# Variable Definition In Statistics: Read This!

Variable definition in statistics in one of the beginning to produce your own number. It becomes an integral part of the research. Sometimes, determining the variables becomes quite complicated for several people.

This is because they themselves do not understand the definition of the variable itself.

In life, we often hear opinions and statements like the best song in the world is rending, the people satisfaction rate of this program is very high, watching television is just a waste of time, etc.

Of course, we will ask, are these statements valid? Can these opinions be trusted?

To measure these things, surely we need a clear measurement tool. Concepts are not enough to formulate and summarize a problem or a phenomenon.

Using variables help us to define something so that it produces valid information. It is very essentials before using inferential statistics.

## Variable definition in statistics

Based on several experts, the following is the understanding of variables:

1. Variables are things that have different values ​​where the value (Kerlinger)

2. Variables are rational units that can be defined in the analysis that can be assumed as a set of measures of value (Black and Campaign)

Variable is a picture, opinion, a concept that is measured with a certain scale whose value can change. A thing or concept that can be measured can be called a variable.

Basically, a variable can always be measured or made its size through various measurement scales.

Tent the next question, can all things in this world be measured? Is it possible to measure everything whether it is things like feelings, judgments, and other aspects of personality?

Many researchers have different opinions on this matter. However, I myself believe that if you want to get valid knowledge and information, then you must try to find the right size and value for each variable.

This measurement is the main difference that distinguishes between concepts and variables.

The concept is a picture or perception that can be different from one person to another. Concepts are often difficult to define or measure and are subjective.

Meanwhile, a variable is something that can be measured even with a variety of sizes and produces accuracy that is relative.

Therefore, it is important to convert a concept into a variable so that it can be measured using a more precise scale and unit of measurement.

Examples of concepts are effectiveness, happiness, wealth, etc. My question is, how do way more effective in measuring these things?

Then, we can change the concept into variables. For example, effectiveness is transformed into an index of effectiveness, happiness is changed to an index of happiness, wealth is changed to an income level.

Thus, you can get a size that is more valid and easily interpreted.

Consider the following picture so you can easily understand it.

## Variable types

In general, there are 3 types of variable classification:

1. Based on cause and effect relationships

2. Based on the study design

3. Based on the measurement scale

### 1. Based on cause and effect relationships

Based on a causal relationship, there are 4 types of variables that we can define:

1. Independent variables are variables that cause changes in an event or event.

2. The dependent variable is the result or value that arises from changes in the dependent variable.

3. Extraneous variables are variables that may have an influence on the dependent variable. This variable is not measured in research but may have an influence on the dependent and independent variables.

4. Intervening variables are variables that connect the dependent and independent variables.

In order to more easily understand the cause and effect relationship variable, please see the following diagram.

Suppose, you will examine poverty or criminality. You can use the following model in formulating the variables that you use.

### 2. Based on design studies

Based on the study design, there are 2 types of variables:

1. Active variables are variables that can be changed, controlled, and controlled

2. Attribute variables are variables that cannot be changed, controlled, or controlled and at the same time represent the characteristics of the population that we examine. For example, age, gender, education level, income, etc.

For example, you research the effectiveness of a promotional method for product customers. Suppose you use the promotion models A, B, and C.

In its implementation, this research can show varied results if the research is carried out by different people. However, researchers cannot control the characteristics of the population of product customers in terms of age, gender, the reason for buying, and others. This is what is called an attribute variable.

On the other hand, researchers have control in determining or changing what promotion methods should be used. This is what is called an active variable.

### 3. Based on the measurement scale

When you look at various measurement scales, you will immediately remember the types of data based on the scale. Well, this type of variable is actually not much different.

In qualitative research, we tend to get results and conclusions in the form of descriptive statistics. In contrast, quantitative research produces output in the form of a certain measurement scale (nominal, ordinal, interval, and ratio).

Based on the measurement scale, there are two types of variable categories:

1. If the data type is a category (on a nominal and ordinal scale) and continuous (on an interval and ratio scale)

2. If the type of data is qualitative (on a nominal and ordinal scale) or quantitative (on an interval and ratio scale).

Next, I will briefly discuss the scale of measurement.

1. Nominal scales allow grouping of samples or study populations based on characteristics. These samples or populations will be grouped into groups or classifications according to the same characteristics.

2. The ordinal scale is a scale that has a nominal scale characteristic, where samples or populations are grouped in the same characteristics, but ranking in a particular order.

3. The Interval scale is a scale that has all the characteristics of an ordinal scale but does not yet have an absolute zero value.

4. Ratio scale is a scale that has all the characteristics of a nominal, ordinal, and interval scale. Also, the ratio scale has an absolute zero. Every difference on this scale is always measured from zero.

Maybe, you will ask, aren’t categorical data and qualitative data the same thing? Continuous and quantitative data are the same type of data?

At first glance it does look similar, but here are the differences between the two.

Categorical variables are measured on a nominal and ordinal scale while continuous data are measured on an interval and ratio scale.

There are three types of category variables :

1. Constant variables: only have one category or group of values, for example, flowers, water, cars, etc.

2. Dichotomous variables: only have two types of value groups, such as male or female sex, rich or poor, etc.

3. Polytomous variables: having more than two groups of values, for example, the level of product sales (high, medium, low), the level of satisfaction (very satisfied, satisfied, neutral, dissatisfied, very dissatisfied), etc.

Continuous variables have measurements that are continuous or continuous, for example; age, income, expenses, etc. There are many values ​​and units that can be defined in this variable. Ages can be grouped in years or months. Revenues and expenses can also be grouped in units of million, thousand, or cents.

In many ways, qualitative variables do have characteristics that are very similar to categorical variables. However, there are slight differences in the two types of variables.

For example, we want to measure someone’s income level. We can classify these revenues into the lower, middle, top categories. Measuring the level of income in the form of money is calculated as a continuous variable while the grouping as a lower, middle and upper category is a qualitative variable.

Take a look at the picture below:

Nominal and ordinal scales are often used for qualitative research measurements while ratio and interval scales are often used in quantitative research.

Nominal and ordinal scales are often considered subjective and produce information that is very relative. This is due to the difficulty of taking measurements, especially in categorical or qualitative data research.

Ratio and interval scales are considered a more valid measurement scale and are able to provide accurate results because of the use of numerical indicators.

Summary

Last, there are 3 types of variable classification:

1. Based on cause and effect relationships

• 1. Independent variables
• 2. The dependent variablevariable.
• 3. Extraneous variables
• 4. Intervening variables

2. Based on the study design

• 1. Active variables
• 2. Attribute variables

3. Based on the measurement scale

• 1. Nominal scales
• 2. The ordinal scale
• 3. The Interval scale
• 4. Ratio scale

Basically, variables definition in statistics change the various concepts that exist in this life into something that can be measured properly and is able to describe something.

After you get a variable, don’t forget to determine the types of data analysis of the variable. Because this will determine the research method you will use.