Ten-minute video recordings of the NPCI were obtained for each child in the families’ homes. Before the video-recorded observation began, a research clinician described the purpose of the observation and the recording procedures. Parents were asked to play with their child the way they typically do. For Study 1 (2016–2018), the families used the child's own toys or other items available in the home; for Study 2 (2018–2020), the dyads were provided with a standardized set of developmentally appropriate toys (a drum, farm animal and tractor set, musical piggy bank, blocks, baby doll with accessories, shape sorter, ball, puzzles, snap lock toy, and picture books). Video recordings of the SCCI were obtained for each child during administration of the CSBS DP. Both the NPCI and SCCI were digitally recorded by a research clinician using an iPad2 for a wide-angle view and hidden camera glasses worn by the adult (parent or clinician) to capture child eye gaze.
The recorded NPCI and SCCI assessments were sent to the lab where the two streams of digitized videos (i.e., iPad 2 and glasses) were time linked, transcribed for communication (including gesture), and coded for various social and communicative measures using the conventions of the Child Language Data Exchange System (CHILDES; MacWhinney, 2000 ). Transcription was conducted at the level of the utterance and included all verbal, vocal, and gestural behaviors bounded by a pause or change in conversational turn (Pan et al., 2005 (link)). For Spanish-speaking dyads, bilingual (English/Spanish) research assistants transcribed in Spanish and provided English translations on a secondary coding line. Transcription and coding were conducted by trained research assistants who were blind to group assignment. Specifically, research assistants trained on practice videos until they achieved substantial inter-rater agreement measured by obtaining a Cohen's kappa coefficient of .75 or above. Cohen's kappa accounts for agreement that occurs by chance (Yoder et al., 2018 ). Once reliable, research assistants were allowed to code study videos.
The recorded NPCI and SCCI assessments were sent to the lab where the two streams of digitized videos (i.e., iPad 2 and glasses) were time linked, transcribed for communication (including gesture), and coded for various social and communicative measures using the conventions of the Child Language Data Exchange System (CHILDES; MacWhinney, 2000 ). Transcription was conducted at the level of the utterance and included all verbal, vocal, and gestural behaviors bounded by a pause or change in conversational turn (Pan et al., 2005 (link)). For Spanish-speaking dyads, bilingual (English/Spanish) research assistants transcribed in Spanish and provided English translations on a secondary coding line. Transcription and coding were conducted by trained research assistants who were blind to group assignment. Specifically, research assistants trained on practice videos until they achieved substantial inter-rater agreement measured by obtaining a Cohen's kappa coefficient of .75 or above. Cohen's kappa accounts for agreement that occurs by chance (Yoder et al., 2018 ). Once reliable, research assistants were allowed to code study videos.
