Terms, names and values
Definitions, key terms, thresholds and figures have tight answers that auto-grade cleanly once variants are listed.
A prompt and an empty box: the participant types a word, number or phrase. No options, no guessing, and it still auto-grades.
A short answer question asks the participant to produce a brief response, typically a word, name, number or short phrase, rather than choose from options you provide. Because nothing is offered to recognise, it tests recall: the participant retrieves the answer from memory, a higher bar than picking it from a list.
That distinction between recall and recognition is the reason to reach for the format. A multiple choice question can be answered by someone who would never produce the answer unprompted; a short answer question cannot. When it matters that a participant knows a term rather than just spots it, short answer is the truer measure.
The trade-off is that free text is messier to grade than a tick box. People spell things differently, add spaces, or phrase the same answer two ways. Short answer only auto-grades reliably when the set of acceptable responses is small and predictable, so it suits questions with a tight, well-defined answer rather than open-ended ones.
What is the minimum width, in millimetres, of a marked pedestrian walkway in the warehouse?
Accepted answers: “1200”, “1200mm”, “1200 mm”. Case and surrounding spaces are ignored, so a stray space or a unit suffix still marks as correct.
Name the document a new starter must sign before they are allowed on the workshop floor.
Accepted answers: “induction checklist”, “site induction checklist”. Normalisation handles capitalisation and extra spacing automatically.
Reach for it when you need recall, and the correct answer is short and predictable.
Definitions, key terms, thresholds and figures have tight answers that auto-grade cleanly once variants are listed.
With no options to choose from, there's nothing to guess: they know it or they don't.
When someone must produce an answer unprompted, short answer measures what multiple choice can't.
If the response could be phrased many ways or needs explaining, use a long answer question.
Long answer questions →Short answer questions grade automatically against a list of accepted answers you define. Add every wording, spelling and format you'll count as right, and a matching response marks correct on submission.
Trivial differences don't cost marks: capitalisation and stray spaces are ignored, so “ 1200” matches “1200” and “Induction Checklist” matches “induction checklist”. The answer is graded on its substance, not its typography.
Auto-grading is only as good as your accepted-answer list, so the dashboard shows the responses participants actually typed. If a legitimate wording you hadn't anticipated shows up, add it and apply a correction, and affected attempts regrade automatically.
You list the answers you'll accept, and TestSetGo matches each submission against that list. Capitalisation and extra spaces are ignored, so formatting differences don't cost a correct answer its marks.
Use short answer when it matters that participants can recall an answer unprompted, rather than recognise it in a list. It removes guessing entirely, but only auto-grades well when the correct answer is short and predictable.
Add every legitimate wording to the accepted-answer list. The dashboard shows exactly what people typed, so you can spot a valid variant you missed, add it, and regrade affected attempts.
Yes. List the forms you'll accept, for example “1200”, “1200mm” and “1200 mm”; spaces and case are ignored, so equivalent responses all match.
Write recall questions with accepted answers, and share a test that grades typed responses automatically.