Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Is there a way to detect random responders in CBC studies?

When we field MaxDiff studies, we detect random responders by looking at the RLH score.  Is there a similar technique possible for CBC studies?
asked Oct 26, 2011 by anonymous

1 Answer

+2 votes
 
Best answer
You can do the same with a standard CBC study.  Take the inverse of the number of concepts in each choice task to find the fit of someone who answers randomly.  

Keep in mind, though, that it is much easier to "game" a CBC study.  In MaxDiff you would need to be consistent across multiple tasks to achieve a decent fit statistic.  In CBC I could still go through your survey very quickly and use a simply rule, such as always picking Brand A, or always picking the cheapest option on each task, to achieve a high fit statistic.
answered Oct 27, 2011 by Brian McEwan Gold Sawtooth Software, Inc. (37,410 points)
Would you please provide more detail on this?
Would you please further elaborate what do you mean by "Take the inverse of the number of concepts in each choice task to find the fit of someone who answers randomly. "?

How would you notice that someone is trying to "game" a CBC study by using ,say always picking the cheapest option, rather than they are extremely price sensitive?
...