We are capturing emotions as survey responses. We need to assign values for the responses(emotions) for analysis purposes. Is there an optimum range that can be assigned to achieve this? (like from -100 to 100). An example of a question and a set of answers are as follow.

Question: "How are you feeling today?"

Answers: Terrible, Sad, Ok, Good, Great

A suitable approach we could think of is to assign values from 1 to 100 with equal distance. Is this statistically valid? What are the things that we should consider when achieving this? In this case, only positive integers are assigned because we need to calculate statistics like weighted average. Can't we assign negative numbers as well?


The final range of emotion is completely arbitrary. No matter the interval [a, b], you can adjust the emotions to fit inside. [-100, 100] is perfectly reasonable and is common. An example of use is from GDELT, which provides this interval for average tone of news documents.

Asking if equally distancing the emotions is statistically correct does not make sense. This entirely depends on your use case and opinions. Also, there is absolutely no reason why you can't use negative numbers in a weighted average.

If you mentioned what you are doing and how you are evaluating emotion, there might be more to say.

| improve this answer | |
  • $\begingroup$ Thank you for the reply. We do customized surveys and the customers who use the surveys can assign any value to represent the answers (emotions). We need to define a range (like [a,b]) then we can map the numbers assign by users to this range so we can compare the behaviors of two users, even they have assigned two different numbers to represent the emotions. $\endgroup$ – user33293 Jun 21 '17 at 5:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy