The data analysis involved three steps: (a) transcribing dyadic conversation sessions; (b) coding CTs; and (c) coding SRs (i.e., AN, AU, and PR) [57 (link)]. The assistants were then paired as R1 with R2 and R3 with R4. For the analysis, first, all the video-recorded conversation sessions were transcribed through transcription notations adapted from the work of Tsai [56 (link)], as documented in Table S1 . All spoken words, nonverbal conversation behaviors (e.g., head nod and eye gaze), and silences within and between utterances were transcribed precisely [62 (link)]. The first two research assistants (i.e., R1 and R3) transcribed the conversation sessions, and the other research assistants (i.e., R2 and R4) examined the accuracy of transcriptions before coding. Any discrepancies in the transcriptions were discussed to gain consensus [76 (link)], and any potential discrepancies were resolved [76 (link),77 (link)].
Second, CTs were independently coded on an utterance-by-utterance basis by R1 and R2 [46 (link)]. A CT-A was coded to a CT contributed by the core HOAs, and CT-B was coded to a CT contributed by the recommended HOAs. Detailed operational definitions of CT coding adapted from the work of Tsai [56 (link)] and extracts illustrating different coding rules of CTs are documented inTable S2 .
Third, three SRs (i.e., AN, AU, and PR) were independently coded on an utterance-by-utterance basis using Goffman’s [43 ] framework by R1 and R2. AN was defined as a person producing utterances (i.e., giving “voice” to the words) [43 ], and AU was defined as a person who selects words or infers meanings from incomplete spoken words, utterances (e.g., go), and/or nonverbal conversation behaviors (e.g., head nod and eye gaze) [43 ,44 (link)]. PR was defined as a person whose beliefs, positions, perspectives, personal information, and sentiments are established during the conversation [43 ,78 ]. An AN-A, AU-A, and PR-A were coded to an SR contributed to by the core HOAs, and an AN-B, AU-B, and PR-B were coded to an SR contributed to by the recommended HOAs. Detailed operational definitions of SR coding adapted from the work of Tsai, Scherz, and DiLollo [46 (link)] and extracted illustrating different coding rules of SRs are documented inTable S3 .
Second, CTs were independently coded on an utterance-by-utterance basis by R1 and R2 [46 (link)]. A CT-A was coded to a CT contributed by the core HOAs, and CT-B was coded to a CT contributed by the recommended HOAs. Detailed operational definitions of CT coding adapted from the work of Tsai [56 (link)] and extracts illustrating different coding rules of CTs are documented in
Third, three SRs (i.e., AN, AU, and PR) were independently coded on an utterance-by-utterance basis using Goffman’s [43 ] framework by R1 and R2. AN was defined as a person producing utterances (i.e., giving “voice” to the words) [43 ], and AU was defined as a person who selects words or infers meanings from incomplete spoken words, utterances (e.g., go), and/or nonverbal conversation behaviors (e.g., head nod and eye gaze) [43 ,44 (link)]. PR was defined as a person whose beliefs, positions, perspectives, personal information, and sentiments are established during the conversation [43 ,78 ]. An AN-A, AU-A, and PR-A were coded to an SR contributed to by the core HOAs, and an AN-B, AU-B, and PR-B were coded to an SR contributed to by the recommended HOAs. Detailed operational definitions of SR coding adapted from the work of Tsai, Scherz, and DiLollo [46 (link)] and extracted illustrating different coding rules of SRs are documented in
Free full text: Click here
