In consumer or B2B marketing research, as in life in general, there is typically more than one way to ask a question. Some of these methods are easier on the part of the survey respondent, while others although a bit more complicated yield richer results for our survey data analysis. When we are interested in measuring consumer attitudes toward companies, brands or products we typically call on the tried and true Likert scale. This scale, and many like it, is multi-dimensional and measures with comparatively higher levels of validity and reliability.
Yet to facilitate analysis, and presentations to senior management, we often end up collapsing these scales into a top-2 box or some derivative. If we know that ultimately we will go there, then why not consider a binary alternative up front and save yourself the trouble of re-coding variables. The primary reason for not going there initially is marketers, like most of us in the survey research business, like variation. We also like having the choice to collapse the data or not.
On occasion a binary measure, especially if it is wedged into a long survey, will serve the purpose. Below is an example of a binary attitude measure from a recent mobile technology survey.
What we give up in the ability to measure variation and create our own sub-groups is offset by the ease of analysis and a less intensive experience for the survey respondent. Even with binary measures we can still create a summary scale if the true/false questions are coded 0/1. I would suggest adding one or more negative measures to encourage a more thoughtful process on the part of the respondent.
Binary measures do have their place on a survey, and they don’t need to be limited to yes/no or demographics (male/female). If you are proposing a long survey then consider them as an alternative to lengthier scales, however keeping in mind your analysis needs.