Objective: To study socio-demographic and functional features related with non-collaboration in a longitudinal design of mental health within a high-risk population of individuals 9 and 13 years old. Method: Regression analyses were used to assess factors affecting the decision to decline participation, and what characteristics both of children and families increase the probability of dropping out once the study had already started. Results: Refusal of participation at the outset is more probable for lower socioeconomic groups, unemployed families (or with Social Security benefits), minority cultures and children having low school performance. The risk of participants dropping out is higher for adolescents, those who need help at school, are unhealthy, have more life-events, receive professional help for mental problems or have had more psychopathology in previous assessments. Lengthy interviews or evaluations without the return of reports to families are also predictive of drop out. Conclusions: This study has practical implications for reducing the lack of collaboration in the prospective studies that assess mental health in children and adolescents. Improvement in the estimation of epidemiological indices requires the planning of special measures for research projects carried out on populations with fewer resources so as to recruit individuals with lower SES, adolescents, individuals with pathologies (physical or psychological) and those with lower levels of school achievement. © Steinkopff Verlag Darmstadt 2007.
- Drop out
- Longitudinal prospective design
- Mental health