Variables and Data

Variable comes from the words " vary " and " able " which means " change " and " can". So, literally means variable is mutable , so each variable can be assigned a value and that value is arbitrary. The value can be quantitative (measured and/or calculated, can be expressed with numbers) can also be qualitative (the number and degree of attributes expressed by quality values). Variables are important elements in research problems. In statistics, variables are defined as concepts, qualities, characteristics, attributes, or properties of an object (people, objects, places, etc.) whose values vary from one object to another and have been determined by researchers to be studied and conclusions drawn . Characteristics are certain characteristics of the object we are examining, which can distinguish the object from other objects, while the object whose characteristics we are observing is called the unit of observation and a certain number or category (quality value) of an object that we are observing is called variate (value). The collection of values obtained from the measurement or calculation of a variable is called data .

The characteristics of an observation are different (variable) or have symptoms that vary from one observation unit to another, or, for the same unit of observation, the characteristics change according to time or place. If the characteristics of each unit of observation are all the same, do not vary, then it is no longer a variable, but a constant .

Example: If you are studying a group of children, the children there are just a concept, not a variable. If you are interested in measuring his height, weight, age, determining gender, and so on, then you have already talked about variables, because their values can vary from child to child. For research purposes, a concept can be converted into one or more variables. For example, regarding the concept of children, among the characteristics that can be measured, you are more interested in weighing, then:

• Concept: is a property/characteristic of Children
• Characteristics : the characteristic you are observing is the weight of the child .
• Variable : because the weight of each child can vary, then the weight is a variable.
• Unit of observation: the unit of observation is each Child (each individual), and
• Value (variate/data) : the measured weight of each child is called data/variate (value).

As another case in point, if you are studying a group of tomato plants (concept), the following variables might come into your consideration: height, width, number of leaves and number of fruits, and weight of tomatoes. Examples of other variables are eye color, IQ, education level, social status, teaching method, type of fertilizer, type of variety, type of medicine, all of which are variables because they have different characteristics. The characteristics of a variable must vary or change. On the other hand, if the characteristics are all the same, then the unit of observation is no longer a variable, but a constant . A constant is a certain number whose value is always constant under all conditions, such as the speed of light, the force of gravity, etc.. However, a variable can become a constant if its values are made the same. For example, gender is a variable, but if the unit of observation that we observe is only limited to female sex, then gender turns into a constant, because the value is the same in all conditions.

Operational definition

Operational definition is an aspect of research that provides information or instructions to us on how to measure a variable. The scientific information described in the operational definition is very helpful for other researchers who want to conduct research using the same variables, because based on that information, they will know how to measure variables that are built on the same concept. In this way, he can determine whether to keep using the same measurement procedure or whether a new measurement is needed. Concepts that have been translated into units that we consider more operational (variables and constructs), are usually not fully ready to be measured. Because these variables and constructs have alternative dimensions that can be measured in different ways. An example of the variable age/age. The way to measure these variables can be different, first you might measure their age directly numerically, for example 4, 12.5, 18, 31 years and so on, or you could measure based on categories, for example Toddlers (0-5 years old), Children (5 - 14), Teenagers (14 – 24), Adults (25 – 54), Old (55-64), and Elderly (> 65) years.

Type of Variable

Variables can be divided based on: role, method of measurement, and whether or not it can be measured directly.

Based on its function/role in research

In quantitative research, operationally defined variables are usually divided into independent variables ( independent : active or attributes), dependent variables ( dependent ), and extraneous /extra/additional variables ( extraneous ).) which is not the subject of the research being studied and is outside the main observation/study of the research. An understanding of extraneous variables is very important, because these variables can compete with independent variables and can be confusing/confusing in explaining the pattern of relationships between independent variables and dependent variables. Therefore, in determining the causal relationship, we should identify the presence or absence of extraneous variables that are proven to affect the dependent variable. If there is, then the extraneous variable is called a confounding variable . Confounding Variables should be controlled or included in the model. Otherwise, we will not be sure that the change in the dependent variable is only caused by the independent variable. To understand the variables in the study, consider the following case examples: If we want to see the effect of different fertilizer doses on plant growth, then:

Independent Variable (IV).

The independent variable is the variable that causes or affects the dependent variable (DV) or causes variation for the dependent variable (DV). If the IV variable changes, the DV variable will also change. The independent variable is a variable whose factors are measured, manipulated, or selected by the researcher to determine its relationship with an observed symptom. If translated into Indonesian, the independent variable is also referred to as the independent variable and is often also referred to as the independent variable, stimulus, factor, treatment, predictor, input, or antecedent.

For example: The effect of teaching methods on student achievement. => The independent variable is the Teaching Method. Effect of Organic Fertilizer on the yield of tomato plants. =>The independent variable is Organic Fertilizer. Teaching methods and organic fertilizers can be manipulated or determined by researchers. Not all independent variables can be manipulated, for example, attributes that are already attached to an object. For example: Gender, Age, Slope, altitude, etc.

Dependent Variable (DV).

The dependent variable is the variable that is influenced or which is the result of the independent variable . Dependent variable , in Indonesian is often referred to as dependent variable, dependent variable, dependent, response, output variable, criteria, or consequent. This variable is the main focus of the research. It is this variable whose value is observed and measured to determine the effect of the independent variable. The value can vary and depends on the magnitude of the change in the independent variable. That is, every time there is a change (addition/subtraction) of several units of the independent variable, it is hoped that it will cause the dependent variable to change (increase/decrease) a number of units as well. Mathematically, the relationship may be described in the form of the equation Y = a + bX. For example, Y = Yield (tons) and X = Urea fertilizer (kg), then each time urea fertilizer is increased/or decreased by b (kg), the yield increases/decreases by b (tons) and if no fertilizer is applied (b = 0), then the result is a (tons). The pattern of the relationship between the two variables is usually examined in association or prediction research, usually tested using Regression Analysis. In contrast to the example of the effect of teaching methods on student success, the measurement scale of the independent variable is not an interval or ratio variable, so to see the effect of the independent variable on the dependent variable, it is more appropriate to use Analysis of Variance (ANOVA). With the Anova, we can determine whether there are differences between teaching methods, and if there are, we can determine which teaching method is better or best. so to see the effect of the independent variable on the dependent variable, it is more appropriate to use Analysis of Variance (ANOVA). With the Anova, we can determine whether there are differences between teaching methods, and if there are, we can determine which teaching method is better or best. so to see the effect of the independent variable on the dependent variable, it is more appropriate to use Analysis of Variance (ANOVA). With the Anova, we can determine whether there are differences between teaching methods, and if there are, we can determine which teaching method is better or best.

Moderator Variables

The moderator variable is a special variable from the independent variable. In relationship analysis that uses at least two variables, namely one dependent variable and one or several independent variables , sometimes the relationship between the two variables is influenced by a third variable, namely other factors that are not included in the statistical model that we use. The variable is called the moderator variable. moderator variable this is another variable that can strengthen or weaken the relationship between the independent variable (independent) and the dependent variable (non-independent). In the Analysis of Variance (Anova), the effect of this moderator variable can be represented as the effect of the interaction between the main independent variable (factor) and the moderator variable (Baron and Kenny, 1986: p. 1174). This variable can be measured, manipulated, or selected by the researcher to find out whether its existence will affect the relationship between the independent variable and the dependent variable. Schematically, the relationship between the three variables can be illustrated as in the following figure:

Case example 1: Consider a study to see the difference between two statistical teaching methods, eg Method A and Method B. If male students do better with Method A, while female students do better with Method B, then gender is a moderating variable. .

Case Example 2: For example, the effect of inorganic fertilizers on rice yields. The results of the analysis show that there is no effect of the use of inorganic fertilizers on rice yields, even though theoretically there should be a difference. Why is that?? After being investigated, it turns out that there are other variables (eg varieties) that are not included in the model or are not controlled (uniformed), thus affecting the diversity of rice yields. These variables are moderator variables, which should also be included in the model. This is, for example, indicated by the difference in responses between rice varieties. Superior varieties are more responsive to inorganic fertilizers, while local varieties are not very responsive and even tend to decrease yields.

Case example 3: The Effect of Training on Job Performance. For example, training followed by administrative staff of a university in the hope of improving skills in completing administrative tasks. All employees involved have the same level of education, D3. After the training was completed, a skill test was conducted. After observing, it turns out that the ability of employees from D3 Management has better skills than employees from D3 Agriculture. It is clear here that the difference is due to differences in the ability to absorb the material presented when carrying out the training. D3 management employees are more enthusiastic in participating in the training compared to D3 Agriculture because they are relatively easy to understand the material (according to their field). In this example, training is the independent variable, job performance is the dependent variable, and educational background is the moderator variable. From the three case examples, it can be concluded that the moderator variable has a significant effect (has a significant contribution) on the ability of the independent variable to influence the dependent variable.

Intervening/mediator variables.

The independent and moderator variables are concrete variables . These variables can be manipulated by researchers and their effects can be seen or observed. It is different with the intervening variable, the variable is hypothetical, meaning that the effect is not visible, but theoretically can affect the relationship between the independent and dependent variables being studied. Research involving intervening variables (mediator/mediating/mediation/disruptor) is very common in the fields of sociology and psychology, as well as the behavioral sciences and other non-experimental research. For researchers in the exact field (especially in experimental research), there may not be too many people who know or involve this variable, because it is abstract and cannot be measured (mysterious, don't take it seriously.. :-) ). Just look at the following statement by Tuckman (1988): " … an intervening variable is that factor that theoretically affects the observed phenomenon but cannot be seen, measure, or manipulate…” . Many students, myself, and even some researchers still have difficulty in distinguishing between the moderator variable and the interfering variable on this one, intervening (mediator) means :-) . Intervening variables defined as a variable that theoretically affects the relationship between the independent variable and the dependent variable, but cannot be seen, measured, and manipulated; the effect must be inferred from the effects of independent variables and/or moderate variables on the symptoms being studied (Tuckman, 1988). This variable is an intermediate variable (interrupt) which is located between the independent variable and the dependent variable. This variable can be used to explain the process of the relationship between the independent variable and the dependent variable, for example X → T → Y, where T is the intervening variable used to explain the pattern of the relationship between IV and DV.. The last terminology, which is as an intermediate variable, is consistent with the methodology and definition in Structural Equation Modeling (SEM) Analysis. For example, X is age and Y is reading ability, a causal relationship between X and Y can be explained by the Intervening variable T, for example Education. Thus, age (X) does not directly affect reading ability (Y), but first through the intervening variable, education (T), or in other words, X affects T and then T affects Y.

Example: Education level → type of work → income level Teaching method → learning motivation → Student achievement New technology → culture → Community response Age → Riding experience → motorcycle riding skill (Valentinno Rossi, for example, :-))

Example in agriculture: Effect of inorganic fertilizer application on rice yield. For example, varieties have been included in the model or the varieties are made the same (seeded varieties), but the results are still not significant. Why?? After careful research, it turns out that the rice plants that are given fertilizer, for example, become susceptible to disease/pest attacks so that most of the land is affected by pests/diseases, as a result rice yields do not increase. Intervening Variable is Disease/Pest Attack.

Relationship of these variables:

Notes: There is some literature that says there are other variables besides the variables mentioned above, namely the Control Variable . The control variable is the variable that is controlled or made constant so that the influence of the independent variable on the dependent variable is not confounded by the influence of other factors that we have not observed. In other words, other variables that can affect the relationship between the independent variable and the dependent variable, try to be eliminated or neutralized or controlled or uniformed! Thus, it is expected that the variable that gives diversity to the dependent variable is only the independent variable whose influence is studied, which is known as treatment!

Based on the measurement method

• Quantitative (discrete/continuous)
  • Ratio
  • Interval
• Qualitative
  • Ordinal > there is order
  • Nominal > no order

Based on whether or not it can be measured directly

• Observed variable (observed variable)
  • Can be directly observed/measured
  • Example: age, gender, weight
• leave variable (latent variable)
  • Cannot be directly observed/measured
  • Example: service quality, customer satisfaction, health
  • Generally measured using indicators in the form of observed variables, usually more than two indicator variables.

Refs/Links:

Statistical Mediation Mediation Moderator

GORDON MARSHALL. "intervening variable." A Dictionary of Sociology. 1998.

Encyclopedia.com. 23 Feb. 2010 <http://www.encyclopedia.com>.

Tuckman, B. W. (1988). Conducting educational research . (3rd ed.) New York: Harcourt Brace Jovanovich.