Receiving actionable data based on people’s attitudes can be challenging. After all, measures of attitude are complex, and more often than not, highly subjective. Luckily, you can use an ordinal scale in your survey to collect useful data about your respondents’ opinions, perceptions, performance, and sentiments. The straightforward ordinal scale is a down-to-earth way to approach abstract questions in your surveys.
An ordinal or “ordered” scale allows you to evaluate a respondent’s attitude towards a subject by using a set of ordered responses. For example, responses can include: “very satisfied,” “satisfied,” “dissatisfied,” and “very dissatisfied.” In an ordinal scale, the order of answer options is what’s significant—you can’t quantify the exact difference between each answer option. The difference between responses like “very satisfied” and “satisfied,” for example, is relative, not exact.
Most of us have plenty of real-life experience with ordinal scales. Ordinal scales can help you do things like:
As you can tell, the ordinal scale works across a variety of use cases. But what does it look like when it’s in use?
While not all ordinal scales are Likert scales like the ones above (or Likert-type scales if you want to get technical) all Likert scales are ordinal. This popular form of survey question offers respondents an ordered range of answers from one extreme to another. Take, for example, these questions from our Employee Satisfaction Survey Template:
How meaningful is your work?
How challenging is your job?
These Likert scale questions measure each employee’s perception of the work they do using various ordinal, i.e. ordered, scales. Other Likert scale questions measure sentiment with a balance of positive, negative, and neutral answers:
Are you satisfied with your employee benefits, neither satisfied nor dissatisfied, or dissatisfied with them?
In the Question Bank, you’ll find numerous survey questions that use ordinal scales. But we’ll walk you through how to write them on your own.
Just follow these steps:
1. Identify a focus for your question by deciding which opinion, perception, performance, or sentiment you’d like to collect data on. Decide whether to use a unipolar scale or bipolar scale. Unipolar scales measure the absence or presence of a single item—”not at all interested” to “extremely interested,” for instance. Bipolar scales ask respondents how their attitudes fall on two different sides of neutrality—”strongly disagree” to “strongly agree,” for example.
2. For unipolar questions, decide which single variable—like the level of “meaning” or “challenge”—to include in your scale. For bipolar questions, decide which two opposing variables—like “agree” and “disagree” or “satisfied” and “dissatisfied”—to include in your scale.
3. Create a set of ordered responses using your variable(s). While the difference between responses is always relative in ordinal scales, try to choose options that are somewhat evenly spaced from each other. For bipolar questions, include an equal number of responses for each opposing variable to avoid skewing your results.
Note: Keep in mind that matrix questions can easily become overwhelming to answer. If you decide to use one, limit its size to 5 rows and 5 columns.
For multiple choice questions that use an ordinal scale, you can look at the responses both individually and collectively. In either case, you can easily compare the relative popularity of each choice to identify key takeaways. Matrix/rating scale questions provide a similar level of analysis but also give you the weighted averages from each choice.
So take the time to write survey questions that use an ordinal scale. The responses will help you align with your respondents’ opinions, perceptions, performance, and sentiments.