CAUSE AND CAUSALITY IN DAY-CARE RESEARCH - AN INVESTIGATION OF GROUP-DIFFERENCES IN SWEDISH CHILD-CARE

One of the most important goals of child care research has been to det ermine whether or not nonparental care has adverse effects on child de velopment. Answering this question involves making causal attributions about the origins of differences between groups of children who have experienced divergent nonparental care arrangements. Some of the probl ems researchers face when trying to demonstrate causal relationships a re illustrated in this paper using data from the Goteborg Child Care S tudy, a comprehensive, prospective, longitudinal study comparing the d evelopmental trajectories of children in exclusive home care, family d aycare, and center daycare. It is argued that the discovery of signifi cant differences between groups must initiate extensive efforts to val idate and interpret the findings. Focusing on the specific characteris tics of prospective longitudinal studies, the article shows what measu res can be taken to avoid the misinterpretation of group differences t hat are actually attributable to pre-selection effects or confounding variables. By rigorously identifying such confounds, researchers can a ccumulate support for hypothesized causal relationships, but they cann ot 'prove' them because unknown and unmeasured confounding variables m ay always have important effects.