alid or invalid on a priori ground*y Thus, de-pending on the problem, the laboratory may bean altogether appropriate setting for an investiga-tion and certain real-life environments may behighly inappropriate. Suppose, for example, oneis interested in studying the interaction betweenmother and child when the child is placed in astrange and unfamiliar situation. Clearly the lab-oratory approximates this condition far better thanthe home. Conversely, if the focus of inquiry isthe modal pattern of parent-child activity pre-vailing in the family, observations confined to thelaboratory can be misleading. As I have docu-mented elsewhere in greater detail (Bronfenbren-ner, in press), patterns of parent-child interactionin the laboratory are substantially and systemati-cally different than those in the home. Specifi-cally, so far as young children are concerned, theresults indicate that the strangeness of the labora-tory situation tends to increase anxiety and othernegative feeling states and to decrease manifesta-tions of social competence (Lamb, 1976b; Ross,Kagan, Zelazo, & Kotelchuck, 1975; Lamb, Note3). Possibly in response to this reaction of thechild, parents tend to exhibit more positive inter-i actions toward their children in the laboratory' than in the home (Schlieper, 1975; Shalock, 1956;Belsky, Note 4). In addition, Lamb (1976b;Note 3) reported that the tendency of the infantat home to display more affiliative behaviors (e.g.,looking, smiling, reaching, vocalizing) toward thefather than the mother was reversed in the labora-tory. Moreover, consistent with the arguments ofSroufe (1970) and Tulkin (1972) that the lab-oratory is especially likely to be an anxiety-arous-ing situation for lower-class families, Lamb foundsocioeconomic differences in father-infant inter-action favoring the middle class in the laboratory,'whereas such differences had not been present inI the h o m e.Again, the fact that