Introduction

An individual’s alcohol use and his problems with alcohol begin and develop in a cultural context. Besides individual biological and psychological factors, cultural factors constitute a major role in determining prevalence of alcohol use problems. Social values can encourage young people’s alcohol use, or actively discourage and encourage involvement in activities in line with adult mainstream values, hopes and expectations. The geographical location of Turkey allows different cultural influences both from the Muslim countries where alcohol use is religiously and/or lawfully forbidden, and from the European countries where alcohol use is highly prevalent. The Turkish Government and its institutions are based on laic laws. Although majority of the Turkish population is formed by Muslims, the great majority does not live up to the principles of the religion of Islam, and alcohol use is a socially acceptable phenomenon in Turkey. In deed, the Turkish population is a multicultural and multiethnic society, so that culture and many cultural interactions determine religious activities, but the reverse is not true. Religion is just a cultural element in the whole picture for a significant part of the population.

According to the World Health Organization Global Status Report on Alcohol [29], the rate of the use of alcohol in the year 2000 was 19.6% in general Turkey population, prevalence of alcohol dependence was 1.3%. Regarding the other Muslim countries, the prevalence of alcohol use within the past year lies between 0.5% in Egypt and 35.6% in Lebanon, and the prevalence of alcohol dependence ranges from 0.2% in Egypt to 7.3% in Iran. Still, these figures are far below those in European countries where the proportion of alcohol users among adult populations and prevalence of alcohol dependence were reported to be 62.0–97.5 and 2.2–12.2%, respectively. Turkey is more open to the effects of the Western culture than that of the East because of the government’s political choices, touristic visits from Europe, and the effect of mass media and Internet connections.

Distribution of the rates of alcohol consumption is usually heterogeneous in a given population where risk groups can be defined. University students may be one of the risk groups in terms of hazardous alcohol consumption because of some particular characteristics of the university life. First of all, the university campus can be considered as a distinct cultural environment where new peer relations and peer groups are formed. The university campus where social contact and drinking are frequent often has increased modeling for drinking, peer pressure to partake in drinking, and easy availability of alcohol that may lend itself to greater rates of problematic drinking. Alcohol use can become a norm in this sociocultural context where students may feel obliged to drink in conformity with the group’s norms for social approval [13]. Moreover, university students are free from demands of work and marriage, giving them the freedom to seek leisure without the adult responsibilities [24].

Life-time prevalence of alcohol use was reported as 74% in Israeli university students [14] and 61% in Hong Kong [1] university students, and changed from 49.2 to 70.8% with an eight-year interval in two universities of Lebanon [15]. In an international study prevalence of alcohol use was investigated among university students in 21 countries, and it was found that prevalence of any more than drinking on dinking occasions was 95% in Ireland, in a range changing from 42% in France to 86% in other European countries, 65% in the United States, 43% in Thailand, and 50% in Venezuela [9]. Thus, there is a considerable variation among different countries, even among the European countries.

In Turkey, students enter the university by passing a central exam after high school, and it is usually the first time that youths leave their protective parent home environment. With respect to alcohol use the university environment is relatively permissive when compared with students’ previous home and school environment. Many of the students either rent a house with their friends or they stay in student dormitories during their university years. Again considering alcohol use, this age group (18–24 years) corresponds to the time of first use of alcohol in Turkey. Clinical studies done in clinical alcohol dependent patients showed that the mean age of first use of alcohol changed between 14 and 20 in Turkey [7, 8, 20]. Peer influence is of paramount importance during adolescence [24] and drinking is generally perceived and imitated as an adult behaviour at this psychosocial developmental stage. Thus, the youth may need to prove himself/herself as an independent “adult”. During this period of time, adolescents generally begin the protracted psychological process of forging an identity and making role transitions one of which is entering the workforce [13]. All of these mentioned psychosocial developmental issues may be critical factors in regulation of alcohol consumption behaviour of university students.

The main aim of this study was to determine the extent of alcohol use problems in the university students in Turkey. This study is based on the observation that alcohol use, especially drinking beer, which was previously an occasional event for the youth entered the daily behavioural repertoire of university students in Turkey. There are several studies in Turkey which surveyed university students to determine the rate of use of alcohol and alcohol use problems. Alcohol use prevalence among first and forth-year university students was between 58 and 82.4% in Istanbul in 1983 [25]. Yüksel et al. [30] reported that 10% of university students from four university centres in Ankara used alcohol in the past month, and 2% used alcohol daily. Life-time prevalence of alcohol use was found as 72.3% and the prevalence of clinically significant alcohol use problems was found as 6.6% in a medical school [2]. Akvardar et al. [4] found that the prevalence of alcohol use problems was 22.4% in junior medical students and 8.9% in senior medical students. The results of the mentioned studies cannot be generalized to the whole country, as they were done in single university centres representing a restricted sociocultural environment. This study was done on five university centers which were located in different social environments with the aim of having more generalizable findings and determining the relationship between alcohol use and various sociodemographic factors. Having findings from different social environments would enable the institution of appropriate and specific prevention and intervention programmes for the alcohol use problems of the university students.

Method

This study is a survey carried out among randomly selected 2,000 students in five universities. Students were drawn from the faculties of political sciences. The reason for preferring these schools was first the ease to reach the sample by the authors. Secondly, a homogeneous sample with regard to the faculty would allow comparability and interpretation of the findings in itself. However, it is still a necessity to include other faculties with further studies.

Questionnaires were distributed to the participants voluntarily by lecturers while they were in attendance of a required course. The lecturers were instructed about introduction of the questionnaire forms. Students were assured that their responses would be confidential and they were informed that they were not obliged to complete the questionnaire. The respondents were instructed to place the completed questionnaire forms in an envelope before returning.

A total of 1,720 participants, or 86% of the sample took part in the study. The mean age of the sample was 21.5 ± 1.8. Distribution of the sample by gender and school is shown in the Table 1.

Table 1 Sociodemographic characteristics of the student sample
Full size table

The questionnaire was prepared by the authors as a self-rating form. It consisted of questions including sociodemographic status, academic status, parents’ education, economic status, and place of birth. Assessment of the pattern of alcohol use was done by asking questions on ever use of alcohol, use of alcohol in the past year, in the past month and past week. Students who drank alcohol in the previous year were asked about frequency of drinking and whether they evaluated their consumption of alcohol as problematic.

The CAGE (Cut-down, Annoyed, Guilty, Eye-opener) Questionnaire was applied by the interviewer to identify the problems of alcohol use. This instrument was also used in previous surveys done in Turkey [3, 4, 6, 26], so it was preferred in the present study with the aim of comparing the results. The CAGE Questionnaire is a scale including four items. Questions covered were: “Have you ever felt you should cut down on your drinking?”, “Have people annoyed you by criticizing your drinking?”, “Have you ever felt bad or guilty about your drinking?”, and “Have you ever had a drink first thing in the morning to steady your nerves or get rid of a hangover (eye-opener)?” Each question was scored with one point. The CAGE Questionnaire was originally developed by Ewing [11]. Probable presence of an alcohol use disorder is indicated by a score of 1+, whereas a score of 2+ was taken as the cut-off point for detecting presence of clinically significant alcohol use problems [23]. Validation of the Turkish version of the CAGE Questionnaire was done by Gul et al. [12] and it was shown that with cut-off score of 2+ the CAGE revealed a specificity of 86% and a sensitivity of 75 to 97% in detecting alcohol use disorders.

Statistical analysis

In the first step univariate analyses were carried out on the relationships between demographic and background variables and alcohol use. Alcohol use was taken as either the presence or the absence of alcohol use problem according to the CAGE2+ score, and drinking alcohol at least once a month or more frequently. Relationships between alcohol use and gender, school, academic year, subjective evaluation of academic performance, place of habitation of the family during the last ten years, place of residence of the student, education of the father and the mother, subjective evaluation of the family income and frequency of drinking in the past year were examined with χ2-square analysis.

Next, the relationships which were found to be significant in the univariate analyses were reassessed using logistic regression models. The presence of alcohol use problems according to the CAGE2+ cut-off score or the frequency of drinking (at least once a month or less frequently) were taken as dependent variables in the logistic regression analysis. Finally, the same univariate and multivariate statistical procedures were performed by taking individual items of the CAGE Questionnaire as dependent variables.

All statistical analyses were executed by using SPSS 13.0.

Results

Of the whole student sample 63.3% (n = 1,088) reported that they had ever tried drinking alcohol (Istanbul 67.5%, Ankara 67.6%, Izmir 76.8%, Mersin 54.2%, Muğla 51.6%). The prevalence of drinking more than once during life-time was 50.6%, whereas 12.7% of the sample stated that they had tried drinking only once during life time. A quarter of the students (25.3%, n = 436) reported that they had been drinking in the past week, and nearly half (48.5%, n = 835) had used alcohol in the past year. The students who had consumed alcohol during past year were inquired about the frequency of drinking, and 65.0% (n = 574) answered that they had been drinking once a month or more frequently. The overall prevalence of alcohol use problems according to CAGE2+ was 9.7% (19.9% among the students who used alcohol during the past year). Regarding the CAGE items, 24.8% of the past year drinkers felt they ought to cut down on their drinking, 16.6% had been annoyed by other people’s criticisms of their drinking, 24.8% felt bad or guilty because of drinking, 4.6% had consumed alcohol in the morning.

The subgroup of the sample who reported that they had consumed alcohol in the past-year excluding the rest of the sample was statistically analysed in order to detect the factors which differed those who had problems with alcohol use and the ones who did not have alcohol use problems despite their use of alcohol during the past year (Table 2).

Table 2 Relationship between alcohol use and sociodemographic factors in past-year drinkers-univariate analysis
Full size table

Male students used alcohol more frequently (χ2 = 37.402, P = 0.000) and more problematically (according to CAGE) (χ2 = 25.160, P = 0.000) compared with the females. As educational status of the father (χ2 = 9.392, P = 0.009) and the mother (χ2 = 18.049 P = 0.000) increased, a higher percentage of the subjects consumed alcohol. A greater percentage of students in the “good” family income group than in the other income groups stated that they drank alcohol once a month or more frequently (χ2 = 8.651, P = 0.034). A significantly smaller number of students who stayed in dormitory consumed alcohol at least once a month when compared with students who were living with their family or with friends or alone at home in the univariate analysis (χ2 = 21.087, P = 0.000).

A smaller number of students who were living with their family were CAGE2+ (χ2 = 10.107, P = 0.039). Presence of alcohol problem according to CAGE2+ was found to be significantly related with lower educational level of both the mother (χ2 = 14.690, P = 0.001) and the father (χ2 = 6.348, P = 0.042) in contrast to the frequency of drinking which was related to higher educational level of the parents.

In order to examine this controversy, univariate analyses were conducted for each item of the CAGE Questionnaire. The three CAGE items, namely Cut-down, Annoyed, Guilty were all related to the educational status of the mother. A significantly higher percentage of students in the lowest maternal educational group (illiterate and primary education) gave positive answers to the Annoyed (χ2 = 16.218, P = 0.000) and Guilt items (χ2 = 15.975, P = 0.000). Positive responses to the Cut-down item increased as the maternal educational level decreased (χ2 = 15.936, P = 0.001). Positive answers to the Annoyed (χ2 = 7.449, P = 0.024) and Guilt (χ2 = 17.153, P = 0.000) items increased as the paternal educational level decreased. Lower subjective academic performance was significantly related to positive Eye-opener item answers (χ2 = 8.822, P = 0.032), and low family income was related with more positive answers to the Guilt item (χ2 = 13.219, P = 0.004).

The statistically significant relationships were additionally tested by constructing logistic regression models. Maternal and paternal education variables were not taken in the same model, since it was assumed that they would be highly correlated. In general both the univariate and the multivariate analyses revealed similar results with a few exceptions. The educational level of the parents was related with the Annoyed item of the CAGE Questionnaire in the univariate but not in the multivariate analysis. The significant relationships between family income and the Guilty item and drinking once a month or more frequently, and the relationship between subjective academic performance and the Eye-opener item in the univariate analysis did not appear in the multivariate analysis.

In the logistic regression models, male students tended to have problems with alcohol about three times more than females (Tables 3, 4). Place of residence was still significantly related with alcohol use problems according to the CAGE, and also frequency of drinking. Although living in the dormitory seemed to be protective in terms of frequent drinking, just the opposite was true when CAGE scores are considered. Likewise, as educational level of the parents increased, the odds of having alcohol problems according to the CAGE decreased, whereas the odds of drinking at least once a month increased (Table 4). In order to be able to elucidate this controversy, individual items of the CAGE Questionnaire were taken as dependent variables in logistic regression models (Table 5). Students whose mothers were illiterate or primary school graduate tended to give more positive answers to the Cut-down, Annoyed and Guilty items. The odds of giving positive answer to the Cut-down and the Eye-opener items among those living alone was greater than the other residence groups. The predictors of positive answer to the Eye-opener item were male gender, living alone at home, and residence of the family being in a foreign country. Paternal educational level being in the illiterate/primary school category was significantly related with more positive answers to the Guilty item.

Table 3 Relationship between the CAGE items and sociodemographic variables-univariate analysis
Full size table
Table 4 Relationship between alcohol use and sociodemographic factors-multivariate analysis
Full size table
Table 5 Relationship between CAGE items and sociodemographic variables-multivariate analysis
Full size table

Discussion

Alcohol consumption by university students in Turkey has not reached the levels reported in Europe and in the United States. Prevalence of any more than drinking on special occasions among university students changed from 42 to 95% in West Europe, and from 52 to 75% in Eastern European countries [9]. In the present study prevalence of drinking more than once during life-time was 50.6%. Prevalence of drinking in the past year changed between 64 and 80% in a sample of American college students [10]. The corresponding figure was 48.5% in the present study. Kraus et al. [16] reported that 91.3% of males and 87.3% females among a general population sample of 18–24 year-old past year drinkers in Germany consumed alcohol once a month or more frequently. In the present study the corresponding rates were 75.4 and 55.8%, respectively. Prevalence of life-time alcohol use in the present study was 63.3%. This rate is still lower when compared with that found among university students in Israel (74%) [14], in Lebanon (70.8%) [15], and in Rio de Janeiro, Brazil (96.4%) [21].

Several studies done in various regions of Turkey indicated that over the last 25 years prevalence rates of life-time use of alcohol among university students changed between 36 and 82% [25]. Ögel et al. [18] found prevalence rate of alcohol use as 53.6% in Istanbul youth between 18–25 years of age. In a general survey done by the Association of Turkish Psychologists lifetime prevalence of alcohol use was found as 38.7% among 15–24 year-old males and it was 16.0% in female population at the same age group [26]. Prevalence of current drinking among individuals between 20–29 years of age was found as 36.7 [4]. When the CAGE scores are considered, the prevalence of alcohol use problems was found as 2.0% between the ages of 15 and 24 in general Turkey population [26], 2.2% among a general population sample over 15 years of age in Ankara [6], 19.9% in an Istanbul sample aged between 20 and 29 [4]. In this study, prevalence of alcohol problems is considerably high (9.6%) when compared with the results of the mentioned studies done in general population samples except the latter one. Akvardar et al. [3] reported that prevalence of drinking problems according to CAGE2+ positiveness was 20% in junior and 22% in senior medical students. Higher prevalence rates of alcohol use problems among the university students may be related to the university life being a particular period of life in terms of beginning a new life far from family control, passing an important milestone in identity development, entering a different social environment, and meeting with the university culture comprising its specific characteristics. All these changes can also bring difficulties and stresses which can facilitate drinking behaviour. Drinking is a way of socializing among peers as well as imitating adult roles.

Regarding the CAGE items, 24.8% of past-year drinkers felt they ought to cut down on their drinking, 16.6% had been annoyed by other people’s criticisms of their drinking, 24.8% felt bad or guilty because of drinking, 4.6% had consumed alcohol in the morning. More or less in parallel with these results, the figures are respectively, 31, 22.5, 27.3, and 8.1% among drinkers in the 20–29 age group in another study [4]. In another study carried out in Turkey, 42% of drinkers among medical students gave positive answers to the Guilt item [3]. In the present study, the controversy between the findings related to the CAGE positiveness and frequency of drinking is probably due to the overrating of the Guilty item in Turkish culture. In Turkey the sociocultural context including traditional, moral and religious reasons and also stigmatization of alcohol use might have contributed to this overrating.

The wide range of prevalence rates found in various studies in Turkey necessitates investigation of variables related with alcohol use in the country. The results of this study show that some characteristics tend to be associated with alcohol use problems among university students.

A consistent relationship was found between gender and alcohol use in many studies [4, 13, 16, 17]. Akvardar et al. [4] reported that men were more likely to be CAGE positive (CAGE 2+) than women. This is related with both biological [22] and sociocultural variables. Women are biologically more intolerant to alcohol, and easily get drunk, however “drunken woman” image and drinking behaviour in women is socially more unacceptable compared with men, and this distinction seems to be more valid for Turkey population. Additionally, males seem to have more unhealthy lifestyles than women as vonBothmer and Fridlund [28] reported that male university students in Sweden tended to drink too much as unhealthy lifestyle habits in general compared to the female students. They also reported that it was more common among male students to have the opinion that drinking is a part of student life and that use of alcohol makes it easier to socialize.

Oksuz and Malhan [19] studied factors associated with risky behaviours in a sample of Turkish University students, and they found that high maternal and paternal educational levels were associated with increased alcohol use. Similarly Dantzer et al. [9] reported that odds of heavy drinking were reduced among students whose parents were less educated, and Passos et al. [21] reported that past-month alcohol use was about 20% higher among university students with college educated parents. In a study on Chinese university students [1] it was found that mother’s educational level was not associated with alcohol use. There are other studies which showed that alcohol drinking problems were associated with high socioeconomic level [5, 14, 27]. According to our results, paternal education and particularly maternal education seem to be predictive for alcohol use and problems of the university youth with alcohol use.

Several studies done in college student samples in the United States showed that students living in on-campus residences and dormitories tended to drink more [13]. University hall non-residents were 2.1 times more likely than those who were staying at the university hall to be ever-drinkers [1]. In contrast, we found that living alone or living at home with friends increased the risk of drinking once a month or more frequently, but this was not true for living in student dormitory. Student dormitories in Turkey have some rules such as strictly determined entrance hours, being unable to stay out without permission, and prohibition of alcoholic drinks. When the CAGE items were taken separately, it was found that living alone increased the Cut-down item positiveness nearly twice, whereas living in the dormitory makes the student more prone to feel guilty about his drinking behaviour. Moreover, living in a dormitory or at home with friends or alone entails diminished exposure to parental controls and more frequent exposure to peer influences, therefore to opportunities to engage in such problem behaviours as drinking.

The male student who lives alone and whose parents’ educational level is low feels the need to cut down on drinking. An interesting point is that although the proportion of frequent drinkers increase as the maternal level of education raises, these students gave more negative responses to the Cut-down item. This may be due to a more tolerant family environment of the student whose mother has higher educational level. The relationship between the low educational level of the family and feeling guilty about his drinking behaviour according to the CAGE may indicate conservative structure of the family. Higher positiveness of the Guilt item of the CAGE Questionnaire was found in several other surveys done in Turkey [3, 4].

The male student who lives alone and whose family lives in a foreign country has the risk of having problems with alcohol according to the Eye-opener item of the CAGE Questionnaire. Dantzer et al. [9] also reported that heavy drinkers among university students were more likely to live away from home. Proximity to parents appear to play a role in protecting the student from alcohol problems, as indicated by our results and lower rates of drinking problems in students who live with their parents in other studies [9, 10, 13].

There may be some underestimation of the prevalence of alcohol use as a result of the absentees being omitted from the student population. As any other targeted study, there are limitations about the representativeness of the sample of the present study, but while the prevalence indicators of alcohol use may vary in other universities in Turkey, the factors associated to alcohol use are unlikely to differ, and this is the first large-scale multicenter university survey which investigated correlates of alcohol use problems.

We suggest that preventive efforts should aim providing alternative ways of socialization other than alcohol use behaviour for the university students. This can be achieved by supporting organization of the youth’s life through activities such as sports, giving responsibility within social projects, and supporting various student clubs of specific purposes such as arts and other intellectual activities.