Representaciones sociales del alcohol

Please quote this paper as: Páez, D., Basabe, N., Igartua, J. J. , Iraurqui, J. & Valdoseda, M. (1998) Social Representations of Drinking and Youth Behavior: Comparing Surveys and Focus Group Data on Social Beliefs. En S.Moscovici, I.S. Markantonis, A.V. Rigas & N Vilias (Eds) Social Representations and Contemporay social Problems. Athens/Paris: D.Danias/Maison des Sciences de l`Homme. LC.

SOCIAL REPRESENTATIONS OF DRINKING AND YOUTH BEHAVIOR: COMPARING SURVEYS AND FOCUS GROUP DATA ON SOCIAL BELIEFS COMMUNICATION TO THE MEETING REPRESENTATIONS TO SOCIAL PROBLEMS UNIVERSITY OF ATHENS LABORATORY OF EXPERIMENTAL PEDAGOGY

APPLICATION

OF

SOCIAL

Athens, 9-10, October, 1995 Authors: Dario Paez, Nekane Basabe, Juan J. Igartua, Joseba Iraurgui, Maite Valdoseda Department of Social Psychology, University of the Basque Country, Spain Peter M. Bentler Department of Psychology University of California, Los Angeles

1

ABSTRACT The relevance of sociodemographic factors in alcohol consumption has been well established by researchers in this field. In this study we intend to study alcohol consumption from the perspective of the theory of Social Representations. Controlling for social demographical variables, shared beliefs and expectancies held about alcohol and its effects can be used to predict the way it is consumed. Data obtained from a study carried out amongst young people confirm this hypothesis. The presence of socially shared beliefs is verified through the structural modelling of a survey and also using a qualitative technique (focused group discussions). Comparing results from surveys with group accounts we found that beliefs about positive alcohol effects were shared and avaliable both in group accounts and in individual responses. Negative effects were shared "privately" but were not displayed in group discussions. Structural causal modelling with survey data show that only positive beliefs were related to the degree of drinking. The same ocurred when analysing group accounts. Higher age group (higher level of drinking) shared more beliefs (largest group's mean) in positive effect. Triangulationn between individual responses (survey), content analysis of group accounts and agreggated data analysis (group as unit of analysis) confirmed that beliefs about positive effects circulated more or were not only avalaible "in the head", but were also accesible in social discourse. These accesible and more actively distributed positive beliefs were more related to behaviour than the negative ones. The triangulation between structural modelling, discussion group and previous questionnaire results, allows us to be sure of the social and explanatory character of these beliefs -and to verify them as social representations. Our research has partially supported a preventive campaign, showing the heuristic value of the social representations approach in order to cope with social problems. KEY WORDS:

1

Alcohol Youth Social Representations Beliefs

Excessive youth drinking during week-ends is perceived in Spain as a problem both by adults and by health authorities. Probably, this concern may be due to the fact that this pattern is associated with an increase in car accidents - one of the most important causes of death in young people. But, we must also bear in mind that this excessive youth drinking is perceived as a problem because of its "deviance" from past cultural mediterranean rules - associated with regularly drinking wine in meals. Superficially, this new youth pattern of drinking is a deviance from the traditional importance attached to sociability, controlled and regular drinking, or wine and alcohol acting as a "medium" and a "milieu" for social interaction. For young people week-end drinking is partially supported by a socially shared knowledge that helps give this practice a satisfactory meaning. In other words, by social representations about alcohol and drinking. In this chapter we will examine some aspects of this young peoples∩ social representations of drinking in order to help improve preventive campaigns. Social representations of alcohol refers to the social knowledge originated, used and shared in order to guide social responses to a social object (which in this specific case would be alcohol). Moscovici (1988) defined "social representation" as a cognitive product about a social object created by, and used to communicate among, a group. For a social representation to exist it must not be limited only to a series of beliefs or cognitive schema liberally extended through the social ecology. A social representation refers to a series of scripts for interaction and the theories which are implicit in them, and are publicly acted out. These beliefs are not only tied to forms of strategic interaction but they also hold social functions, such as defending the subjects'social identity, explaining socially relevant events, and justifying and guiding actions in their light. Apart from being socially shared, originated and used, and taking as objects only events of a certain social weight, these beliefs must be structured. These beliefs are based on ideologies and values and are also emotionally charged (Doise & Palmonari, 1986; Doise, 1988; Di Giacomo, 1987; Elejabarrieta, 1991; Farr, 1987; Ibáñez, 1988; Jodelet, 1991; Páez, 1987; for a critical view Jahoda, 1988; Potter & Wetherell, 1987). In other words, while researching the social representations of alcohol we will be interested in how different social groups, defined by their proximity to the social object (alcohol consumption) have different beliefs about the effects of alcohol. There is evidence to corroborate our view on this point-differences have been found between alcoholics and light drinkers as to what they believe the effects of alcohol are. Alcoholics and heavy drinkers expected more globally-positive changes, more assertiveness and more intense physical and social pleasures when compared to lighter drinkers (Brown, Goldman & Christiansen, 1985). It has also been found, proving the normative or prescriptive character of 1

beliefs about alcohol, that the level of consumption considered to be normal varied according to the drinker's sex and age. The most negative opinions about consumption were held by the very young and by women (Franco, Gili, Giner, et al., 1988). When studying social representations of alcohol it is important to mention the concept of "expectancies about the effects of alcohol". Brown, Goldman, Inn, et al. (1980) define expectancies about alcohol as the beliefs that a subject has with regard to the effects of this substance. There is a line of thought which has led to studies trying to discover the predictive value of expectancies or beliefs about alcohol when compared to consumption levels (Cárdenas & Moreno-Jiménez, 1990; Christiansen & Goldman, 1983; Leigh, 1989; Rohsenow, 1983; Stacy, Widaman & Marlatt, 1990). The study of alcohol expectancies has been carried out mainly using field studies. Of the existing instruments the most popular is the AEQ (Alcohol Expectancy Questionnaire) developed by Brown, Goldman, Inn, et al. (1980). It is composed of 90 items and the various factorial analyses which has been applied to it have produced six belief dimensions, on the effects of alcohol: global positive changes; social and physical pleasure; sexual enhancement; social assertion; tension reduction and arousal with feelings of power. Other authors have added two new factors which refer referring to negative expectancies: alcohol as an inducer of cognitive-motor deterioration and personal irresponsability (Rohsenow, 1983). Cárdenas & Moreno-Jiménez (1990) found that in Spain the dimensions were more of a global nature than those of Brown, Goldman, Inn, et al. (1980). The factors found were: a) alcohol as an agent of positive individual change (on cognitive, behavioural, emotional and well-being levels); b) alcohol as an agent of negative individual change (cognitive and motor impairment, loss of control, irresponsability and agressive behaviour); c) alcohol as a modifying factor in social intercourse. The results of both longitudinal and cross-sectional studies (Brown, 1985; Christiansen & Goldman, 1983; Christiansen, Smith, Roehling, et al., 1989) have confirmed that beliefs about alcohol's effects are related to consumption. In particular, positive effects are better predictors of future consumption than negative ones (Cárdenas & Moreno-Jiménez, 1990; Christiansen, Goldman & Inn, 1982; Leigh, 1989). Social Representations and Justification of Alcohol Consumption The use of structured and closed instruments (scales, questionnaires) in psychological research has been acknowledged as limiting. Focus discussion groups and in-depth interviews, so long as the content analysis codes are validated, allow a triangulation of empirical results, in other words, it allows us to compare questionnaire results with open and intensive research methods. The motivational and dynamic (processual) elements of the phenomena are also better appreciated using qualitative methods (Schwartz & Jacobs, 1984; Taylor & Bogdan, 1

1986). Research in social representations tries to correlate quantitative and qualitative views of group beliefs and attitudes. To summarize previous literature (Jodelet, 1991; Doise & Palmonari, 1986; Di Giacomo, 1987) we will be able to talk about social representations, and not just about expectancies (individual beliefs) or collective representations (shared beliefs) or ideologies (dominant beliefs), if the beliefs have a structure, if they have emotional or affective resonances, if they are associated to different social behaviours and groups, and if they are reproduced publicly in group situations as justifications of when, how and why we think or behave in (such a way in) response to a social object or event. If these conditions are fulfilled one can talk about social representations of alcohol. These beliefs would then be useful in order to explain drinking patterns and to plan public health, or specifically anti-alcohol, campaigns. Sex, Age, Social Consumption

Integration,

Previous

Behavior

and

Alcohol

Sociodemographic variables of sex and age have repeatedly showed their capacity to predict alcohol consumption. There is evidence to show that men drink more alcohol, and more often, than women. However, recently this difference has begun to disappear (Alvira, 1986; Comas, 1985; Elzo, González de Audikana, Ayestrán et al., 1990). With regard to age, it has been found that the youngest age group (15-17 years of age) has the largest percentage of non-drinkers and the lowest percentage of heavy drinkers (Elzo, González de Audikana, Ayestarán, et al. 1990). It is also noteworthy that the fundamental change in the consumption pattern of alcohol takes place from the age of 16 onwards (Ajanguiz, Apodaka, González, et al., 1988). Turning to social integration, various studies have suggested a relationship between alcohol consumption by the young, and the socialization process (Hualde, 1990). Supporting this, Seeman & Anderson (1983) found that being part of a social network correlated positively with higher alcohol consumption. Newcomb & Bentler (1988, 1989) show that alcohol consumption is positively related to social integration in teenagers - in this research the scale of alcohol consumption has a ceiling effect, (because higher ranks are «to drink one or more time by day»). Finally, research on attitudes shows that previous practices is one of the strongest predictors of behaviours (Bentler & Speckart, 1981). It is possible to speculate that positive beliefs about the effects of alcohol are rationalizations and justifications post hoc, a secondary phenomena without explanatory power. We will examine the explanatory power of shared beliefs, confirming his structured and publicly displayed features, and statistically controling the influence of sociodemographic factors. Study 1: Quantitative Research Method 1

Subjects and Procedure Data was obtained from 430 youngsters as part of a bigger sample which was obtenied using a random selection procedure (proportional and stratified by age, sex and district). The age ranged between 13 and 25 years and average age was 19 years and 11 months. The sex distribution was 49.9% men and 51.1% women. Sample error was estimated at 3.5% with a confidence level of 95.5%. Data was collected by means of a personal interview held at home. Measures For this study we have used models from other longitudinal studies already mentioned. Four are presented: 1) Sociodemographic variables (age, sex). Sex as a dichotomous variable (1=men;2=women). 2) Social integration and social support: a measurement of social integration-number of friends-(«How many friends can you usually count on to go out with to have fun?») ranged from 1 (from 0 to 2 friends) to 4 (9 or more friends); and a scale of social support (Vaux, Phillips, Holly, et al., 1986), composed of two dimensions: family relationships and peer relationships, which had both 5 items and ranged from 1 (low support) to 4 (high support). 3) A scale of expectancies and beliefs about the effects of alcohol. A synthesis of the main items of the AEQ scale developed by Brown, Goldman, Inn, et al. (1980), the Rohesnow scale (Rohsenow, 1983) and Christiansen & Goldman's AEQ-A scale (Christiansen & Goldman, 1983) has been used. The final scale was composed of 34 items (the answer format was, 1=«totally disagree» to 5=«totally agree»). A previous exploratory factor analysis was extracted nine factors (Basabe & Páez, 1992), in this study and for the confirmatory factor analysis were selected the four main dimensions, composed by 20 items. 4) Habitual consumption Alcohol consumption variables. (frequency of alcohol use during the last month, ranging from 1=«never» to 3=«four or more times»), this is a measure used in the Basque Health Survey (Gobierno Vasco, 1987). Consumption of alcohol on weekends (refering to quantity or amount of alcohol drunk during the last weekend) (and measured by c.c. of alcohol). Three items from a scale about frequency and circumstances surrounding alcohol use (Franco, Gili, Giner, et al., 1988): habitual consumption with friends, at parties and at family gatherings. These items ranged from 1=«never drink» to 5=«always drink»; and finally a scale of drinking problems during the last year (Robbins, 1984), with four items: discussions or fights with the family, peer and other people and health problems, these items ranged from 1=«never» to 3=«some times». Analytical Procedure Direct effects of sex, age, social support and expectancies 1

about the effects of alcohol on «social youth drinking» and on «drinking problems» were analyzed using a structural-equation modeling (SEM) procedure (e.g. Bentler, 1989; Bentler & Bonett, 1980). The EQS program (Bentler, 1989) was used for all SEM procedures using the maximum likelihood (ML) estimation. Three indexes of model fit from this program, the normed fit index (NFI), the nonnormed fit index (NNFI) and the comparative fit index (CFI) were used to evaluate the fit of the models. The chi- square test was used for the overall fit of the model. The Bentler-Bonett normed fit index (NFI) takes the sample size into account and acceptable values of the NFI should equal at least 0.90 (Bentler & Bonett, 1980). Results Confirmatory Factor-Analysis Measurement Models Before the structural models of effects were examined, initial measurement models were estimated to evaluate the latent construct. A confirmatory factor analysis was used to evaluate the adequacy of factor structures of family support, peers' support, beliefs about alcohol, social youth drinking and drinking problems variables. We present the models corresponding to the latent variables, in accordance with the notes given out by the EQS programme, the observed variables appear as «V» and the latent variables as «F», the notes «E» correspond to the measurement errors and the residual variances of the observed variables. Neither model reached statistical nonsignificance, except the latent construct «drinking problems», but the fit indixes are higher than 0.90, in every case and for the different indixes of adjustment which have been used (see graphics 1, 2, 3 for the confirmatory factor analysis of the measurement models). Social support is composed of two dimensions, social support from family and integration into peer group. Goodness of fit indexes are 0.91 for NFI and 0.93 for CFI (see Graphic 1). Social youth drinking is characterized by alcohol use at parties, with friends, habitual consumption, alcohol use on weekends and use at family gatherings. The fit indexes are high, NFI=0.98 and CFI=0.98 (see Graphic 2). The drinking problems construct includes fights with other people and with peers, and with smaller factor loadings, health problems and fights with the family. The fit indexes are NFI=0.98 and CFI=0.99 (see Graphic 3). INSERT GRAPHICS 1,2,3 Dimensions of Beliefs Held About Expected Effects of Alcohol: Confirmatory Factor Analysis:"Beliefs About Expected Effects of Alcohol" INSERT TABLE 1 The confirmatory factor analysis is composed of the four main dimmensions extracted in a previous exploratory factor analysis (Basabe & Páez, 1992) (see Table 4 for the confirmatory 1

factor analysis of the beliefs about alcohol). The first factor, termed belief in the individually disinhibiting effects of alcohol, refers to a series of effects such as: alcohol is a good antidote for shyness (V21), it helps overcome bashfulness and inability to talk (V27), it makes us happy and joyful (V25), it helps to express feelings (V11) and with alcohol is easier to chat with others (V14). Other items have smaller factor loadings (see results for V23, V29, V24, V15 and V17). The second factor (beliefs in positive psychological and interpersonal effects) is defined as the belief in the expected positive effects such as: becoming more optimistic (V1), makes us feel good (V6), people are more more amusing (V2), feeling better or it's a pick-me-up (V9), alcohol helps flirting (V29) and it gives us more self-confidence (V17). Other beliefs are less important (see results for V23, V4, V14, V32 and V31), we should mention that negative beliefs have negative and small loadings. The third factor (negative effects at the cognitive level and predisposition towards social rule breaking) has as it basis the idea that alcohol has negative consequences, both at the cognitive- it is difficult to know what are you doing (V12), cannot understand others (V13) nor think clearly (V10)- and social level-break rules (V24), have fights (V31) and disobey (V15)-. The last item (V21)-«alcohol is a good antidote to shyness»- has small loading. The fourth factor (alcohol produces aggression and feelings of power) is defined as believing that alcohol induces aggressive behaviour: breaking rules (V24), disobey (V15), to mistrust and be suspicious of others (V32), feel more powerful (V4), to have a fight (V31). Less important is the belief that people are not more amusing with alcohol (V2). Finally, beliefs about disinhibiting effects and the positive psychological and interpersonal effects are correlated. There is a covariation between positive beliefs (F2, F1). Structural Model Analyses A structural model analysis was carried out to verify the hypothesis as to which specific group of beliefs best predicted future alcohol use and the impact of different forms of social integration and support -negative influence of family support and positive influence of peer support and integration on social youth drinking. The dependant variable was «social youth drinking» and predictive variables were those associated with the extent of beliefs about alcohol, an index of social integration, social support and the sociodemographic variables sex and age. The confirmatory structural model posits that social youth drinking is influenced by the following latent variables: positive and negative beliefs about effects of alcohol (expectancies), social support-family and peer-; and observed variables: number of friends, age and sex. The general hypothesis states that the social youth 1

drinking increases the lower the family support is, and the hihger the support and integration into peer groups is. It also increases the higher the number of friends is. Moreover, consumption is higher the stronger the positive beliefs are, and the weaker the negative beliefs about alcohol are. Men show higher social youth drinking than women, and it also increases as age increases. It is in this group of young people in which we may find higher social youth drinking. In a second general model drinking problems were included. In other words, the latent dependent variable drinking problems was introduced into the previous model. In this case, we posit that higher social youth drinking will positively influence the higher number of problems related with alcohol consumption. Also, less family integration (family support) will influence a higher amount of problems rel ated to social youth drinking. In order to know the impact of the variables posited, we analyzed the different models, starting off with the most simple one and finally finishing with the most complete one. In this way, we can check the direct effects of the independent variables, and the percentage of explained and non-explained (or residual D1) variance of the latent dependent variable «social youth drinking» when we include step by step each explanatory variable. INSERT GRAPHIC 4, 5, 6 We can see (Graphics 4, 5, 6) that from model number 1 to model number 3 the explained variance of the latent variable «social youth drinking» increases. In model number 1 we only include variables such as age, sex and number of friends, in model number 2 we added family support and group of peers, and in model number 3 we also included the four dimensions of the positive and negative beliefs on alcohol (see the decrease of the residual D1 for the three models). In model number 3 (see Graphic 6), the only significant effect of the four dimensions of beliefs on alcohol on social consumption is related to variable «Beliefs 2» (t=6.1 p