Blog

Questionnaire Design: Response Scales - Part III

Part III: Choosing a Response Scale – Anchors aweigh!

So far, I’ve addressed why to use a response scale, and some things to consider in deciding the number of points in the scale. Now, how to label it?

Do you put verbal definitions on every point in the scale (AKA ‘fully anchored’), or use a partially anchored scale, say, only defining the two opposite ends, or maybe the two ends plus a definition for the mid-point if you have one? 

A picture’s worth a thousand words

Some scales don’t even use verbal descriptions, and rely on diagrams for conveying meaning, such as simple line drawings of faces with various expressions (e.g., big smile, medium smile, straight-line mouth, medium frown, and big frown). These types of scales may be successful with audiences that are pre-literate, illiterate or for whatever reason have difficulty expressing themselves with words – of course these situations may require an interviewer to assist in administering the questions. However, even with respondents who are able to read and understand the language of the questionnaire, such scales may make the experience more fun or engaging, depending on the situation. 

Too many words spoil the broth

However, most scales use words to define their points. A general rule of thumb is that it seems appropriate to fully anchor a scale if there are five or fewer points on it, and for scales with six or more points, a bi-polar anchor can do nicely. There may be cases where 6+ point scales can still be completely anchored on all points, but in some cases this can become tedious for respondents to read through (or in telephone surveys, to have it read out loud for every question).  On self-completed surveys, having so many labels may compromise the visual appeal, and the questionnaire may appear more cluttered-looking, longer, or confusing to the respondent.

Two’s company, three’s a crowd

You may wonder if it is helpful, when using bi-polar anchors, to also label the mid-point in order to provide another point of reference for the respondent. Here, I would argue that you may be introducing risk by calling unnecessary attention to the mid-point: this emphasis on the mid-point may tempt respondents to use it as a ‘way out’ from having to think about their response and provide an opinion. Mid-points can also be a tempting fall-back for lazy or fatigued respondents, and in my opinion the less we do to call attention to that half-way point in the scale, the better.

1 Comment

Just now coming across your blog, great stuff.  Could be I didn’t look hard enough, but I didn’t see anywhere I could subscribe to an RSS feed for your blog.  Love to add you to my reader, appreciate you sending along the feed info.

Leave a Reply