Have you ever been asked to rate your satisfaction for a product or service based on answers ranging from "Very Satisfied" to "Very Dissatisfied"? If so, then you've encountered a Likert scale.
The question asked with associated response choices, and the individual answers (e.g. "Very Satisfied", "Very Dissatisfied") are the Likert items. The Likert scale is a combination of related Likert items that are similar in topic.
Let's get started with an interactive example below!
The questions and related set of answer choices are referred to as a "Likert item"
Question #1 and #2 combine to make the Likert scale; meanining the combination of multiple "Likert items"
The number next to the answer choice is a weight used to give responses a quantitative or numerical value
(these values are hidden from respondents in a real survey)
|Dissatisfied (1)||Neutral (2)||Satisfied (3)|
|Dissatisfied (1)||Neutral (2)||Satisfied (3)|
Each item (Question #1 and #2 in our example) may be analyzed separately or can be summed to create a group total score. Because of this summed total, Likert scales can also be called summative scales.
If you voted on the above example, you'll see there are results for each Likert item and for the total of both items. The results show people (who voted) are currently more satisfied with their personal growth compared to their professional growth. If you look at the "totals" row, you can see people appear to be leaning toward satisfied for both items.
This "total" is what a summative Likert scale is. It is trying to measure an overall feeling or sentiment towards a product, service, or object. Since a numerical value was assigned for each answer choice, you can sum the selected answers. To illustrate, let's say you voted "Very Satisfied" for both questions. This would give you 10 total points from 2 items. This turns into an average of 5.
The results for this scale are related to "growth at your current job". Each Likert item is a very specific question related to "growth at your current job". A researcher can look at the above results and say "looks like people are satisfied with growth at their current job (from the summative total), but seems like professional growth is lagging. If professional growth was improved people would be more satisfied with overall growth". To draw this conclusion though, the Likert scale must be setup properly.
As mentioned above, the Likert scale should be a set of specific questions (Likert items) all related to the same topic. In our example we are trying to look at "growth from your current job". The two questions (Likert items) are very specific to this. Personal and professional growth are two areas of growth.
An incorrect Likert scale item for our example, would be asking a third question "are you satisfied with your current pay". While this is related to your job, and plays into job satisfaction, it is not related to growth.
Likert item answer choices are assigned in a constant progressive nature, for example from a strong dislike to a strong like. This structure, along with assigning numerical values to the answers, help to ensure that each choice is in fact progressing at some constant level.
Another aspect of setup is the type of answer choices you select. Answer choices should be similar in nature and ask for either satisfaction, agreement, likelihood, importance or any other standard metric. SurveyKing has premade answer banks for the following:
The amount of answer choices is also important. Choosing an even number doesn’t allow for a neutral option. Because of this all of our predefined answers are odd (with the exception of 10 because it contains a neutral option). Having too many choices makes respondents feel overwhelmed. Too few answer choices might not provide enough variance. Generally 5 items is the consensus, based on giving respondents a neutral option, and two extreme options for either a strong like or strong dislike.
Since October 2016, more than 4,000 Likert type questions have been made on SurveyKing. Over 39% of these have used 5 answer choices.
Likert scales are not perfect. Data can sometimes be skewed from respondents doing the following:
Central tendency bias can be explained by respondents "saving up" their "extreme" answers for later questions. Respondents may start the survey with the idea that a later question will meet the need for an extreme answer. Because of this a respondent will answer in more of a "neutral" fashion until they find that later question. Be sure to explain the rating scale to give respondents a clear expectation.
Knowing your audience can in part help to remedy some issues. For example, if you're conducting an employee satisfaction survey, employees may not want to give any extreme answers that are negative for fear of punishment, even if the survey is labled "anonymous". In this example, using three answer choices without extremes may work best.
The Likert scale is best used to measure and evaluate a general topic and then drill down into specifics. Most common uses are for customer sentiment on a specific product, service or experience.
A Likert scale for product feedback for example could have items asking for price, quality, color, and usefulness. You wouldn’t want to include items related to shipping for example, because that’s not directly related to the product itself. Always make sure your items are similar in nature and have an overall theme for the scale.
You have two options here. The best option is to setup a matrix question with rows and columns. The columns are the answer choices and the rows are the questions you want to ask. The results would be like the above example and analysis would be a breeze.
|Very Dissatisfied||Dissatisfied||Neutral||Satisfied||Very Satisfied|
The other option is to setup multiple rating questions. Each rating question would be a Likert item. For this scenario, be sure to include all related questions on the scale on one page. Results for this setup only show the one question. Calculating the overall Likert score would need to be done with a data export outside of the software.
The scale was developed by psychologist Rensis Likert in 1932 for an article in the "Archive of Psychology" titled "A Technique for the Measurement of Attitudes". He was attempting to build a scale that made it easy to judge attitude and opinions.
Previous to this, the Thurstone scale was used. This method required a lot of work and the use of judges. It starts by asking a test group of respondents to write down their feeling towards a subject. The responses are collected and then a set of judges would rate each response from 1-11, with 1 being very unfavorable and 11 being very favorable. The responses are then filtered down to 10 or less. Each response now has a value that is the average from the judges. The final questionnaire is now given to collect responses, asking them to mark what answer choices they agree with. This time, scores are calculated using the judges' values.
Likert was able to prove that while an attitude can take form of many words there can only be a certain range that feeling can be in; generally from one extreme to the other. Likert was able to prove that his method gave similar results to the Thurstone scale but with less work. The results were confirmed in another paper by Likert in 1934.