To assess the effect of setting on mother-infant interaction, 24 twelve-month-old children and their middle class mothers were observed on two occasions, a week apart. Half of these dyads were seen twice at home under naturalistic conditions or twice in the lab in a free play situation. The remaining 12 pairs were observed once in each location (order counterbalanced). General level of maternal functioning, but not infant functioning, was greatly affected by setting; mothers attended to, talked to, responded to and stimulated their children more frequently in the lab than at home. In addition, a change in setting tended to disrupt the individual differences in maternal rates of responsiveness observed consistently within a given setting. These results are discussed in terms of the differential demands placed upon mothers in each setting and their tendencies to behave in a more socially desirable manner in the lab. Individual differences in infant behaviors (e.g., vocalization, cry, smile) were more stable across settings than within settings, suggesting that infants were responding consistently to the strangeness inherent within each observational context. The total corpus of results is discussed in terms of the generalizability of laboratory findings to real world settings. (Author/MS)