Director, International Center for Comparative Criminology, University of Montreal, Canada
Professor, Department of Psychoeducation and of Psychology, University of Québec, Hull, and Researcher, International Center for Comparative
Criminology, University of Montreal, Canada
Director General, Research, Correctional Service of Canada
Senior Research Scientist, National Institute for Alcohol and Drug Research, Norway, and Department of Social Medicine, Uppsala University, Sweden
The present article explores the connection between alcohol and illicit drug use on the one hand, and crime on the other. The data were collected in a survey of 8,598 male inmates in (on average, about two weeks into the sentence) Canadian federal penitentiaries using a Computerized Lifestyle Assessment instrument. This survey instrument poses questions about various aspects of an inmate's life before incarceration. Results show that 79.4 per cent of inmates had used an illicit drug at least once in their lifetime. More recently, 50.3 per cent of inmates had used such drugs at least once during the six months prior to their most recent arrest; this was true for 45.8 per cent of most recent inmates during the four weeks prior to their most recent arrest.
In the six months prior to their arrest, 76.3 per cent of inmates had used alcohol at least once; this was true for 56.8 per cent of inmates during the preceding four weeks. Nineteen per cent of all inmates had used drugs every day or almost every day during this six-month period and about 34 per cent had used them at least once a week. Results also indicate that substance use is highly prevalent on the day of the most serious crime for which federal inmates were admitted to Canadian penitentiaries. However the Drug Abuse Screening Test (DAST) and the Alcohol Dependence Scale (ADS) indicate that the majority of inmates who use drugs and alcohol are not alcohol or drug dependent but that substance use on the day of the crime may facilitate criminal activity. Yet, statistics show that one half of the most serious crimes are committed in the absence of alcohol or other drugs. Despite a strong association, it cannot be concluded that the ingestion of an intoxicant causes crimes. Substance use and criminality are related in a complex way.
Numerous scientific studies show that there is a statistical association between the use of illicit drugs and criminality [1-5]. The strongest link is generally found with simple acquisitive crimes of various kinds, such as theft, break and enter and robberies.
Less information is available about how the use of the various types of illegal drugs are related to criminal activity and if there is an association between specific drugs and the type and intensity of such activity. Considering differences in the cost of drugs, addictive potential and other pharmacological properties, one could expect such differences to exist.
There is also a statistical association between the use of alcohol and crime, the strongest link being with homicides and assaults. Studies show that often both the perpetrator and the victim had been drinking prior to the violent criminal act [6-7]. It has also been shown that a large proportion of other types of crimes were preceded by drinking on the part of the offender .
Alcohol is a legal substance in most countries and its sale is monitored by governmental agencies. Thus, information is available that allows assessments of the aggregate level of association between alcohol and criminal acts. Studies have been undertaken in various countries and other kinds of jurisdictions. Owing to the controls exercised over the sale of alcohol, it is also possible to study the impact of sudden changes in the availability of alcohol on the level of crime. Such changes come about when, for example, the legal drinking age is lowered, when the sale of beer or wine is allowed in grocery stores or when a strike stops the transportation or sale of alcoholic beverages. Studies at this kind, however, have been overwhelmingly limited to the link between alcohol use and violent crime.
The possibility of establishing aggregate levels of association between the use of illicit drugs and crime is severely limited. Only rough estimates can be made of the availability and levels of use of the various types of drugs at any point in time. Indirect indicators must be used, such as the cost of a drug on the street or the quantity of drugs confiscated by authorities. Such estimates do not have the accuracy needed for cross-sectional or time trend analyses which try to assess the causal impact of drugs on crime. General population surveys are used from time to time to assess levels of use of various types of drugs but costs would become prohibitive if this method were used for assessing the strength of the relationship between drugs and crime. Other methods must therefore be used.
Official records, such as court documents and police records, are often used. In the case of victims of homicide, it is possible to gain access to information on drug and alcohol use through the coroner's inquest [9-10]. The survey method has, however, been used to fill the gap in available information: through victimization surveys and self-report studies it is also possible to make comparisons between alcohol use and drug use in relation to deviance, delinquency and crime. The most cost-effective self-report studies are those carried out with known offenders, in particular with offenders who are readily available for surveys in jails and prisons.
Obtaining estimates on statistical associations between the use of psychoactive substances and crime is only a first, relatively easy, step. The ultimate goal is to estimate the strength of the causal relationships that exist between the two. A particularly useful measure of the causal impact of substance use and abuse on criminal activity would be to estimate the proportion of crimes in a society that can be said to be caused by the use of alcohol and illicit drugs, respectively. In the health sciences, it is common to look for this attributable fraction, or etiologic fraction, for an illness or other health condition. This is done through the use of standard statistical procedures which vary somewhat, depending on the nature of the data. Only recently has the logic of attributable fractions been applied to causal estimates in the area of alcohol, drugs and crime. *
A research programme is under way, aimed at estimating attributable fractions of alcohol and drugs on crime, specifically, at establishing the proportion of various types of crimes that can be said to be caused by the use of alcohol, cocaine, heroin and cannabis. The authors are restricting their assessments to examples in Canada but a more general objective is to find methods that can be standardized so that this procedure can be replicated in other countries.
Data for the present article, which explores the connection between alcohol and illicit drug use on the one hand, and crimes on the other, were collected in an ongoing survey of inmates entering Canadian federal penitentiaries. The article therefore deals with crimes detected by the police and processed by the courts, and for which the individual charged was found guilty and given a sentence of at least two years in prison. Several different selection processes were applied to the totality of criminal activity in Canada before the data used were collected.
The generalization of the findings to all criminal activity and all individuals committing crimes in Canada is therefore unknown. However, the sample inmate population comprises a greater number of individuals who have committed serious crimes and who have relatively long criminal careers than any other population available for study. The social problems and costs linked to their criminality are more serious than for any other group and extend far beyond their individual lives. For these reasons, the findings presented below have a great deal of social relevance in their own right.
Upon intake, all inmates in Canadian federal penitentiaries are required to respond to a survey instrument that poses questions about different aspects of the inmate's life (see table 1). The Computerized Lifestyle Assessment instrument (CLAI) is a computer-driven questionnaire administered at the time the inmate is being assessed for treatment needs and for placement in a suitable facility at the beginning of the prison sentence. This is done, on average, about two weeks into the sentence.
Table 1. Main content areas of the Computerized Lifestyle Assessment
|Health||Relationships||Drug use||Alcohol use||Criminal activity|
Overall impact on various life areas
Patterns prior to age 18
Patterns 6 months prior to arrest
Patterns 28 days prior to arrest
Dependence scale (DAST)
Overall impact on various life areas
Patterns prior to age 18
Patterns 6 months prior to arrest
Patterns 28 days prior to arrest
Dependence scale (DAST)
Assessed drug and alcohol impact on crime
Number of crimes on current sentence
Most serious crime on current sentence
The Correctional Service of Canada (CSC) started using CLAI with a few inmates in 1989. During the following two or three years, penitentiaries in other parts of the country were included in this type of survey. Administration of the survey continues and, at the time of writing, the number of cases in the computer file is approximately 20,000. The number of inmates available in the data file is 16,957. Of these, 284 (or less than 2 per cent) are women. This sub-group has been omitted from the analyses but will be the subject of a subsequent report.
Although coverage of CLAI was inconsistent in the five administrative regions of CSC during the first years of implementation, 12.5 per cent of cases in the CLAI data file were from the Atlantic region, 40.9 per cent were from the Québec region, 32.6 per cent were from Ontario, 10.1 per cent were from the Prairie region and 3.8 per cent were from the Pacific region. The best geographical coverage in CLAI data was achieved between the years 1993 and 1995. The number of new male inmates assessed with CLAI was 8,598.
The CLAI questionnaire contained a number of questions concerning the demographic characteristics of the individual inmate. The proportion of new inmates born outside Canada was 18.0 per cent; and 26.2 per cent were members of a non-white racial or ethnic group. Aboriginal offenders made up 5.5 per cent and Black offenders 8.6 per cent of all new admissions. The mean age at admission was 32 and the median was 31. Six per cent of inmates had never attended school and 7 per cent had six years or less of formal schooling.
It took the inmates approximately two hours to complete the CLAI questionnaire, which contained a few questions on their views concerning the instrument. Eighty per cent of the inmates considered that the task of filling in the questionnaire was "easy", while 16 per cent considered it "a little difficult", 2 per cent considered it "quite difficult" and the remaining 2 per cent considered it "very difficult". As for the length of the questionnaire, 81 per cent considered it about right, 12 per cent considered it too long and 7 per cent considered it too short. Slightly more than one half the inmates (53 per cent) said that they understood the instructions and questions "very well", 42 per cent understood them "quite well", and 5 per cent did not understand them "too well"; 91 per cent said they liked doing the survey and 90 per cent said they would encourage a friend to take it.
Robinson, Porporino and Milison  conducted a reliability and validity assessment of some components of CLAI on a sample of 503 inmates and concluded that the instrument had good psychometric properties and a high concordance with information on an inmate's file.
The Computerized Lifestyle Assessment instrument contains detailed questions concerning the past habits of drug and alcohol use of the newly incarcerated individual: his initiation to use and his patterns of regular use prior to age 18 are probed. More recent use is covered by questions pertaining to the six months and four weeks immediately prior to arrest. The situational association with criminality is measured by a series of questions on the use of drugs and alcohol on the day that the crime was committed. Here a descriptive account is given of the use of drugs and alcohol by inmates, starting with initiation and ending with the narrowest focus and the most relevant to direct causation: use of psychoactive substances on the day of the crime.
In addition to the factual information on the use of psychoactive substances, CLAI contains questions that ask the inmate to make an overall assessment of the impact of drugs and alcohol on all of the crimes he has committed, as well as detailed assessments of the influence that his use of these substances has had on his judgement, aggressive tendencies and the likelihood of his committing the crime etc. Such estimates are naturally coloured by the perpetrator's subjective perceptions and, sometimes, wishful thinking. Subjective estimates, however, provide useful comparisons with more objective statistical estimates of the causal impact of drugs and alcohol on crime. Subjective estimates of causal impact and the factors affecting them form the second part of the results presented herein.
The prison population had a history of relatively extensive alcohol and drug use. Nearly four in five inmates (79.4 per cent) had used non-medical psychoactive drugs, almost exclusively illicit ones, at least once in their lifetime and 52.1 per cent had used them with a frequency of at least once a week over a period of time. Nearly 19 of 20 regular users (93.9 per cent) had quit drugs voluntarily at least once. Alcohol had been used at least once by 95.1 per cent of the inmates. Using the same measure as that for drug use, 62.7 per cent had used alcohol regularly, at least once a week, and 78.1 per cent of the regular users had stopped using it voluntarily at least once.
The mean ages at which the inmates reported initiation to drug and alcohol use and to regular use of drugs and alcohol is shown in table 2. The proportions of users in the total inmate group of 8,598 individuals are shown in parentheses. About two-thirds of initial users of both alcohol and drugs later became regular users. The time elapsed from experimental drug use to first regular drug use (1.5 years) was less than half that of the corresponding time elapsed for alcohol use (3.5 years). This may be an indication of the relative strength of the social, psychological and pharmacological factors which determine the formation of dependence on drugs versus alcohol.
Table 2. Mean age at first use and at the beginning of first regular use
of alcohol and illicit drugs
|First use||14.5 years
(95.1 per cent)
(79.4 per cent)
|First regular use||18.0 years
(62.7 per cent)
(52.1 per cent)
By far the most common drug of initiation was cannabis (either in the form of marijuana or hashish) which was chosen at first use by 85.6 per cent of the inmates who had taken drugs at least once in their lifetime (see table 3). The use of cocaine (including crack) and lysergic acid diethylamide (LSD), which were the second and third drugs of choice, was very low by comparison. Marijuana/hashish also dominates (but not as strongly as at initiation) as the drug of choice among the 52.1 per cent of inmates who had used illicit drugs regularly before the age of 18; 56.8 per cent of them had used at least marijuana on a regular basis (see table 4).
|Table 3. Type of drug used at initiation to drug use|
|Drug||Percentage of users|
Glue or gas
Table 4. Percentage of inmates who used major drugs on a regular basis
prior to age 18 (multiple responses possible)
All inmates who had
used drugs regularly
Glue or gasoline
As noted above, 79.4 per cent of the inmates had used an illicit drug at least once in their lifetime. More recently, 50.3 per cent of inmates had used such drugs at least once during the six months prior to their most recent arrest, and this was true for 45.8 per cent of the inmates during the four weeks prior to the most recent arrest. In the six months prior to their most recent arrest, 76.3 per cent of inmates had used alcohol at least once and 56.8 per cent had used it at least once during the four weeks prior to the most recent arrest.
Drugs had been used by 19 per cent of all inmates every day or almost every day during the six-month period, and about 34 per cent had used them at least once a week (see table 5).
Table 5. Frequency with which various types of drugs were used
during the six months prior to the most recent arrest
(Percentage of all inmates)
Every day or
almost every day
|At least once a day|
Dugs with alcohol
Drug users dominated over alcohol users in daily or near-daily use (19.1 per cent compared to 13.3 per cent). There were more inmates who had used alcohol on at least a weekly basis than there were drug users with the same frequency of use (44.4 per cent versus 33.7 per cent). Cannabis, cocaine and heroin were the three drugs most commonly used on a daily basis. It is probably a reflection of the addictive property of heroin that the difference between daily or near-daily use and use at least once a week is small, compared to the other drugs. An indication of the prevalence of the combined use of alcohol and drugs is the finding that, during the six months prior to incarceration, 9.7 per cent of inmates (i.e., 19.3 per cent of inmates who had used drugs during that period) had always or often used their drug together with alcohol and that 5.7 per cent of inmates reported that they had used alcohol together with drugs every day or almost every day during that period (see table 5). During the most recent four-week period prior to arrest, 11.1 per cent of inmates had used drugs every day and 25.2 per cent had used drugs on average at least every other day. The mean number of days of use among those who had used drugs at least once during the four-week period was 14.9 per cent. This translates into a drug use of 50.3 per cent of all days during that period. In the total population of inmates, at least one type of illicit drug was used on 24.3 per cent of days. The 11 per cent of inmates who had used drugs daily accounted for 45.4 per cent of all drug use days in the population.
The inmates in the present study were convicted for the first time (including for juvenile offences) at the mean age of 17.9 years. Close to one quarter (24.7 per cent) were serving their first sentence, while 20.9 per cent had one or two previous convictions, 21.6 per cent had been convicted three to five times before, 11.7 per cent had been convicted six to nine times and 21.4 per cent had been convicted more than ten times.
In assessing the crimes committed over their lifetime, 53.2 per cent of the inmates reported at least some of their crimes had been committed under the influence of drugs and 28.8 per cent considered that most or all of their crimes had been committed under the influence of drugs. In the same way, 50 per cent of the inmates considered that some of their crimes had been committed under the influence of alcohol and 24.1 per cent that most or all of their crimes had been committed under its influence.
In considering the total sample, inmates had been convicted for a mean number of 4.1 crimes on their current sentence. Those who were under the influence of drugs when committing at least one of those crimes had been convicted for 5.3 crimes while those who were under the influence of alcohol had been convicted for 4.5 crimes on their current sentence and those who were not under the influence of either drugs or alcohol had been convicted for 3.9 crimes. This is another indication of the somewhat more intensive criminal behaviour of those using drugs and alcohol in connection with their crime.
The focus of the present paper is the use of alcohol or drugs on the day that a crime was committed. Looked at first will be the proportion of inmates who reported the use of alcohol, drugs or both alcohol and drugs on the day of a crime. The last section will report on the assessment by the inmates of the causal role of alcohol and drugs when committing a crime and the probability of their committing the crime under the influence of drugs given their habits in the four weeks prior to their arrest.
The crime selected for special study is the current crime for which the inmate received the longest sentence, which can be said to be the most serious of the crimes for which he was serving time. Table 6 shows the distribution of the types of crime in our sample.
|Table 6. Most serious crime on the current sentence|
|Robbery or attempted robery
Break and enter
Assault or wounding
Murder or manslaughter (homicide)
Fraud or forgery
Driving while intoxicated
Possession of stolen property
Possession of a weapon
Other traffic offence
Abduction or kidnapping
Other type of offence
Just over one half of the inmates (50.6 per cent) had used either illicit drugs or alcohol or both on the day of committing their most serious crime. Table 7 shows the patterns of use for various major crimes on the sentence. There was a rather clear distinction between acquisitory crimes and violent crimes in the prevalence of use of drugs and alcohol. While homicides and, more pronouncedly, assaults and woundings were predominantly alcohol-related, crimes such as thefts and break and enter showed a higher prevalence of drug use on the day of the crime. Robberies and attempted robberies, however, showed the clearest preponderance of drug use. There was little involvement of drugs in sex offences and, where drugs had been used, this had occurred mostly in combination with alcohol.
Table 7. Use of alcohol or drugs on the day of the most serious crime,
by type of crime
|Crime (and number committed)||Drugs||Alcohol||Alcohol and drugs||No substance used|
|Robbery or attempted robery (1,811)
Drug offence (1,427)
Break and enter (1,146)
Sex offence (1,052)
Assault or wounding (581)
Murder or manslaughter (535)
Fraud or forgery (262)
Driving while intoxicated (235)
Possession of stolen property (174)
Attempted murder (130)
Possession of a weapon (128)
Other crimes (632)
|Total, all crimes (8,579)||16||21||13||49|
Given the way the question was phrased in referring to the day of the crime, it is possible that in some cases the inmate may have referred to the use of alcohol or drugs that occurred after the crime was committed but on the same day. It was possible to control for this by analysing responses to a question asking how many hours prior to the crime the use of a substance had occurred. The results showed that a maximum of 1.9 per cent of the alcohol use, 3 per cent of the drug use and 2.1 per cent of the combined use of alcohol and drugs had not occurred until after the crime.
It may seem surprising that drug offences had such a low representation in drug-related cases. This can be explained by the lesser drug crimes of possession and minor trafficking which receive sentences that are milder than the minimum two-year sentence required for incarceration in a federal penitentiary. Most of the more serious drug crimes are committed by individuals who are not themselves regular drug users.
Cocaine was the drug most often used on the day of the crime (see tables 8 and 9). Inmates had consumed cocaine during the day in close to one half (46 per cent) of all drug-related crimes (see table 8). The bottom line of table 9 shows that the overall cocaine involvement rate per 1,000 most serious crimes was 135 (i.e. 13.5 per cent of all most serious crimes were such that the inmate/offender had used cocaine on the day of the crime). For cannabis, the rate was about one half of that (66 of 1,000 crimes) and for heroin it was one sixth of the rate for cocaine (22 per 1,000 cases).
Table 8. Drug use on the day of the most serious crime, including drugs
used together with alcohol, by type of crime
|Crime (and number committed)||Marijuana||Cocaine/crack||Heroin||Other drugs|
|Robbery or attempted robery (754)
Drug offence (342)
Break and enter (413)
Sex offence (153)
Assault or wounding (176)
Murder or manslaughter (148)
Other crimes (318)
|Total, all crimes (2,524)||22.3||46.0||7.4||24.4|
Cocaine was used most frequently in connection with thefts and, second, with robberies and attempted robberies. Heroin was most prevalent in connection with drug crimes and robberies. Break and enter was most often connected with marijuana use, while the crimes of assault (murder/manslaughter and assault/wounding) were linked with drugs other than the three specified in tables 9 and 10: tranquillizers (14 per cent), LSD (5 per cent), PCP (4 per cent) and barbiturates (4 per cent).
Table 9. Rates per 1,000 inmates who had used marijuana, cocaine/crack, heroin
or other drugs on the day of the most serious crime on their current sentence,
including drugs used together with alcohol, by type of crime
|Crime||Marijuana||Cocaine/crack||Heroin||Other drugs||No drugs used|
|Robbery or attempted robery
Break and enter
Assault or wounding
Murder or manslaughter
|Total, all crimes||66||135||22||72||706|
Interesting patterns could be observed in the combinations of alcohol and the different types of drugs. Cocaine (especially crack/cocaine), and even more so heroin, was used predominantly without alcohol (see table 10). Use of marijuana and tranquillizers was, on the other hand, more prevalent in combination with alcohol. Also of interest are the combination of LSD with alcohol, and the fact that when more than one type of drug was used, they were more likely to have been used together with alcohol than without alcohol in a truly multi-substance combination.
Table 10. Types of drugs used on the day of the most serious crime,
alone and together with alcohol
Drugs and alcohol
Glue or gas
Other types of drugs
More than one type
All inmates who had used drugs at least once in their lifetime were asked to assess the overall effect that drug use had had on their involvement in crime. A small proportion (4.8 per cent) said that their use of drugs had decreased their criminal activities, while 44.2 per cent reported that their drug use had increased their involvement in crime, and the remaining 51.1 per cent judged that their drug use had had no effect on their criminality. The same assessments for the causal impact of alcohol showed that about the same percentage as for drugs attributed a decrease in criminal activity to their alcohol use (3.7 per cent), fully 63.6 per cent considered that alcohol had had no effect one way or the other, and 32.7 per cent considered that their alcohol use had contributed to their criminal behaviour.
Inmates who had used drugs, alcohol or both on the day of committing their most serious crime were asked how these substances had affected them. Among those who had used drugs, 78.7 per cent said that the use of the drug had made them more likely to commit the crime, 83.4 per cent said the same about alcohol, and 90.9 per cent said so about the combination of alcohol and drugs that they had used.
Among those who said that the use of a substance had made them more likely to commit their most serious crime, a large majority considered that its consumption had affected their judgment at the time of the crime (see table 11). A much greater proportion of inmates reported that alcohol had made them more argumentative, aggressive and violent than was the case for users of illicit drugs. The users of both alcohol and drugs on that day fell somewhere in between, but closer to the pattern for drugs than for alcohol.
Table 11. Drug and alcohol affected inmates who reported being more prone
to argue and be rude to others, more aggressive or violent, and affected in their
judgement by the substance taken on the day of their most serious crime
|Drugs||Alcohol||Drugs and alcohol|
More prone to argue
Questions on the effects of having been under the influence or having used alcohol or drugs on the day of the crime imply a natural-causal type of model in explaining the relationship between substance use and crime: having used the substance in some way caused the perpetrator to commit the criminal act. Another major type of explanation of the association is that of means unto an end: a crime was committed in order to get one or more psychoactive substances for personal use. In response to the question "Was this crime committed to get or while trying to get drugs for your own personal use?", about two thirds of the inmates who had used drugs at the time of their most serious crime indicated that this was the case. Acquisitory crimes, such as theft (83.6 per cent), robbery (78.1 per cent), fraud (70.4 per cent) and break and enter (68.1 per cent) were the crimes most commonly committed in order to obtain drugs among those who had used drugs on the day of the crime. Those who either used alcohol or both substances reported this reason in lower percentages (7.8 per cent and 38.7 per cent, respectively).
Evidence of an association between alcohol and drug use and crime abounds. The link between drugs and crime can take many different avenues. Some individuals who experience cognitive impairment or disinhibition while intoxicated may manifest impulsive behaviours ; others experiencing financial problems arising from heavy use of an expensive drug would be inclined to commit crimes that will provide the financial means for obtaining their drug . There are several additional pathways that link alcohol and drugs to crime. For the purpose of investigating these, and their contribution to the overall associations, it is important to obtain information on the joint prevalences of the various types of psychoactive substances and major types of crimes in key populations. The CLAI data bank offers several key insights into the alcohol, drug and crime relationship among Canadian federal inmates during different periods of their lifetime.
One half of the inmates were under the influence of alcohol or a drug on the day that they committed their most serious crime. This figure is comparable to data compiled by the United States Bureau of Justice Statistics  which, in 1997, recorded that 51 per cent of prisoners in the United States of America had reported using alcohol or drugs while committing their offence. The data in the present paper are also comparable to the statistics on drug involvement reported by the United States Office of National Drug Control Policy  which estimated that between 17 per cent and 36 per cent of inmates had used drugs at the time of their offence. When asked if their offending was affected by their substance use, nearly three quarters expressed the belief that there was a link. Although personal assessments of the role of substance use in the commission of crimes must be regarded as prone to information bias, it can be noted that this proportion does not differ very much from the estimates obtained for the present study. Still, caution is necessary: some subjects may have preferred to portray their crimes as substance induced in order to minimize their responsibility. More objective information, however, is available which supports subjective assessments of this kind. For instance, the present analysis showed that those who were under the influence of drugs had committed, in the past, more crimes than those who were not. Apart from driving while intoxicated, assaults and thefts were the most important crimes to be committed under the influence of alcohol or drugs; driving while intoxicated and assault were more likely to be committed by alcohol users, while theft was predominantly committed by illegal drug users.
Despite the heavy involvement of drugs, alcohol was the substance most often used by federal inmates. Alcohol is also more strongly associated with crimes that have raised the level of concern in society more than any other: crimes of violence. In addition to the objective information on the type of crime committed by the inmates, the observation was made in their self-reports which indicated a greater propensity towards argument and aggressiveness than was the case for illicit drugs. According to Roth , it must be concluded that examination of the link between mind-altering substances and crime, especially violent crime, should not be done without a careful evaluation of alcohol use. The majority of those who used alcohol on the day of a crime exhibit a pattern of regular alcohol use. The association between crimes and patterns of use and dependence on alcohol will be examined in a subsequent paper.
Cocaine was second in prevalence of use after cannabis among federal inmates in the period immediately prior to arrest. It was, however, the illicit drug most commonly used by federal inmates on the day that they committed their most serious crime. The proportion of cocaine users on the day of the crime was double that of cannabis users. Moreover, the proportion of cocaine users among the study population was astonishingly high in comparison with national statistics. While about one per cent of the Canadian population used cocaine over a 12-month period, 28.8 per cent of federal inmates had used it during the six months prior to incarceration. Robberies represent the most common serious crime committed by those who used cocaine on the day of the offence (break and enter and theft were the second and third most common). Robberies made up one third of all crimes committed by cocaine users, while they represented less than one quarter of all crimes committed by federal inmates in the population of newly admitted inmates. The need for money is strongly associated with the criminal activities of users of this expensive drug. This interpretation is supported by their responses: 68 per cent of cocaine users said that their need for drugs made them commit their most serious crime. This figure is much higher that the one reported by the United States Bureau of Justice Statistics  or the United States Office of National Drug Control Policy  where between 10 per cent and 20 per cent of male inmates indicated that they had committed their offence in order to obtain money for drugs. The difference may in part reflect the incarceration policies of the two neighbouring countries. Nonetheless, it must be kept in mind that a considerable proportion of the cocaine users did not put the blame for their criminal activities on the drug used. In agreement with Hunt , it must be concluded that drug use is not the only factor to consider when analysing the link between drugs and crime. Other factors, such as the price of a drug in relation to marginal income from illegal sources, the other sources of income available and the level of addiction should be carefully considered.
Compared to other studies, especially from Europe [18-23], there were few heroin users in the population surveyed for the present paper. Like cocaine users, they are over represented among robbers. In addition, more than one quarter of heroin users had been admitted to a Canadian penitentiary for a drug offence. These two criminal activities represent more than two thirds of the most serious crimes among heroin users newly incarcerated in Canadian federal prisons.
Substance use is highly prevalent on the day of the most serious crime for which federal inmates were admitted to Canadian penitentiaries. Preliminary analyses on the DAST and ADS scales used in CLAI indicate that the majority of the inmates using drugs and alcohol were not alcohol or drug dependent. Instead, substance use on the day of the crime may have facilitated their criminal activities. At the same time, one half of the most serious crimes were committed in the absence of alcohol or other drugs. Despite a strong association, it cannot be concluded that the ingestion of an intoxicant causes crimes. Substance use and criminality are related in a complex way. In any event, the high rates of alcohol and drug use in the inmate population suggests that this issue deserves more systematic exploration. One step in this direction would be further examination of how criminality and the level of dependence on drugs and alcohol are related. Beyond the role of substance use and abuse, it is necessary to highlight the importance of controlling for individual factors (e.g., propensity towards criminality) and contextual factors (e.g., the possible role of drug control laws in determining the price of illicit substances and the level of criminality) in studying the relationship between alcohol and drug use and crime.
*Estimates of causal impact, such as attributable fractions, can be assigned with much greater ease and confidence in the case of risk factors related to illnesses. The direction of causality can often be based on almost irrefutable causal models. For example, if the task is to assess the impact of an occupational hazard-such as exposure to asbestos fibres-on the risk of lung cancer and to compare this to the risk of smoking, there is not much scientific basis for entertaining the competing hypothesis that a propensity towards lung cancer explains past smoking patterns and/or choice of an occupation where asbestos is handled and fibres are inhaled. This means that measures of association can, with some confidence, be used as measures of causal impact. This aspect of the assignment of attributable fractions is non-controversial but it is not so with psychoactive substances and crime. These two entities essentially refer to forms of behaviour with no necessary time-ordering; central causal processes in the models must take into account human free will and a host of social and psychological factors.