How happy are you with the life you have?
How satisfied are you with the latest album of your favorite band? What is your general attitude toward your current government? How frequently do you have negative thoughts?
Regardless if in psychology, marketing or simply life in general, it is difficult for us to express with precision how we think about abstract concepts, ideas, feelings, emotions, intentions and so much more.
Rensis Likert was an American social psychologist (PhD in Psychology from Columbia Univeristy). During his lifetime, Rensis Likert had some pretty interesting accomplishments. He worked at the Department of of Agriculture, Office of War Information and event at the United States Strategic Bombing Survey Morale Division.
However, Rensis is most renowned for creating psychometric (the study of psychological measurements) scale, useful for allowing us to express ourselves in an objective manner.
They have been named after it creator, being coined as: Likert Scales.
In order to understand Likert scales, it is important to differentiate two types of scales:
1. Scales to Measure Theoretical Constructs
As described before here on the site, the Marketing Scales Handbook for example, contains a compilation of scales that measure theoretical constructs (or concepts). These include concepts such as: satisfaction, attitude, purchase intention, risk, quality and many more.
The measurement of each concept is made by the evaluation of multiple “items” (normally at least 3 items). Each “item” is normally framed as a statement or a question.
For example, Lovibond and Lovibond (1995) developed a scale with 7 “items“, to measure “anxiety“. The items are:
1. I am often aware of the action of my heart in the absence of physical exertion (e.g,
heart racing, skipping a beat).
2. I often experience dryness in my mouth.
3. I often experience difficulty breathing (e.g. excessively rapid breathing,
breathlessness in the absence of physical exertion).
4. I often experience trembling (e.g. in the hands).
5. I worry about situations in which I might panic and make a fool of myself.
6. I often feel close to panic.
7. I feel scared without any good reason.
And each of these 7 “items” of the scale is measured through a ranking.
2. Liker Scales: The Ranking Through Which Responses are Scored
The rankings through which the items of construct scales are measured is what we refer to as: Likert scales.
Essentially, Likert scales are rating scales which contain a series of “anchors” (displayed numerically or in words) that allows numerical measurements of an item or question.
Here are a few common Likert scales examples of:
-
- Agreement (5 points): Fully Disagree – Disagree- Neither agree nor disagree – Agree – Fully Agree.
- Acceptability (7 Points): Totally unacceptable – Unacceptable – Slightly unacceptable – Neutral – Slightly acceptable Acceptable – Perfectly Acceptable.
- Desirability (5 point): Very undesirable – Undesirable – neutral – Desirable – Very desirable.
- Difficulty (5 point): Very difficult – Difficult – Neutral – Easy – Very easy.
- Frequency (5 point): Never – Rarely – Sometimes – Often – Always.
- Likelihood (5 point): Extremely unlikely – unlikely – Neutral – likely – Extremely likely.
- Satisfaction (5 Points) : Fully Dissatisfied – Dissatisfied – Indifferent – Satisfied – Very Satisfied.
DOWNLOAD: In this fantastic review (Vagias, W. M. (2006). Likert-type scale response anchors. Clemson International Institute for Tourism & Research Development, Department of Parks, Recreation and Tourism Management. Clemson University) you can download a long list of Likert-type anchors for your work.
Video of the Article
How Are Likert Scales Used in Marketing?
Exactly as in psychology, Likert scales are extremely often used in marketing to measure perceptions, emotions and behaviors (in this case, mostly of consumers or employees).
The Likert scales are then inserted into “structured questionnaires” (questionnaires that contain pre-defined or “closed” questions and answers) in association with scales which measure specific theoretical constructs, or simply to help answer closed questions.
Thus, two frequent applications of Likert scales in marketing are:
1. Descriptive surveys
-
-
- e.g. customer satisfaction, employee motivation, brand perception and awareness.
-
2. Questionnaires for experiments and evaluation of product testings
-
-
- e.g. evaluation of product performance, prototype variations or uses under different conditions.
-
How Many Points Should a Likert Scale Have? 4, 5, 7 Points?
This is something my students often ask. You will find a great deal of works that will contain 5-point measurements, others will contain 7-points and a few will even go as far as 100. It is not “wrong” to measure it with certain point numbers, but you must be aware of a few issues.
Here are important aspects which must be considered:
-
- Odd Numbers (allowing a middle-point): You will find often Likert scales with “odd numbers” (e.g. 5,7,9). This way it will allow a “neutral” value, usually labelled as “indifferent”, “neutral”, “neither agree nor disagree”, depending on what the scale is measuring. This is especially important when you are measuring consumer perception in which respondents need to position themselves towards a topic or subject.
And why? Because we do not necessarily have to “agree” nor “disagree” with something. We can simply not want to position ourselves because we do not have a position. Or, it can also be that the respondent is simply unsure.
Therefore, when measuring opinion or perception, it would be a mistake to “force” respondents to either position themselves positively or negatively.
-
- Even Numbers (not allowing a middle-point): However there are moments in which you can use Likert scales with “even values” (e.g. 2, 4, 6, 8). For example, when measuring intensity of a feeling or of a problem.
For example, see this 4-point Likert scale for “problem”: Not at all a problem – Minor problem – Moderate problem – Serious problem.
In this case the scale only contains 4-points. And it is completely ok, as it represents the intensity of a perception. The respondent would not have to position themselves in on way or another.
-
- Consistency: If in your project you measure a concept with 5-points, you must also measure the others with a similar range. Otherwise, you will not be able to compare, for example, a concept that was measured on a 5-point with another on a 7-point. Moreover, when you plot them on a graph it will be misleading as they were measured differently. So on a single project, I suggest being consistent!
- Consistency: If in your project you measure a concept with 5-points, you must also measure the others with a similar range. Otherwise, you will not be able to compare, for example, a concept that was measured on a 5-point with another on a 7-point. Moreover, when you plot them on a graph it will be misleading as they were measured differently. So on a single project, I suggest being consistent!
Are Likert Scales Ordinal or Interval?
My dear friend, you entered the space of controversy. Some scholars defend that Likert scales as interval, as others argue that it has the properties of an ordinal scale.
The main aspect of discussion lies in the “distance between the values”. In interval scales, the distance between values are required to be the same and constant (e.g. we would have to assume that the distance between “Fully disagree” to “Disagree” would be the same distance as of “Disagree” to “neither agree nor disagree” and so on).
In ordinal scales, the distance between values is unknown or not constant.
And this is especially important because the property of values resultant from the scale will impact which kind of test (e.g. non-parametric or parametric tests) can be applied and how the data should be treated.
Thus, a Likert scale would only be said to be interval if the labeling of anchors allows for a symmetric and equidistant distribution. And this is especially difficult when the interpretation of anchors might be different. So in essence, a Likert scale is ordinal.
Nevertheless, you will find an abundant amount of works in marketing and across different fields using Likert-type scales as interval data. This is supported by further studies which indicate the robustness of statistical tests in dealing with this violation.