of their involvement in campus, youth culture activities and some of their nonconforming conduct. The students' participation in the research was voluntary. Anonymity was complete. At least 40 minutes of class time was required. Assistants were hired to administer questionnaires and to answer any questions that the students might have.
C. Administering Research Questionnaires
Our strategy was to contact selected classes of students as early in the Fall term as possible.* For this purpose five additional part-time research assistants were hired. They were required to make themselves available whenever they were needed. The questionnaire was thoroughly discussed with them. They were trained in handling class situations in the actual administration of questionnaires. Their duties included: (a) distributing questionnaires at the appointed time, (b) being present in classes and answering any questions, (c) returning the questionnaires, and (d) reporting and recording any problems that they encountered.
Data collection was started in the second week of September, and it was completed by October 30, 1973. A total of 4879 questionnaires were administered, 91.0% of our expected sample had completed questionnaires. We had anticipated some case mortality in our expected sample (5364) due to students being absent from classes, changing courses, or simply refraining from participating in research activity. A 9.0% rate of mortality was noted (5364 - 4879 = 485).
D. Checking the Completed Questionnaires
Each questionnaire was thoroughly checked. This resulted in a 4.94% (265) drop in the sample. These questionnaires were discarded because they were: (a) incomplete, (b) not taken seriously (e.g., incomplete, scribbled and illegible), and (c) carelessly completed and showing response sets. A total of 4614 completed questionnaires remained. Table 2 shows the distribution of discarded questionnaires.
The 4614 cases comprised 44.6% of the total undergraduate student population.
Our timing strategy was based on our effort to minimize the influence of some external factors which might possibly bias the results of our survey. Some of these factors were: (1) "exposure effect"; any delay in gaining information from newly registered first year students might change the quality of data because of their exposure to the campus youth culture, thereby confounding our estimates of the effects of campus youth culture on students' behaviour. (2) "Interaction effect"; prolonging the data gathering process might affect the quality of data because of students' interaction and communication about our research instrument. (3) The progressing term in time brings in its wake a multitude of mid-term examinations, tests, etc., which might interfere with our plans for administering questionnaires and students would likely give divided attention to the questionnaire. Besides, case mortality rates in the selected sample might be higher. Some of these problems were overcome by first gathering the data from the larger first and second year classes.
Table 3 indicates the proportions of six faculties, five educational levels, and males and females in the total population and in the sample.
II. ANALYSIS OF DATA
A. Making the Sample Proportionate
Initial computer runs on the data were performed to establish the quality of data. In order to make our sample more homogeneous the following categories of cases were excluded from the analysis:
Total number of cases in the sample = 4614
Part-time registered students = 64
Registration status not indicated = 6
Cases not classified in, socioeconomic categories according to the Blishen Occupational Scale = 27
Cases having missing values on dependent variable (Involvement in marihuana use) = 7
Balance = 4510
One of our objectives was to generalize our findings to the total University of Commons undergraduate student population. For this purpose we needed to have a proportionate representation of students (male and female) from all faculties, and all educational levels. Our selected sample approximated the total population of students at the university matched according to year of study and sexual status. In order to get a proportionate representation from the various faculties the following weighted formula was adopted:*
Analyses for marginals were performed on a weighted sample of 4507t cases or 43.5% of the total undergraduate student population.
B. Validity and Reliability of the Data
No formal tests of validity and reliability of the data were conducted. But some gross indications of reliability are listed below: (1) Expost facto checking of responses to the same or similar-in-content questions in various parts of the questionnaire permitted a multiple check on the respondent's truthfulness. This checking led us to believe that the data were very consistent. For example, the statement "I don't smoke marihuana" appeared sixteen times in the questionnaire. In aggregate terms only 2% of the responses to this statement were "inconsistent". (2) One item "During your last school year, how many times have you told a lie?" provided a further check. The fact that only a small fraction (7.2%) of the sample indicated not having told lies suggested that a large proportion of the sample were probably trying to answer questions about illegal acts truthfully. (3) The student's participation in this research was voluntary. Complete anonymity of the questionnaire was assured, and clear indications of the use of the data in aggregate form were supplied, so that there was little reason to believe that students would give false information on the questionnaire through fear of exposure. (4) During the collection of data, a good number of students informally assured the researcher that the amount of lying in completing the questionnaire would not be more than the amount of lying practiced by students in conducting their day-to-day affairs. (5) The completed questionnaires were carefully checked, and those which gave serious indications of inaccuracy were discarded. This screening improved the quality of the data.
* Spearman's Rank Order Coefficient of Correlation (rs) between rank orders of faculties in the total population and in the sample was .80 before the faculties were weighted.
t Three cases were lost in the weighting process.
C. Scales and Indexes
In their effort to explore the patterned variations in human conduct sociologists often construct scales and indexes which show the response patterns of a population. These aggregated measures are convenient tools for organizing human information and testing relationships between variables. In this study a battery of scales and indexes were constructed for testing our proposed hypotheses. A brief description follows:
Measure of Socio-economic Status: A person's occupation is the single most important determinant of one's style of life in North American society, and is a major criterion by which one is evaluated in the community. The literature on social stratification indicates that as a measure of socio-economic status the use of occupation has many advantages: it correlates highly with educational level, income, and subjective class affiliation. It is more accurately and easily obtained from students than information on income and home value and it is convenient to use in research.10
Data were gathered on father's occupation and formal level of education. These data were coded according to two scales: Hollingshead Two Factor Index of Social Position" based on American data, and the Blishen Occupational Scale '2 based on Canadian data. We chose the Blishen Scale for the following reasons: (1) the scale is based on Canadian data, and its use will allow us to compare our findings with other Canadian research, (2) the scale correlates very highly with other socio-economic indexes, e.g., the Pineo-Porter,13 and Blau-Duncan scales,14 etc., and (3) when tested against our own data the scale correlated highly (r = .86) with Hollingshead's Two Factor Index of Social Position.
Method of Classifying Subjects: The Blishen "Occupational Class Scale" based on Canadian data was used for classifying subjects. The following steps were taken:
Step 1. Using father's occupation each questionnaire was assigned a score from the Blishen Occupational Scale and later classified into one of six groups. The six groups corresponded to six socio-economic levels. Table 4 indicates the socio-economic levels used in this study corresponding to the names of the occupations, and the range of occupational scale scores on the Blishen Scale.
We faced the problem of classifying (a) reported occupational titles not listed on the Blishen Scale, and (b) ambiguous occupational titles, e.g., "works in a steel factory" or "works for government". Such questionnaires were put aside for further consideration.
Step 2. Reported occupational titles not listed on the Blishen Scale were reviewed and coded according to father's educational level, amount of skill implied'in the title, size of the organization or business, and its apparent similarity to existing occupational titles on the scale. For ambiguous occupational titles, and where father's occupation was not reported but his education was indicated, level of education was used as a main criterion for classification. Admittedly, some discretion and judgement were exercised by the coders.
There were a number of "discrepancies" noted in the questionnaires with reported occupational titles that were unlisted on the Blishen Scale. The reported father's occupation and his reported education sometimes did not "fit". For example, "University Professor", "Manager of a Bank" were reported having finished high school and finished grade school respectively. Such questionnaires were discarded.
The following scores were given to some occupational titles unlisted on the Blishen Scale.
Step 3. Questionnaires having poorly reported occupational titles, e.g., "labourers", "professional", or "worker" and not having included either father's education or place of work were also discarded. In these cases there was no basis for classifying them into any socio-economic categories.
Table 5 gives the distribution of subjects classified according to socioeconomic positions and their comparison with Ontario and Canada populations classified according to socio-economic positions.
Subjects located in socio-economic positions I and II, III and IV, and V and VI were combined into three socio-economic statuses as listed in Table 6. This was justified on two grounds: (1) in performing multivariate analyses we gain more by working with a larger number of cases than we lose in combining socio-economic positions. To check this practical consideration, distributions of six socio-economic categories were examined against our dependent variable (Involvement in marihuana use). The proportions in categories I and II, III and IV, and V and VI were closer to each other than they were to each in every other category. This provided an empirical rationale for collapsing adjacent socio-economic categories. (2) For purposes of analysis the grouping of these socio-economic categories is justified in accordance with the belief that the "life-styles" of subjects in the combined categories are not especially different in a university setting.
Self-Confidence Index: Self-confidence is an important "personality" variable that is used sometimes to understand participation in youth culture activities. Two items (14 and 15) were used to measure the degree of self-confidence experienced by respondents in managing social encounters. Four response categories to each item were arranged to indicate a continuum from low to high self-confidence. From combined data on both items the following index of self-confidence was constructed:
Involvement in Selected Self-Reported Illegal Acts: In this study involvement in illegal conduct is treated as a variable rather than as an attribute. The students were considered either more or less involved in illegal behaviour. We were interested in identifying configurations of illegal behaviour and establishing differential involvement of students in illegal behaviour. Following Guttman's suggestion that "measurements to be meaningful should be along one dimension at a time",15 we tried to construct scales of involvement in illegal behaviour. The procedure used for scale construction is described below:
A list of 24 illegal acts was constructed (items 16-41). All items, translated from the Canadian Criminal Code,16 were violations of law on the basis of which the students could be adjudicated. The items ranged from minor to serious offenses.
Two sets of twelve items were selected for Guttman scaling. Criteria for the choice of items were the following: (1) items were committed by a sizeable proportion of the sample, (2) the proportions admitting offenses for different items were differentiated by at least five percent, (3) the items seemed likely to measure a common dimension of involvement in illegal conduct." Four attempts were made to scale the items each time reducing and changing the number of items for the total sample. This was done separately for males and females from all socio-economic levels. However, the Guttman scale procedures did not produce acceptable coefficients of reproducibility.
*A variety of analyses was performed on the data. The scores of scales and indexes were divided into limited categories mainly for describing the results of cross tabulation analyses. In fixing ordinal categories of scores the following three principles were applied: (1) there should be a sufficient number of cases in each category to allow multivariate analyses, (2) there should be a sufficient dispersion of the sample over all categories to minimize error in estimates, and to avoid "category sealing" of sample, and (3) the constructed categories should facilitate description and meaningful interpretation of the data in terms of aggregated measures—scales and indexes.
Likert Type Summated Scores: Having failed to find an acceptable Guttman scale of illegal behaviour we decided to convert the data to Likert Type Summated Scores.18 The responses to each of the 24 items of illegal conduct were scored from 0 to 2. Total scores ranged from 0 to 48. These scores were trichotomized to form an index of illegal involvement as follows:
Involvement in Marihuana Use and Other Self-Reported Illegal Drugs: Involvement in each drug was measured with two items: frequency and duration of use (items 42-43). The responses to both items were weighted from 0 to 8 and 0 to 6 respectively. These weights were translated into Likert Type Summated Scores ranging from 0 to 14. Total scores were divided into the following three categories:
Indexes of involvement in other illegal drugs (LSD, Amphetamines, Barbiturates, and Opiates) were also constructed and the scores were dichotomized into never used = 0, used = 1 + categories.
Attitudes Towards Involvement in Selected Self-Reported Illegal Acts: Responses to 12 items (items 44-56) were subjected to a Guttman scalogram technique to measure students' attitudes towards involvement in illegal conduct. Four trials were made, each time reducing and changing the number of items for the total sample. This procedure was used separately for males and females, from each socio-economic level. None of the Guttman scales produced an acceptable coefficient of reproducibility. The data were converted to Likert Type Summated Scores. Each of the 12 items was scored as follows:
Attitudes Towards Involvement in Self-Reported Illegal Drugs: Responses to 10 items (items 57-66) were subjected to a Guttman scalogram technique to measure students' attitudes towards involvement in illegal drugs. The Guttman scales produced acceptable coefficients of reproducibility (.946) for the total sample (males and females) for all socio-economic levels. (Table 7). The cutting point for each item included in the scale was 3. The scale scores were translated into the following categories:
Attitudes Towards Marihuana Laws: Responses to six items on penalties for first, second and third offenses for unauthorized possession and sale of marihuana were subjected to a Guttman scalogram technique (items 68, 69). The scales produced acceptable coefficients of reproducibility (.929) for the total sample (males and females) for all socio-economic levels. (Table 8). The scale scores were translated into the following categories:
Peer - Parent Orientation Indexes: The degree of peer and parent orientation was measured by three items (items 70-72). Two response categories were provided for each item, one favouring peers the other parents. Each response was given a score of 1. The scores ranged from 1 to 3 for both peer and parent orientation. The scores were divided into three categories to form peer and parent orientation indexes:
Involvement in Selected Formal and Informal Youth Culture Activities: Six items (items 73, 93, 103, 113, 123, 133) measured the formal, and eight items (items 78, 83,88,98, 108, 118, 128, 138) measured the informal youth culture involvement. Five response categories were provided for each item. The responses for all items were converted to Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 30 for formal, and 1 to 40 for informal youth culture involvement. These scores were divided into the following categories to form indexes of involvement in formal group, and informal youth culture activities.
Social Intimacy of Participants in Selected Formal and Informal Youth Culture Activities: Measures of the status of participants are conceived as measures of social intimacy among participants in youth culture activities. Six items (items 74, 94, 104, 114, 124, 134) measured the degree of social intimacy among participants in formal group activities. Eight items (items 79, 84, 89, 99, 109, 119, 129, 139) measured the degree of social intimacy among participants in informal youth culture activities. Responses to all items were translated into Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 18 for formal and 1 to 40 for informal youth culture. These scores were trichotomized to form social intimacy indexes of participants in formal, and informal youth culture as follows:
*These may also be considered social roles.
Attitudes Towards Involvement in Selected Formal and Informal Youth Culture Activities: Attitudes towards youth culture involvement was measured by the degree of a student's liking to participate in formal and informal youth culture activities. Six items (items 75, 95, 105, 115, 125, 135) measured the degree of liking to participate in formal group activities. Eight items (items 80, 85, 90, 100, 110, 120, 130, 140) measured the degree of liking to participate in informal youth culture activities. Responses to all items were converted to Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 24 for formal and 1 to 32 for informal youth culture. These scores were divided into the following categories to form indexes of the degree of liking to participate in formal group, and informal youth culture activities:
A Summation of the Separate Measures of Involvement in Formal and Informal Youth Culture Activities: In the initial contingency control and multiple regression analyses, the six indexes of youth culture involvement, i.e., the indexes of involvement in selected, formal and informal youth culture activities; the indexes of social intimacy of participants in selected, formal and informal youth culture activities, and the indexes of attitudes towards involvement in selected, formal and informal youth culture activities, were used as separate measures. Results revealed that the use of these variables separately added to the complexity of the data and path models without appreciably increasing the amount of explained variance in the dependent variables, i.e., involvement and attitudes towards involvement in marihuana use (Table 9). The relationships among these variables were all in the predicted direction.** Since nothing was gained by keeping these variables separate, a summated measure of involvement in informal youth culture plus a summated measure of involvement in formal group activities were constructed by combining scores on these mutually related variables. The constructed measures are listed below along with their subdimensions.
*The responses of those who had participated in formal group and informal youth culture activities, but said "I don't know" to some items were treated as missing data. The scores for missing data were predicted from the respondent's average score on other items.
1. Involvement in selected, legitimate, informal youth culture activities = the score on frequency of participation in selected, legitimate, informal youth culture activities + the score on participation in selected, legitimate, informal youth culture activities with differentially statused persons (social intimacy score) + the score on attitudes towards involvement in selected, legitimate, informal youth culture activities with differentially statused persons.
2. Involvement in selected, legitimate, formal group activities = the score on frequency of participation in selected, legitimate, formal group activities + the score on participation in selected, legitimate, formal group activities with differentially statused persons (social intimacy score) + the score on attitudes towards involvement in selected, legitimate, formal group activities with differentially statused persons.
Our strategy was aimed at (a) reducing the complexity of the data and relationships among a multitude of variables, (b) increasing the efficiency in interpreting the results while simultaneously increasing our ability to grasp the social and cultural processes underlying the involvement in youth culture activities and illegal conduct, (c) avoiding the possible problems caused by statistical interactions, and the overlap of these variables, and (d) handling the problems involved in establishing the causal priorities among these variables in the path analysis.
We realize that combining data on attitudes and behaviour is sometimes theoretically unjustifiable since both can vary independently to some degree. Because our main aim was to single out common constellations of verbally expressed attitudinal and behavioural dimensions of youth culture involvement, we examined empirical relationships among these variables. The strong positive relationships among these possibly overlapping sets of variables suggested that these variables were measuring more or less common configurations of verbally expressed attitudes and self-reported behaviour concerning involvement in informal youth culture and formal group activities. Besides the observed relationships it was intuitively appealing to speculate that the use of combined measures in further analyses would not greatly distort our conclusions. This line of reasoning constituted part of our rationale for combining these variables into single measures. In stating the operational hypotheses, and in the final analyses of the data we have used these summated measures of differential involvement in informal youth culture and formal group activities.
Status Indenticality of Participants in Selected Formal and Informal Youth Culture Activities and Marihuana Smoking: Six items (items 76, 96; 106, 116, 126, 136) measured the status identicality of participants in formal group activities, and participants in marihuana smoking with the respondents. Eight items (items 81, 86, 91, 101, 111, 121, 131, 141) measured status identicality of participants in informal, youth culture activities and participants in marihuana smoking with the respondents. Data were converted to Likert Type Summated Scores. Each item was scored as follows:*
Total scores ranged from 0 to 12 for formal and 0 to 16 for informal youth culture participants. These scores were trichotomized to form indexes of status identicality of participants in (a) formal group and informal youth culture activities and (b) marihuana smoking.
*The respondents who had smoked marihuana, but who did not answer one or more of the items were treated as missing data. The scores for missing data were predicted from the respondent's average score on other items.
It is noteworthy that these status identicality indexes are also indexes of social intimacy of identically statused participants in marihuana smoking.
Social Intimacy of Non-Identically Statused Participants in Marihuana Smoking: Students seldom smoke pot alone. Some students smoke pot with other persons with whom they are socially active in youth culture affairs. Other students smoke pot with another category of person. Who are these other persons? Are they friends, acquaintances, etc.? Six items (items 77, 97, 107, 117, 127, 137) measured the degree of social intimacy among participants in marihuana smoking for formal group activities, and eight items (items 82, 87, 92, 102, 112, 122, 132, 142) for informal youth culture activities. Responses to all items were converted to Likert Type Summated Scores. Each item was scored as follows:*
Total scores ranged from 1 to 24 for formal, and 1 to 32 for informal youth culture activities. These scores were translated into indexes of social intimacy of non-identically statused participants in marihuana smoking as follows:
Perceived Sanctions for Failure to Participate in Selected Informal Youth Culture Activities and Marihuana Use: Eleven items (items 143-153) in Part V measured the extent to which students are sanctioned by their peers for failing to participate in common youth culture activities. All items were subjected to a Guttman Scalogram technique. The scale did not produce an acceptable coefficient of reproducibility on the first trial. When the number of items was reduced to six (items 143, 144, 145, 151, 152, 153) on the third trial, the Guttman scale produced an acceptable coefficient of reproducibility (.908) for total sample (males and females) for all socio-economic levels. (Table 10). The cutting point for each item on the scale was 4. The scale scores were translated into the following categories:
*Those respondents who had smoked marihuana, but who did not answer one or more of the items were treated as missing data. The scores for missing data were predicted from the respondent's average score on other items.
This scale was used in some initial contingency controls and multiple regression analyses. Although the relationships were all in the predicted direction its use was abandoned. Since our aim was to explore some underlying processes concerning differential involvement in legitimate, youth culture activities and marihuana use it seemed logically more appealing to use two subscales: one on the perceived likelihood of being sanctioned for failure to participate in legitimate informal youth culture activities—dating, partying, drinking with friends; another on the perceived likelihood of being sanctioned for failure to participate in marihuana use—smoking, sharing and selling "pot". Two Guttman scales on perceived sanctions were constructed.
Guttman scale 1: perceived sanctions for failure to participate in dating, partying, drinking with friends (items 143, 144, 145). Cutting point for each item was 4. The coefficient of reproducibility was .894.
Guttman scale 2: perceived sanctions for failure to participate in smoking, sharing, selling "pot" (items 151, 152, 153). Cutting point for each item was 4. The coefficient of reproducibility was .964.
Some further analyses were performed using these two Guttman scales of sanctioning. All relationships were more or less in the predicted direction. But it was realized that a Guttman scale of only three items may be technically untenable, and that a Likert Type Summated Score might be a reasonable alternative. Responses to all six items (items 143, 144, 145, 151, 152, 153) were converted to Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 15 for each of the two sanctioning indexes and they were used in the final multiple regression analyses.
Beliefs and Experiences Concerning the Effects of Marihuana Use: Two sets of five items each (items 154, 155) measured beliefs and experiences concerning the effects of marihuana use. The respondents were asked to indicate: (1) whether they believed that using marihuana generally affects the user's appreciation of art, appreciation of music, sexual enjoyment, "self-awareness", and religious communion (item 154), and (2) whether their use of marihuana affected their own personal experiences concerning appreciation of art, appreciation of music, sexual enjoyment, "self-awareness", and religious communion (item 155). The data were converted to Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 25 for beliefs, and 1 to 25 for experiences. These scores were trichotomized to form indexes as follows:
"Positive" and "Negative" Effects of Marihuana Use: Two sets of four items each (items 157, 158) were used to construct indexes of "positive" and "negative" effects of marihuana use. The data were converted to Likert Type Summated Scores. Each item was scored as follows:
Total scores ranged from 1 to 16 for "positive", and 1 to 16 for "negative" effects. These scores were trichotomized as follows to form indexes: -'
Role Oriented Types: Eleven items were used to measure two dimensions of student role: academic-intellectual (items 159, 160, 163, 164, 166, 167) and interpersonality (items 161, 162, 165, 168, 169). Arbitrary scores of 1 to 2 were assigned to all items as suggested by Bolton and Kammeyer.* Summated scores ranged from 6 to 12 for academic-intellectual dimension and 5 to 10 for interpersonality dimension. These scores were dichotomized to form indexes as follows:
*See questionnaire in Appendix D for scores assigned to each response category of the items.
A low score on each dimension indicated a favourable attitude toward activities implied in each dimension. A high score indicated an unfavourable attitude.
Role oriented types were a result of the crossproduct of values on academic-intellectual and interpersonality dimensions. The frequency distribution of the sample into various role-oriented types is presented according to the subjects' sexual status and their socio-economic position.
D. Correlation Analyses
Throughout the planning and execution of this study, our aim was to explore and illustrate a set of social processes and social conditions that would help explain the students' involvement in selected forms of illegal conduct. A number of general propositions and operational hypotheses were formulated. The hypotheses indicated specific relationships among a multitude of independent and dependent variables. Multivariate analyses were performed to test the hypothesized relationships. In order to obtain an empirical estimation of the relationships in quantitative terms, the following statistical techniques were considered suitable and used in the analyses: (1) Contingency Control Analyses, (2) Multiple Regressions and Path Analyses.
The main assumptions underlying the relatively more parsimonious techniques of multiple regressions and path analyses are summarized by Kerlinger and Pedhazur. They are listed below.
1. The relations among the variables in the model are linear, additive, and causal. Consequently, curvilinear, multiplicative, or interaction relations are excluded.
2. The residuals are not correlated among themselves, nor are they correlated with the variables in the system. The implication of this assumption is that all relevant variables are included in the system. Endogenous variables are conceived as linear combinations of exogenous or other endogenous variables in the system and a residual. Exogenous variables are treated as "givens". Moreover, when exogenous variables are correlated among themselves, these correlations are treated as "givens" and remain unanalyzed.
3. There is a one-way causal flow in the system. That is, reciprocal causation between variables is ruled out.
4. The variables are measured on an interval scaleP
* Faculty weighted.
(i) Curvilinear Relationships
Before performing multiple regressions and path analyses we explored the data to investigate the curvilinear relationships among major dependent and independent variables. Appropriate modifications in scales and indexes were made where curvilinear relationships were discovered. In multiple regressions, continuous scales and indexes were used. Where variables were measured on a nominal scale, dummy variable regressions were performed. While the assumption of linearity underlying the multiple regression and path analyses was not met in ideal terms; in no case was it seriously violated.
(ii) Statistical Interaction Among Variables
Contingency control analyses indicated that there was a statistical interaction between the involvement in informal, youth culture activities and the involvement in formal group activities. The involvement in formal group activities operated as a specifier of the relationship between involvement in informal youth culture activities and involvement in marihuana use (Tables 11, 12, 13). We attempted to include this statistical interaction in the multivariate regression analysis, using a number of different techniques, some of which are listed in Appendix C. However, the addition of extra variables in order to take the interaction into account did not contribute significantly to the variance explained. As a result no "interaction terms" were introduced in the final multiple regression analyses.
(iii)Major Variables Used in the Final Multiple Regression Analysis
The following major independent and dependent variables were used in the final multiple regression and path analyses:*
X1 = Peer-group orientation.
X2 = Active involvement in selected, legitimate, informal youth culture activities.
Frequency of participation in selected, legitimate, informal youth culture activities.
Participation in selected, legitimate, informal youth culture activities with differentially statused persons (Social Intimacy Index).
Attitudes towards involvement in selected, legitimate, informal youth culture activities with differentially statused persons.
X3 = Active involvement in selected, legitimate, formal group activities.
Frequency of participation in selected, legitimate, formal group activities.
Participation in selected, legitimate, formal group activities with
differentially statused persons (Social Intimacy Index).
Attitudes towards involvement in selected, legitimate, formal group activities.
X4 = Perception of sanctions for failure to participate in selected, legitimate informal youth culture activities.
X5 = Perception of sanctions for failure to participate in marihuana use.
X6 = Attitudes towards involvement in selected, self-reported, illegal acts.
X7 = Involvement in selected, self-reported, illegal acts.
*For Means and Standard Deviations of these variables see Tables B.1 and B.2 in Appendix B.
X8 = Attitudes towards marihuana laws.
X9 = Attitudes towards involvement in marihuana use.
X10 = Involvement in marihuana use.
X11 = Attitudes towards involvement in selected, self-reported, illegal drugs.
X12 = Involvement in other self-reported, illegal drugs.
X13 = Involvement in all self-reported, illegal drugs.
X14 = Perception of "positive" effects of marihuana use.
X15 = Experiences with the effects of marihuana use.
X16 = Beliefs concerning the effects of marihuana use.
X17 = Self-reported, formal academic performance.
X18 = Participation in formal religious activities.
X19 = Self-confidence experienced in social encounters.
X20 = Sexual status.
X21 = Usual living arrangements: Rural-Urban; Off campus-On campus.
X22 = Role-Orientations: Privatist; Conventional; Vocational; Academic.
X23 = Interpersonality scale.
X24 = Academic-Intellectual scale.
X25 = Major field of study.
X26 = Socio-economic status based on Blishen Occupational Scale.
X27 = Year of study at the university.
X28 = Status identicality of participants in selected, legitimate, informal, youth culture activities and marihuana smoking.
X29 = Status identicality of participants in selected, legitimate, formal, group activities and marihuana smoking.
Coeff of Alien = Coefficient of Alienation.
1. Sophia Robinson, Can Delinquency be Measured? Welfare Council of New York: Columbia University Press, 1936; Fred J. Murphy, Mary M. Shirley and Helen L. Witmer, "The Incidence of Hidden Delinquency", American Journal of Orthopsychiatry, XVI, 1946, 686-696; Austin Porterfield, Youth in Trouble, Fort Worth: The Leo Potishman Foundation, 1946 ;James F. Short, Jr. and F. Ivan Nye, "Extent of Unrecorded Juvenile Delinquency: Tentative Conclusions", The Journal of Criminal Law, Criminology and Police Science, 49, 4, 1958, 296-302; E. W. Vaz, A Sociological Interpretation of Middle-Class Juvenile Delinquency, op. cit.; 18-24.
2. For an excellent review of the problems and prospects of various data gathering techniques, see G. Nettler, Explaining Crime, McGraw-Hill, Inc., 1974, Chapters 3, 4, and Chapters 4, 5 and 6 of Second Edition, 1978.
3. M. C. Urquhart and K. A. H. Buckley, eds., Historical Statistics of Canada, Toronto: Cambridge University Press, 1965, "Section Y: Justice" by Nicolas Zay, 634-659; see also Nicolas Zay, "Gaps in Available Statistics in Crime and Delinquency in Canada", Canadian Journal of Economics and Political Science, 29, 1963, 75-89.
4. Le Damn Commission, Interim Report of the Inquiry into the Non-Medical Use of Drugs, op. cit., 178.
5. Boydell and Grindstaff had similar items in their research. See Craig L. Boydell, et al., eds., Critical Issues in Canadian Society, op. cit., 579-597.
6. Harry M. Johnson, Sociology: A Systematic Introduction, New York: Harcourt, Brace & World, Inc., 1960, 20.
7. Marion J. Levy, The Structure of Society, Princeton, New Jersey: Princeton University Press, 1952, 104. Reprinted by permission of Princeton University Press.
8. E. W. Vaz, "Explorations in the Institutionalization of Juvenile Delinquency", The Journal of Criminal Law, Criminology and Police Science, 62, 1971, 396-406.
9. Several items were adapted from Charles D. Bolton and K. C. W. Kammeyer, The University Student, op. cit.
10. Bernard Barber, Social Stratification: A Comparative Analysis of Structure and Process, New York: Harcourt, Brace and World, Inc., 1957, 183-185; James E. Curtis and William G. Scott, eds., Social Stratification: Canada, Scarborough, Ontario: Prentice-Hall, Inc., 1973; John Porter, The Vertical Mosaic: An Analysis of Social Class and Power in Canada, Toronto: University of Toronto Press, 1965.
11. A. B. Hollingshead, Two Factor Index of Social Position, New Haven: 1957. (mimiographed)
12. Bernard R. Blishen, "A Socio-Economic Index for Occupations in Canada", The Canadian Review of Sociology and Anthropology, 4, 1, 1967, 41-53.
13. Peter Pineo and John Porter, "Occupational Prestige in Canada", The Canadian Review of Sociology and Anthropology, 4, 1, 1967, 24-40.
14. Peter M. Blau and Otis Dudley Duncan, The American Occupational Structure, New York: Wiley, 1967.
15. Louis Guttman, "The Problem of Attitude and Opinion Measurement", in S. A. Stouffer, et al., Measurement and Prediction, Princeton, New Jersey: Princeton University Press, 1950, 49.
16. Martin's Annual Criminal Code, Canada Law Book, Ltd., 1973, 1977.
17. Matilda White Riley, Sociological Research II: Exercises and Manual, New York: Harcourt, Brace and World, Inc., 1963, 73-92.
18. Rensis Likert, A Technique for the Measurement of Attittides, New York: 1932, (Archives of Psychology), No. 140; H. M. Blalock, Jr. and Ann B. Blalock, eds., Methodology in Social Research, New York: McGraw-Hill Book Company, 1968, 94-97.
19. Powhatan Wooldridge, "Handling Missing Data", mimeo., n. d.
20. Fred N. Kerlinger and Elazar, J. Pedhazur, Multiple Regression in Behavioral Research, New York: Holt, Rinehart & Winston, Inc., 1973, 309. Reprinted by permission of Holt, Rinehart and Winston. See also R. Boudon, "A Method of Linear Causal Analysis", American Sociological Review, 75, 1965, 461-480; Otis Dudley Duncan, "Path Analysis: Sociological Examples", American Journal of Sociology, 72, 1966, 1-16.