In the first data analysis, the answers from both informants from each country were compared by the two authors (D.K.and I.S.) and inconsistencies and discrepancies were identified.
In the second stage, the interpretations with a list of inconsistencies were sent back to the key informants with some additional information; in some instances def-initions were added (e.g. ‘mostly’ was over 50%, ‘almostall’ was more than 80%) and consensus was sought. When consensus was not possible, this was noted and described later in the paper.