The Parent-Child Early Relational Assessment
Introduction
The Parent-Child Early Relational Assessment (ERA) (Clark, 1985, 2006, 2010, 2015) was designed to measure the quality of affect and behavior in parent-child interactions. The instrument uses ratings that are based on observations of 5-minute videotaped interactions during feeding, structured task, or free play. The purpose of the ERA is to attempt to capture the child’s experience of the parent, the parents’ experience of the child, the affective and behavioral characteristics that each bring to the interaction, and the quality or tone of the relationship. The ERA was developed to classify both clinical and research observations and assess in a systematic manner the behavioral style and affective quality believed to be indicators of disturbed relationship patterns. It includes an assessment of the areas of concern as well as areas of strength in the interaction for the parent, the child, and the dyad. Profiles may be developed for use in focusing clinical intervention efforts with families at risk for early relational disturbances, in treatment outcome and program evaluation studies, and in research seeking to document the interactional patterns of high-risk groups of parent-child dyads.
The initial instrument was originally developed to assess and describe patterns of interaction in mothers with psychiatric disorders and their infants and young children and to focus early intervention efforts aimed at improving parenting skills and relationship quality (Musick, Clark, & Cohler, 1981). ERA items were selected and developed based on theory, developmental studies, and clinical observations of high-risk and well-functioning mother-child dyads.
An earlier version, the Mothers’ Project Rating Scales of Mother-Child Interaction, consisted of 53 items (Clark, 1983; Clark, Musick, Stott, Klehr, & Cohler, 1984). The scales were refined to measure therapeutic change and to describe differences in interaction quality between children and mothers with and without a history of psychiatric disorders. The original scales were added to and revised to extend ratings downward to allow for assessment of parental interactions with very young infants. Some components of earlier existing items were dropped. The instrument was further modified in 1985 for use with other high-risk groups in which the infants and/or children brought risk to the relational quality (e.g., premature or medically high-risk infants, children with developmental delays or disturbances), and with additional high-risk parent populations (e.g., mothers and fathers with substance abuse problems, adolescent parents) (Clark, 1985). Items with no variability were dropped, components of some items were removed to improve inter-rater reliability and additional child and parental items were added to improve the sensitivity of the instrument to possible differences in other clinical and community populations.
The revised instrument was named the Parent-Child Early Relational Assessment (PCERA) (Clark, 1985), and includes a total of 65 individual items composed of 29 parental, 28 child, and 8 dyadic items. Further revisions have been made in 2006, 2010 and 2015. The instrument has come to be referred to as the Early Relational Assessment or ERA (Clark, 1999).
The ERA has been used in more than 400 programs and research projects internationally. Psychometric work is ongoing, and reliability and validity have been addressed in both high-risk and normative populations (Clark, 1999; Clark et al., 2004).
Conceptual Background
The specific items or behaviors selected for inclusion in the ERA were derived from psychodynamic, self-psychology, attachment, and soviet cognitive-linguistic theories, as well as results from previous developmental studies and clinical observations of parent-child interactions of high- and low-risk samples (Clark, Paulson, & Conlin, 1993; Clark, Tluczek & Gallagher, 2004). The conceptual framework and particular items thought to be important for determining areas of strength and concern in parent-child relationships were drawn from the theories of Winnicott (1965) (genuine visual regard); Sander (1964) (negotiating a fit); A. Freud (1970) (auxiliary ego/mediating the environment); Kohut (1971) (mirroring); Vygotsky (1978) (scaffolding/structuring); Bruner (1974, 1975) (joint attention); Emde (1980) (emotional availability); and Stern (1985) (affective attunement) and the empirical studies of Ainsworth, Bell, and Stayton (1974) (sensitivity and responsiveness); Clark-Stewart (1973) (social responsivity); Lewis and Rosenblum (1974) (contingent responsivity); Demos (1982) (positive affect); and Tronick (1982) (affect matching).
The quality of interaction, rather than the quantity of interaction, may be more relevant to the young child’s experience of his primary relationships. For example, mothers can often “go through the motions” (e.g., verbalize a great deal of commands but not use language to communicate, to extend, or to elaborate the child’s verbalizations or may relate with flat or negative affect). Thus, a macro (e.g., global rating of 5 minutes of interaction) rather than a micro (e.g., time sampling, frequency count) assessment approach was used to attempt to capture meaning and content. The ERA extends the work of previous investigators by asking raters not only to look at concrete and discrete behaviors but also to “organize and synthesize from a greater number of cues to a more abstract construct” (Yarrow & Anderson, 1979). Although acknowledging the contribution of the child to the quality of parent-child interactions in the earliest years, the mother’s behavioral patterns and affective involvement are viewed as contributing significantly to the child’s development by facilitating optimal growth of emerging abilities and a sense of competence in his or her interactions with people and things. In particular, a mother’s positive affect has been found to serve an organizing function for her infant, contributing to the infant’s emotional and behavioral regulation (Demos, 1982).
Review of Validity Studies
Discriminant validity of the ERA has been established through demonstrated differences between various groups of high-risk and well-functioning parent-infant dyads (see Table 1). For example, mothers with psychiatric diagnosis were differentiated from mothers without diagnosis on ERA scales of maternal positive affective involvement and responsiveness and maternal consistency (Clark, 1983); low-income parents of non-organic failure to thrive children as compared to parents of children with adequate growth have been found to exhibit less nurturant, more neglectful behavior as measured by ERA scales (Black, Hutcheson, Dubowitz, & Berenson-Howard, 1994); and mothers with histories of drug-abuse during pregnancy as compared with drug-free controls were rated lower on enthusiasm and responsivity to infants’ cues, and their infants were found to express less positive affect (Burns, Chethik, Burns, & Clark, 1997). Concurrent validity has been documented with ERA subscales correlating significantly with measures of similar constructs including the security of an infant’s attachment to the mother as defined by the Attachment Q-Sort (Waters & Deane, 1985) and the Ainsworth Strange Situation (Ainsworth, Biehar, Waters, & Wall, 1978), as described by Teti, Nakagawa, Das, and Wirth (1991); with maternal internal working models (George, Kaplan, & Main, 1985), as described by Das Eiden, Teti, and Corns (1995); with the Parenting Expectations Profile (Stoiber, 1993); with the Early Coping Inventory (Zeitlin, Williamson, & Szczepanski, 1988), as described by Stoiber and Houghton (1993); and with the Sense of Competence and Reinforcement of Child subscales of the Parenting Stress Index (PSI) (Abidin,1986), as described by Grych and Clark (1999).
The ERA’s sensitivity to affective and behavioral change has been found in numerous therapeutic intervention outcome studies, for example, with sleep-disturbed toddlers (Minde, Faucon, & Falkner, 1994), with depressed mothers and their infants (Clark, Keller, Fedderley, & Paulson, 1993; Clark, Tluczek. & Wenzel, 2003; Goodman et al 2008; Clark, Tluczek, & Brown, 2008), and with adolescent mothers and their young children (Stoiber, 1993). The ERA also has been found to be sensitive in a non-clinical population to differences in mother-infant and father-infant relational quality in families of employed and non-employed women (Clark, Hyde, Essex, & Klein, 1997; Grych & Clark, 1999).
An exploratory factor analysis conducted for the study of mother-infant interactions during feeding at 4 months of age supported 6 scales (two parent, two child, two dyadic). A replication with an older age group and with an age-appropriate task (free play) using both exploratory and confirmatory factor analytic methods was then designed. Although the ERA is widely used in clinical and research programs to discriminate parent-child dyads with and without early relational disturbances, a study that clarified the factor structure of the three components of this instrument was required to determine the construct validity of scores on the ERA in a normative sample of 12-month-old infants and their mothers during free play and presents data using more sophisticated levels of analysis.
In the study to determine the factorial validity of scores on the Parent-Child Early Relational Assessment in a normative population of mothers and their 12-month-old infants, parent, child and dyadic items scored from free play interactions were analyzed as separate components of the instrument (Clark, 1999). Scores on three parent, three infant, and two dyadic subscales were examined for reliability and convergent and discriminant validity using confirmatory factor analysis. To provide confidence in the results, a factorial invariance study using multi-group confirmatory factor analysis was conducted using a separate sample. All 8 subscale scores demonstrated high levels of internal consistency, with coefficients ranging from .75 to .96. Evidence was also found for convergent and discriminant validity.