The right scale for an employee survey

Today we are taking a closer look at one methodological aspect of surveys. In this article, we clarify questions relating to the topic of response scales, i.e. whether there is the right scale for a survey.

Is there the right scale for an employee survey?

We are often asked by our customers what the "right scale for an employee survey" is. It's mainly about the number of response categories and whether there should be a "don't know" or "no answer" response category.

Results from the online survey by ValueQuest

In an online survey, ValueQuest investigated how different forms of presentation influence the response behavior of employees. The employees surveyed were shown different variants of the scale at random. The following graphic shows the schematic representation of the different variants:

Fig. 1: Different response scales for employee surveys

In our online survey, we applied the selected variant to the entire questionnaire. The "Don't know" response category was not taken into account for the average.

If the assumption that the scale does not play a role were correct, the results would have to be statistically identical for all groups with approx. 170 respondents per variant. However, as the following graph shows, there were actually differences in response behavior, and quite consistently across all questions.

Fig. 2: Differences in response behavior for different scale types

Why do variant 2 (from right = bad to left = good) show the highest values? Respondents answer questions intuitively based on experience. They do not think long and hard about which score they want to give, but tend to answer on the left or right. There are therefore clear differences depending on the answer scale - this must be taken into account when creating questionnaires. Caution is also required when comparing with other studies.

In the two variants with the category "Don't know" (variants 3 and 4), there are only minor differences between variants 3 and 4. With regard to "Don't know" in general - i.e. tick "Don't know" if it is present or leave the question blank if "Don't know" is not present - there are hardly any differences between all variants. The number of "Don't knows" was around 2-3% in all four variants. The answer category "Don't know" therefore offers hardly any advantages over variants 1 and 2. We therefore usually recommend that our customers keep the questionnaire visually leaner and do without this option.

Conclusion on the different types of scales

The study shows that there are indeed differences between the various scales, but they are not very pronounced. It is important that the selected scale is retained within a questionnaire - otherwise this confuses the respondents and distorts the results. Cultural aspects also play a role. In Switzerland, scales are generally presented from left (bad) to right (good) (analogous to variant 1), while in Germany the scales tend to be presented from left (good) to right (bad) (analogous to variant 2).