Psychology has generally dealt with child development in terms of the socialenvironment. This is no longer appropriate. We have to look at both thephysical and the social environment, and not only in home but at day carecenters as well, where children in the western society get a large propor-tion of their daily care. In order to do so we may use well-proven techniquesin new ways. The studies carried out so far show that it is possible to employsemantic scales with only two or three subjects making the judgementswhich also opens up the possibility for large-scale studies.This present study consists of two pans. Firstly, a study of the environmentat the day care centers using assessment scales in the conventional way, andsecondly a study of the social and the physical environment of the home usingsimilar scales, but in a new way. The study is based on the research modelproposed by KUller (1987) where both the physical and the social environ-ment is taken into consideration. While the child is considered to be activat-ed by both the environment and the social network, this will be mediatedthrough the child's resources, constitution, earlier experiences and strate-gies. In order to assess the physical environment we used semantic scaleswith both the parents and the staff at two day care centers, one situated in ablock of flats and one situated on freehold land (Laike in prep). We foundnodifferences between the assessment carded out by parents and staff butdifferences in four out of eight dimensions between the two day care centers.This indicates that the test instrument is sensitive even with rather smallgroups (n 11). Turning to the question of the physical and social environ-ment of the home one major problem was that the number of judges was evensmaller. The aim was to find an appropriate way to assess the environmentwhich was not cumbersome to handle but still reliable and valid. A brief lit-erature review showed that most existing methods were complicated to ad-minister and very time-consuming. Only one method met our demands, thesemantic scales. However, these were never before used with only a fewjudges. We therefore conducted a study involving fifteen families where bothmothers and fathers made the assessment. The physical environment was as-sessed by the interviewer as well. The results were then analyzed accordingto the interjudge reliability and construct validity with Pearson's product-moment correlations. We found high correlation between the judges for thesocial environment scales in four out of five dimensions. Also the correla-tion between the size of available space and the dimension of social intensitywas high and negative as predicted by KUller (1988). The results for thejudgement of the physical environment were less clearcut. However, a highcorrelation was obtained between assessed affection and the age of the build-ing as predicted by KUller (1972). To summarize, the results indicate that,used in an appropriate way, it is possible to use semantic scales for smallsamples and still get reliable-group results when assessing the home envi-ronment both seen as physical and as social. Thus it becomes possible to in-clude large groups of participants because of the easiness of administeringthe tests.