2

I have data for a product rating survey, which requires the respondents to rate a product in five levels: Very Bad, Bad, Regular, Good and Very Good. This survey was applied to several communities of clients.

After taking a look on the data, I observed that a lot of communities have an overwhelming "Regular" preference for the product. Later I found out that "Regular", was the default response for the survey question. Apparently, in some communities user's just didn't bother and left all the questions in the default value.

Is there a proper way of identifying the communities that are indifferent to the survey, so they can be excluded? I was thinking in a simple threshold-based approach (If the community has more than 50% of non-default responses, it is not indifferent), but I was wondering if there is a more principled way.

Guillermo Guardastagno
  • 121
  • 3

1 Answers1

1

I'd recommend converting your responses to the Likert Scale and going from there.

It's a widely researched area so there should be plenty of research on how to best deal with 'so-so' responses.

plumbus_bouquet
  • 340
  • 1
  • 4