Poster #81 - Making sense of behavior- behavioral coding schemes: approaches, strengths, and comparisons

• In Event: Poster Session 17

Abstract

Background
Adapting observational coding could increase measurement validity of relationship quality and parent-child behaviors. Parent-child observational coding schemes are often limited by insufficient cultural sensitivity, the utility of global ratings of behavior versus intensive moment-to-moment coding, laborious and costly training to achieve reliability, threats to external validity, and more. Critical investigation can address the unique added value of specific parent-child observational coding schemes and what they can elucidate about behavior and development. This flash talk provides comparison of 5 coding schemes used across play and stress paradigms within one set of videotaped interactions in a diverse sample to compare the relative utility, strengths, and drawbacks of schemes.

Methods
Videotaped parent-child interactions occurred during home visits. Parent-child interactions involved 20-25min of semi-structured book-share and playtime followed by a series of stress tasks adapted from the Lab-TAB battery. Table 1 presents details and citations for each validated coding scheme.

Parental sensitivity: The Three Bag Task was used to rate the parent-child interaction quality during semi-structured play. Interactions were coded using 7-point ratings of intrusiveness, sensitivity, positivity, negativity, animation, stimulation, and detachment.

Affect: Global ratings on relationship quality and time-bound epochs (20s) were coded using Lab-TAB rating systems for emotional expression across tasks to identify child and parent affect, actions, and verbalizations, as well as qualitative ratings of warmth, responsiveness, and rapport between caregiver and child.
.
Emotional Availability (EA): Parental EA scores were coded using the book-share, play task, and separation-reunion task in four adult EA Scales (sensitivity, structuring, non-intrusiveness, non-hostility) from 1 (low) to 7 (high) and along global ratings as a summary score 1-100 via the Emotional Availability System.

Mentalization-Related Behaviors: Two indicators of parental mentalization, Mental state language (MSL) and Mind-Mindedness (MM) were coded using book-share and play tasks. MSL refers to comments which identify, label, or express words reflective of internal emotional (e.g., afraid, happy), cognitive (e.g., remember, know), and desire states (e.g., want, like) and is assessed during parent-child interactions, including book shares. MM refers to parental use of mental state terms (e.g., emotions, cognitions, desires) to make mind-related comments regarding behavior during parent-child interactions for both appropriate attuned (e.g., "You are frustrated because you can't play with the toy you like," "You like this game" and non-attuned (e.g., "You don't like mommy" as the child moves away to explore toys) comments.

Two trained coders coded videotaped parent-child interactions for all coding schemes. Coders maintained agreement of .80 or higher on double-coded cases (between 20-70%) with team resolution of discrepancies.

Results and Discussion
Coding schemes that consider dyadic interactions explicitly provide a potential advantage to coding schemes focused only on one dyad member. Our variability across schemes and linguistic contexts suggest that coding schemes must be sensitive to cultural contexts, individual variability, and sensitivity to change in parent-child relationship quality and that measurement invariance across culture cannot be assumed. We recommend conceptually driven and culturally sensitive coding schemes, keeping in mind pragmatic and rigorous elements of individual coding schemes to train coders and observers.

Authors

• Andrew (Drew) B. McGee, University of Denver
  Presenting Author

• Neda Senehi, United States Office of Planning, Research & Evaluation (OPRE)

• Amy Dominguez, University of Denver

• April R Vollmer

• Liliana M Bolton

• Eliana Hurwich-Reiss

• Sarah E Watamura, University of Denver