To collect some data for restaurants recommendation I want to ask users if two restaurants are similar or not.

I want to keep the question as simple as possible, so I would like to stick to simple Yes/No/Don't Know question. However, I'm afraid it might be to rigid: people may think that if there's slight difference between restaurants, they will mark it always as No (not similar). For instance, they may consider Chineese and Thai restaurants as dissimilar, even if they both offer Asian cuisine.

Likert scale, on the other hand, may give more flexibility:

  1. Very dissimilar
  2. Slightly dissimilar
  3. Neither similar nor dissimilar
  4. Slightly similar
  5. Very similar

However, it can make people less likely to contribute because it's more complex: it might be hard to decide whether a restaurant is very similar or only slightly similar.

Which one should I choose?

  • 1
    What are you going to actually do with that data? If a restaurant attracts several 'Slightly Dissimilar' rating and a few more 'Very dissimilar' ones, does that actually make a difference to what you do with that information compared to if they'd attracted far more 'very dissimilar' ratings?
    – JonW
    Mar 16, 2020 at 15:52

2 Answers 2


Instead of "slightly", "somewhat" is better fit for scales. Also with nice design, and not overwhelmingly large number of question, you should get similar interest like with yes/no.


Since the goal is to create recommendations through comparisons, you could do something like:

  • Which restaurant has better food, overall? [Restaurant A] [Restaurant B] [Both are about the same]

  • Which restaurant is better for kids and families? [Restaurant A] [Restaurant B] [Both are about the same]

And so on.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.