Score, open text and single choice

A scored question renders the instrument scale and produces the number the survey rolls up. An open text question takes an answer in the visitor’s own words. A single choice question offers a list you write and takes one selection from it.

A survey can hold up to five, and almost never should. The shape that actually gets answered is two: the score, then why. Every question after that costs completion, and a survey with a high response rate on two questions tells you more than one with a low response rate on five.

The single choice kind is the one to be careful with. Offering options constrains the answer to things you already thought of, which is efficient when you are confirming something and blinding when you are trying to discover it. Open text is slower to read and the only kind that can tell you something you did not anticipate, which is what sentiment classification exists to make manageable.

Bounds are enforced rather than advisory. Scores are validated against the range of the instrument the survey is using, not against the widest range shared by all three, so a nine submitted to a CSAT survey is rejected rather than stored.