Currently setting up a satisfaction survey for a local Council.
The question of how to deal with the neutral response in 5 point satisfaction questions has created a discussion in the office. i.e. how do we treat the neutral midpoint?
Clients like a % score i.e. 66% of respondents are satisfied but using this method could be misleading as the neutrals are neither satisfied nor dissatisfied, i.e., it might lead the reader to assume that 44% of respondents were dissatisfied where in fact 15% of respondents may have answered neutral.
I'd be interested to know how other researchers deal with this issue.
A Neutral category is considered good for participants who do not have an opinion about a particular statement.
Agree a midpoint is good. The question is more how do we report them. In the Exec Summary we have to say x% were satisfied or x% not satisfied.
Which group do we put into?
Generally, if you aggregate, three groups are used: satisfied + very satisfied; neutral; dissatisfied + very dissatisfied. Obviously, you lose some precision of the 5-point scale, but you gain in parsimony? Is this worth to you? Is the gain important vis-à-vis the loss? You may be tempted to go further in this direction, by subsuming the middle category in one of the two groups. Ask again the former questions. Is the gain worth the loss? Does it make sense to subsume the central category in the satisfied or insatisfied group? The question is mainly theoretical.
Best regards from Madrid
Thanks Andres, discussed using the neutrals as their own group for reporting and actually think this has a lot of benefits.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey