Child Development, March/April 2002, Volume 73, Number 2, Pages 528–543 Predictability of Observed Mother–Child Interaction from Preschool to Middle Childhood in a High-Risk Sample Nancy S. Weinfield, John R. Ogawa, and Byron Egeland This study examined predictability of observed parent–child interaction from preschool to middle childhood in 283 mother–child dyads. Participants were welfare recipients enrolled in the Observational Study of the Job Opportuni- ties and Basic Skills Training Program. Structured observational sessions were conducted both at preschool age and middle childhood, and were coded for maternal social behavior, child social behavior, and dyadic interaction. Anal- yses explored direct relations between the assessments; relations between the assessments with possible third- variable influences, such as maternal literacy, covaried out; and moderated relations. Results indicated that ob- served mother–child interaction in middle childhood could be significantly predicted from observed interaction 4 years earlier. Risk status moderated the relations such that those families with greater risk factors tended to show more stability, although this stability was, at times, through maintaining suboptimal functioning. INTRODUCTION Historically, social development research on middle childhood has not been focused on parent–child rela- tionships. From the psychoanalytic designation of this period as one of latency (Freud, 1964), to Erik- son’s (1964) focus on industriousness, to the current proliferation of peer and school research, social devel- opmentalists have turned their attention to other as- pects of children’s lives during this developmental period. Particularly within the realm of structured or semistructured observational research, few para- digms exist for examining parent–child relationships at this point in development. This gap in observa- tional research is particularly surprising, given that observational schemes for examining parent–child relationships are plentiful during infancy, preschool, and adolescence. In this study, we introduced a struc- tured observational assessment and coding scheme that were developed to examine affective and behav- ioral qualities of mother–child relationships during middle childhood. The present research demon- strated that this middle childhood observational cod- ing scheme showed interesting and logical patterns of continuity and discontinuity from an observational assessment of mother–child interaction conducted 4 years earlier. Observational Research: Current Paradigms Observational research provides a perspective on mother–child relationships that can differ in impor- tant ways from the more commonly used maternal- report measures (Weinfield & Egeland, 1997; Zaslow, Dion, & Morrison, 1998). Because they are a part of the relationship being assessed, mothers provide unique “insider” perspectives. The mother’s perception of the relationship may influence her behavior toward her child; thus, her perspective is a valuable piece of information. Observational assessments are generally “outsider” perspectives, in which trained observers who are unaware of other characteristics of the dyad or family code the dyad’s interactions. Structured ob- servational ratings can provide a more objective view of interactions, and have been shown to predict later child outcomes above and beyond the variance pre- dicted by maternal or interviewer report (Weinfield & Egeland, 1997; Zaslow, Dion, & Morrison, 1998). Structured or semistructured observations of parent– child interactions that take place in home or labora- tory settings are widely used in social development research during infancy, the toddler/preschool pe- riod, and adolescence. Within each developmental period, these observations focus on some of the sa- lient issues of the period. During infancy, much ob- servational research has focused on attachment rela- tionships (Cassidy & Shaver, 1999). Other structured observational research during infancy has included infant responses to maternal emotion (Tronick, 1989) and infant expressions of temperament during dy- adic interactions (Rothbart, 1986). During the toddler and preschool periods there is an ongoing effort to capture attachment relationships (Greenberg, Cicchetti, & Cummings, 1990). Research has also focused on ob- servations of toddlers’ compliance to parental direc- tives (Kochanska & Askan, 1995; Rothbaum & Crock- enberg, 1995; Schneider-Rosen & Wenz-Gross, 1990), on maternal vocal responsiveness to toddler cues (Bar- ratt, Roach, & Leavitt, 1996), and on teaching-oriented © 2002 by the Society for Research in Child Development, Inc. All rights reserved. 0009-3920/2002/7302-0013