In statistics, research variables can be measured in parametric and non-parametric variables. Variables measured by interval scale and ratio scale are grouped in parametric variables.

Meanwhile, the variables measured by the nominal and ordinal scales are grouped in non-parametric variables. On this occasion, Kanda Data will discuss variables included in the non-parametric variables.

Based on the scale of the data, we can easily distinguish between parametric and non-parametric variables. In non-parametric group variables measured by an ordinal scale, many researchers use a Likert scale.

The Likert scale is a scale created by researchers to measure non-parametric variables on an ordinal scale. Examples of variables measured by an ordinal scale are variables of attitude, competence, behavior, motivation, etc.

For those who want to get a more in-depth explanation of the difference between the ordinal scale and other data measurement scales, you can read the previously written article entitled “Nominal, Ordinal, Interval, And Ratio Scales | Types Of Data Measurement.”

The Likert scale initiated by Rensis Likert was created on a scale of 1-5. Researchers have widely used the Likert scale to measure non-parametric variables on an ordinal scale.

Suppose you observe a consumer behavior variable to make it easier to understand. You cannot directly get the quantitative data on the consumer behavior variable.

Next, to analyze the consumer behavior variable, you need to use an approach that represents the consumer behavior variable. The Likert scale can be used to measure consumer behavior variables.

**How to Create Variables using a Likert Scale**

To create non-parametric variables using a Likert scale, the first step you need to do is create a statement or question item. Many researchers ask, what is the limit on the number of questions on the Likert scale?

The number of statement items or questions is adjusted to the items representing the examined variables. For example, if you create a consumer behavior variable, you can develop the following items:

(a) purchasing behavior based on consumer habits

(b) purchasing behavior based on product variations

(c) consumer behavior based on product quality

(d) consumer behavior based on affordable product prices

(e) etc

You can arrange items according to theory and previous research that represent consumer behavior variables. For example, we have determined that 15 question items represent consumer behavior variables.

The next step is to arrange 15 question items that respondents easily understand. Next, you ask the respondent to answer the question/statement using a Likert scale.

The Likert scale can be created with positive or negative statements. For the Likert scale with positive statements, the answer scores can be created as follows:

Score 5 = Strongly Agree

Score 4 = Agree

Score 3 = Uncertain

Score 2 = Disagree

Score 1 = Strongly Disagree

For the Likert scale with negative statements, the answer scores can be created as follows:

Score 1 = Strongly Agree

Score 2 = Agree

Score 3 = Uncertain

Score 4 = Disagree

Score 5 = Strongly Disagree

**How to Input Variable Data Using a Likert Scale**

The next step is data entry if you have collected data from respondents. Data is inputted per item with a score according to the answers given by the respondent.

For example, we have created 15 positive statement items, so the score entry is adjusted to a score of 5 for strongly agree answers and 1 for strongly disagree answers.

You need to know one thing: before you collect data from respondents through questionnaires or direct interviews, you need to test the validity and reliability.

Validity and reliability tests aim to ensure that the research instrument/questionnaire is valid and reliable.

**How to Analyze Variables Using a Likert Scale**

There are at least two things that can be elaborated on to analyze variables measured by a Likert scale. You can use descriptive statistical analysis and inferential statistical analysis.

For descriptive statistical analysis, you can obtain values for Max, Min, Mode, etc. As for inferential statistical analysis, you need to develop hypotheses to be tested.

Hypothesis testing on variables using a Likert scale can include correlation tests, comparison tests, and other tests following statistical rules.

Suppose we analyze the correlation of consumer behavior with the decision to buy a product, then the data used is the total score for each variable.

The consumer behavior variable consists of 15 question items so that if the respondent answers all the question items, a minimum score and a maximum score will be obtained.

The minimum score is obtained from the multiplication of 15 question items with the lowest score (1). In contrast, the maximum score is obtained from the multiplication of 15 question items with the highest score (5). Therefore, the minimum score is 15, and the maximum score is 75.

Next, you also need to create a purchasing decision variable with the same steps as a consumer behavior variable. In the last stage, you can analyze using the Spearman rank correlation.

Based on the results of the Spearman rank correlation analysis, you can find out whether consumer behavior has a significant relationship with purchasing decisions or not. Then, we can determine the size of the correlation coefficient and the sign of the correlation.

Well, that’s the topic of discussion on this occasion. Hopefully useful for all of you. See you in the following article!