How do you process open-ended questions?
We first process open-ened questions to ensure that there are no nonsense answers in the data. There is an instruction to respondents included next to the question that advises respondents to answer “no” if they cannot think of any answers to the question provided.
Once all responses are collected, we run a script to check for any nonsense answers and clean anything out that is not a proper answer/looks like nonsense. We then sense-check the results following this cleaning, to ensure that no valid responses have been inadvertently cleaned and that no nonsense answers still remain in the clean data.
In terms of how we deliver the responses to the open-ended questions, we have different possibilities for SME clients and Enterprise. For SME clients, we classify the answers according to the list of brands provided for tracking for the survey. The results of this are then grouped and can be seen in the Latana dashboard provided at delivery after MRP calculations. For Enterprise clients, we can offer a raw data file providing the cleaned open-ended answers from respondents or we can classify answers given according to a pre-agreed list, and inject the data into a separate question. This would then show percentages of answers/numbers of answers per answer option in the separate question.