Skip to main content
. 2019 Aug 28;4:12. doi: 10.1038/s41539-019-0051-3

Fig. 4.

Fig. 4

Coding scheme 2: fine-grained scheme. a Coding scheme. Responses to questionnaire questions are used to code 44 knowledge traits in three channels each. Some of these traits (boxed in gray), with predefined codes at two or more levels of knowledge, are scored from 0 to 1. Finally, to prevent some aspects of knowledge from being over- or underrepresented later in score differences, scored traits belonging to similar topics are combined into 11 dimensions and their scores averaged. b Comparison with Coding scheme 1. We compared Coding scheme 1 used to form dyads and Coding scheme 2 used for data analysis. Each point represents one participant’s interview, with its Coding 2 or mean dimension score and its Coding 1 or mean axis score. A strong correlation between scores from both coding schemes was found. Histograms next to each axis represent score frequencies, with kernel density smoothing function fit in dashed lines