Survey responses – do the “don’t know” really know?
November 4, 2008 at 3:02 pm 2 comments
I’ve written before about survey respones and the use of “don’t know” as an option on a Likert scale. What I said was that in some situations, a person may not have an opinion on a subject – and cannot say if they agree or disagree – so it may be wise to include a “don’t know” option. Well, i just read an interesting article that suggests that people who respond “don’t know” may actually have an opinion – it’s just that they may require a longer amount of time to develop confidence or awareness of their choice. The article gives an example of how the opinion of undecided people can be acurately predicted by creative means:
In a recent study, 33 residents of an Italian town initially told interviewers that they were undecided about their attitude toward a controversial expansion of a nearby American military base. But researchers found that those people’s opinions could be predicted by measuring how quickly they made automatic associations between photographs of the military base with positive or negative words.
Entry filed under: Evaluation methodology, Evaluation tools (surveys, interviews..). Tags: don't know, likert scale, survey responses.
intelligent measurement 
1.
Philo | November 22, 2008 at 4:44 am
Dear Glenn,
How should I handle the empty cells in the 5-point likert-scale questionnaire while totaling the individual responses for the hypothesis testing. Do I need to assign a score like ‘0’ to it or just treat it as neutral and give it a ‘3’? Or, should I discard the respective item in calculation for the testing? Would you kindly help me out for the solution? Thank you.
2.
Glenn | November 22, 2008 at 1:34 pm
Hello,
If I understand your query correctly, you mean where people either select “don’t know” or don’t answer the question at all? In these cases, you would disregard them from your calculation – treat them as “0”. I wouldn’t give them a “3” as then you are interpreting the positions of the respondents. If you do have a significant non-response or “don’t know” response then you should indicate this in your findings – as it could be important in interpreting the results.
kind regards
Glenn