A special type of 'None of these' is one that asks the respondent to choose their current alternative as an option (e.g., “your current car”, “the airline you flew on your most recent trip”, “Stay with your current energy provider”).
At analysis time, these can be analyzed as any other “none of these” option. However, this is generally invalid. Different respondents’ “current” alternatives will differ in their attributes. By treating the option as being a “none of these” we are ignoring this relevant information.
However, a better approach is to:
- Collect detailed information about each person’s current alternative (e.g., brand, price, and any other attributes).
- After completion of the data collection, modify the design, so that it contains the attribute levels that the respondent saw.
- Proceed to analyze the data as if the alternative is any other alternative (i.e., not a “current” alternative).
While this approach sounds good in theory, it is not really a like-with-like comparison, and likely suffers from the same problems as using 'None of These' Alternatives.
Please sign in to leave a comment.