Introduction to Likert Scale Data Analysis
Business will have to constantly look for transformations in their processes, Product lines, Customer services and brand image to thrive in the system. Data plays an important role in its transformation journey to achieve the end goal. Business cannot sail in this exercise with subjective and hearsay information. They will have to collect objective information on their current products, services, environment and competition from all its stakeholders to start the journey.
Surveys, Questionnaire and interviews are some of the channels through which feedback from all stakeholders are elicited. The response in the feedback may be quantitative data or non-quantitative. Both these types of data will have to be analyzed to gain useful insights. The insights will be used to drive business transformation.
Data analysis can be done in multiple ways viz.,
- Descriptive analysis just by looking at the results and act upon it,
- Diagnostic analysis to study the root cause and correct,
- Predictive analysis to research data and forecast the future,
- Prescriptive analysis to suggest plan of action.
Similar Surveys and Questionnaires are deployed as psychometric tests in measuring the skill, knowledge, traits and capabilities of the people in the Job selection, promotion and special assignments.
Likert scale is the way of measuring and analyzing the responses in the surveys and tests. In this article let us study the features of Likert scale.
What is Likert Scale Data Analysis?
- This scale is widely used in measuring the responses to the questionnaire in a research survey or aptitude test. It is also known as rating scale by default though other rating scales are there. This scale was invented by psychologist Rensis Likert hence it was named after him. This scale helps to investigate the underlying phenomenon of the surveyor test using the responses obtained and bring out the results. Another dimension in this scale is the design or the format which is used to capture the response.
- The survey process involves the design of line items or individual questions, the list of responses options for each line item and the participant will have to select one of the responses. Participants in the survey, while responding to a Likert item or question indicate their level of satisfaction or dissatisfaction / Approval or disapproval/agreement or disagreement on a range of responses.
- The list of responses will mostly be a range with symmetric between negative and positive extremes and many intermittent values in both positive and negative sides including a neutral value. The response set will have 5 or 7 components. It will have equal no of positive and negative responses (Symmetrical).
- Each Likert item will have its own set of responses designed exclusively for the item and the ranges in the responses be it positive side or negative side will be equal distanced (Balanced) and highly correlated. The responses are paired to each line item and they are discrete and non-numeric. The analogous range will lose its meaning when the responses are consolidated across questions.
Likert scale is used extensively in Business, Social Sciences, Statistics, Psychology and Marketing applications.
How does it work?
Each Likert item is a simple statement and the participant has to evaluate the statement and fill up response against the subjective or objective options which expresses his level of agreement of disagreement for the Likert item. The objective of the question may be to elicit responses under various categories. The possible responses under various categories are listed below
| Category | Possible Responses | ||||
| 1 | 2 | 3 | 4 | 5 | |
| Agreement | Strongly Against | Against | Neither | Agree | Strongly agree |
| Quality | Very Poor | Poor | Fair | Good | Excellent |
| Approval | Strongly Reject | Reject | Neither | Approve | Strongly Approve |
| Performance | Far Below par | Below par | On par | Above | Far Above |
| Quality | Far Below level | Below level | Acceptable | Above | Far Above |
| Frequency | Never | Rarely | Some times | Frequently | Always |
| Importance | Not at all | Somewhat | Important | More | Extremely |
| Satisfaction | Not at all | Somewhat | Satisfied | More | Extremely |
| Intensity | None | Very Mild | Mild | High | Severe |
| Awareness | Not at all | Somewhat | Aware | More | Extreme |
| Familiarity | Not at all | Somewhat | Familiar | More | Extreme |
Likert scaling mostly is bi-polar and it has negative, positive and neutral values. In some cases even values scale is used without neutral value and it is called as forced choice method.
This Likert scaling has few distortions that might produce bias results. Participants may
- Follow Central tendency bias and skip extreme values due to fear of being branded as extremist and also being apprehensive that next question will have more
- Agree always with the statement due to the institution mind
- May always disagree by being defensive due to fear of going
- May give incorrect answers with apprehensions of being evaluated as
- Provides wrong answers with the intention of being evaluated weak
- Chooses incorrect answers due to social compulsion to support a cause
How to Data Analysis?
- Once the survey questionnaire is done, responses to each item may be analyzed separately or responses are analyzed for a group of items. The Likert scale provides a common unit of measure to combine a group of items and it is called as summative scales. Responses in the Likert scale are not numeric and they should be Symmetric and balanced so multiple questions responses can be combined on a common scale.
- Parametric tests can be carried out with the data collected and analysis of variance test in particular and if it follows normal distribution cycle, 2 sample T-test can be carried out. If the responses do not fit normal and continuous distributions nonparametric tests like Mann-Whitney test can be conducted.
- The responses are summarized using median or mode but not on mean since these data are ordinal. Observations are shown in bar chart not as histogram since the data is not continuous.
Importance
- Participants in the survey don’t have the option of answering yes or no and not numeric in most of the cases. Degrees of perceptions or experience or opinion or feeling about an entity is taken as a response in the survey. Hence the quantitative data resulting out of the Likert scale can be analyzed without much of difficulty.
- Due to its uniform distribution of negative and positive responses with a neutral value and equal distance in its responses in both positive and negative side gives Likert a great advantage of grouping the responses of group of items can be analyzed together.
- Anonymity and self-administered questionnaire removes the social bias, faking good and faking bad responses and a free, unbiased responses can be ensured. The results in a survey is biased if the name and identity of the participant is insisted on.
Conclusion
Likert scales facilitates unbiased data collections by designing a compelling questions and thoughtful responses. These data can be used further to get insights and informed decision.
Recommended Articles
This is a guide to Likert Scale Data Analysis. Here we discuss what is Likert Scale Data Analysis, how to analysis data, with importance. You can also go through our other related articles to learn more –
