• Think about whether you want qualitative (descriptive) or quantitative (measurable) data – or indeed both. If quantitative, remember that there are dangers in a five-point scale. Learners might tend to neutrality – saying that it was OK – not great and not bad. Denying them the middle point in the scale seems mean, but at least you get a more decisive result.

• When using a quantitative scale for questions, ensure there is a balance of options – for example a scale of ‘Outstanding, Excellent, Very Good, Poor’ doesn't provide an option between Very Good and Poor.

• Qualitative data requires a box to complete. Online this can be expandable but on paper you are forced to have a finite space. Think carefully about whether you are asking too many questions or leaving too much blank space: both can intimidate!

• Only ask questions to which you want an answer. Few people enjoy filling questionnaires in and it is even more tedious analysing them (unless you have an optical mark reader or it can be done online). A few well targeted questions are better than a number of vaguely relevant ones.

• Ask learners to comment on the score they have given if you do use a quantitative scale. Receiving a poor score is useless – you will want to know what was wrong so that you can rectify it.

• Make the text attractive to read – ensure you use a good clear font and provide instructions on how to complete the form.

• If you have asked for self-assessment, e.g. ‘on a scale how experienced are you in —’ then try to find a way to check the score through subsidiary questions. Learners will often inflate or underestimate their experience and knowledge based on a perception of what the teacher wants or on what it might mean for the subsequent session if they answer in a particular way.