Introduction

Behavioural problems in adolescents are thought to be relevant as strong predictors for the detection of other psychological disorders. For this reason and due to the importance they present by themselves, carrying out an adequate assessment of them is fundamental. Mental health professionals have diverse opinions about the value and importance of the different informants. The majority choose of their sources according to the disorder and necessities of each evaluator. On the other hand, the need to obtain data about adolescents' functioning from multiple resources has been emphasised and numerous reasons have been exposed. Concretely, the fact of carrying out the most objective and complete evaluations as possible has been considered essential in those studies aimed at evaluating behavioural alteration in adolescents. For this reason, Achenbach developed three versions of his scale: one for the parents, another for the teachers and a third one for the adolescents themselves. Numerous investigations have studied the concordance between groups of informants about different behavioural alterations in adolescents, but none have carried out a complete analysis of all informants in all subscales (not only the total ones). For this reason, the current study has been developed with the aim of contributing to obtain an enriching vision for the professional in the field.

Objectives.

a) To systematically explore agreement patterns between adolescents, teachers and parents who inform of behavioural problems in adolescents in the general population and b) in those cases in which no agreement is found, to analyse the level of disagreement between each pair of informants for each subscale.

Methodology.

Cross-sectional and descriptive study

Participants. The study was formed by 160 triads of parents, teachers and 13-16 year old adolescents selected from several schools in Barcelona.

Instruments. The three forms of the Achenbach scale to measure behavioural alterations were applied. The scale was translated into Spanish by the Unit of Epidemiology and Diagnostic in Psychopathology of the Development of the Universidad Autónoma de Barcelona: Youth Self-Report, self-evaluated, Child Behaviour Checklist/4-18 and Teacher's Report Form, both heteroevaluated and completed by parents and teachers, respectively. These three forms contain a 89-item set that evaluates the same behaviour, where eight items are organised in scales of syndromes derived empirically and which are invariant throughout informants. The eight subscales are: withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, delinquent behaviour and aggressive behaviour. Some of them are grouped in second order factors: the first three in internalising, the last two in externalising, and the rest of them provide a total problems punctuation.

Statistical procedure. Agreement values were analysed for each pair of informants and each subscale through the Intraclass Correlation Coefficient (ICC). A value below 0.40 indicates low concordance. In these cases, the statistical analysis proceeds with the discordance analysis by pairs of informants and for each subscale through the Bland Altman Method.

Results.

A low concordance (below 0.40) between informants was found especially in internalizing scales (0.230). A slightly higher value was found in attention (0.334), aggressive behaviour (0.371), externalizing (0.357), and total subscales (0.327). Secondly, it was observed that, when informing about somatic complains, thought and attention problems, internalising items and the total scale, parents reported more alterations, followed by adolescents and teachers. Also, parents indicated more withdrawal problems in adolescents, although in this case they were followed by teachers and adolescents themselves. Finally, in the evaluation of the anxiety/depression scales, social problems, delinquent behaviour, aggressiveness and externalising conducts, adolescents informed of more alterations followed by their parents, and then by the teachers. Regarding the agreement/disagreement variability throughout the scales scores, the discordance between different informants was higher when the punctuation was further away from normality, generally when the scales were scored higher.

Discussion.

The normative criteria of comparison and the reference frames for each group of informants are different. For instance, the fact that teachers report less behavioural alterations could be explained because of their familiarity in dealing with adolescents and a higher tolerance towards some behaviors. In general terms, this result fits in with most conclusions from investigations carried out in this field. On the other hand, the fact that parents inform of more internalising problems could be attributed to adolescent behaviour which would in turn alter the family context. Another explanation might be that parents are on the whole more implicated and more sensitive in detecting certain conducts or behavioural alterations in their offspring. In any case, it is disputable whether the lack of concordance between the different informants does really exist or, on the contrary, adolescent behaviour changes depending on the context. Finally, a result contradicting those found in the studies reviewed is that adolescents are the ones who report more externalising problems. Other authors have found that adolescents inform more about internalising problems, something which should be expected taking into account that they are the ones who know themselves better. This could be possibly explained by the presence of more social desirability/undesirability among the adolescents of our sample in front of their pair group when answering to the evaluation scales; this may be due to the group context in which the case was applied. The main limitation of the present study that it was carried out with a general population sample, although from another point of view this may be considered as a gain of the study. We recommend carrying out explicative studies about discordance, which could clarify the predictive validity of each informant group and make variations in the type of sample under study.

Conclusions.

Data from different sources contribute with specific information of relative validity. This is why a multidimensional, multisituational and mulitiinformant approach is fundamental. This is necessary not only to evaluate behavioural alterations in adolescents within a research context, but also when taking diagnostic decisions in a clinical context, because, depending on the chosen informant, the diagnostic criteria for one disorder or another might change. Also, our results imply that there may be an underdetection of behavioural problems in adolescents by the adults, which would result in a lower psychological demand than the necessary.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )