Wednesday, 18 March 2015

Translation and coding of interviews

Back in January Syeda discussed her project (Exploring information literacy practices in primary schools: APakistani case study).  Unfortunately I couldn’t attend, but when I talked to her about it afterwards, she mentioned that, at one point, the discussion had taken an interesting turn, but that there had not been time to explore it.

The unexplored turn revolved around questions of when, where, and what to translate.  Many of Syeda’s interviews were in Urdu, but she translated them into English, and then coded.  Apparently her decision raised questions about whether or not she should have coded her interviews before translation. 

We continued the discussion in the February meeting.  Translation proved to be an issue in more ways than one, since those researchers with a programming background understood coding in a different way from the qualitative researchers attending.  However, it emerged that the qualitative researchers who interviewed in a language other than English had a range of views.  Their positions, it appears, depend on the nature of their research.

The clearest difference was between Syeda and Kondwani, whose project (Experiencing HIV and AIDS information: a phenomenological study of serodiscordant couples in Malawi) was discussed last November.  Kondwani codes first, and then translates.  However, as he pointed out, a key difference between his research and Syeda’s is that his study, being phenomenological, focusses on meaning, while hers focusses on processes. 

Many of Kondwani’s interviews use metaphors and euphemisms that don’t readily translate into English, while Syeda’s involve descriptions of materials and practices.  For her therefore, there was far less chance of important findings getting lost in translation.