Measurement: Single vs Multiple Item Measure

In my research course, we recently delved into the discussion of scaling and measurement. No lack of material for discussion in this chapter! One of the students asked what my preference was in questionnaire design - single-item measures or those with multiple items. A fair question I thought and certainly one that deserves sharing.

Single item measures are low on cognitive load by design. One question yields one answer, what could be simpler? Well in their simplicity lies their undoing. Yes there are domains that can be addressed by a single question, but most constructs faced in marketing research are more complex. An example I gave to my class related to the experience of staying in a hotel. Certainly we could ask our respondents their overall level of satisfaction and be done for the day, but what are we leaving on the table? This idea is very similar to just reacting to the tip of the iceberg, as I've discussed in a previous post.

A single-item measure such as overall satisfaction may serve as a topline metric, but it can easily overlook the nuances which can make a hotel stay either memorable or horrid. Virtually all products and services have multiple levels of experience and we should be prepared to measure them. This is the cornerstone of key driver analysis, identifying those variables that impact satisfaction or intent to purchase.

Customer loyalty is one of those marketing constructs that cannot be easily answered by a single survey question. Burke, Inc. with their Secure Customer Index® employs three questions based upon a theoretical view of loyalty including: likelihood of repeat purchase; willingness to recommend; and level of satisfaction. The point where these three questions overlap is the domain of the secure customer. They are satisfied, likely to return and will communicate their satisfaction with friends, family and colleagues.

On occasion a single-item measure will suffice, with advocates of Net Promoter Score® agreeing, but more often than not it is best to look at a problem from multiple angles. This spreads the burden of proof across multiple questions and can provide a more reliable and valid response. Before deploying your next survey, step back and reflect on the nature of the problem at hand, and ask if justice can be done by a single question. If not then there are several published multiple-item scales useful for the range of marketing constructs.
blog comments powered by Disqus
Crimes in Design Webinar
Subscribe to our Monthly Newsletter