Variables have been a major issue in the development of several researchers in the current system of social and scientific studies. In most scientific researches conducted, the researchers use observation and measurement techniques. In most quantitative studies, they use observable systems. This could be in a direct or indirect manner. Allen and Yen (1979) observed that it enables the researchers to assign and evaluate precise scores of numerical value to the insertions that are made while conducting a study.
On the other hand, researchers use measurements when conducting a quantitative study. These qualitative researchers can systemically observe this social context of the world, but they do not measure the social phenomena in a numerical approach. Therefore, observation becomes the most suitable approach which may be written in records. Observations may also be recorded as verbal texts, by collecting and analyzing texts which are obtained from respondents which form the population under study. The main point to note is that these researchers practically do not make descriptions of their measurement variables; hence, they must be able to describe the measurements they are conducting.
In determination of how variables in this study hypothesis transform, and whether the changes in a single variable influences the other variables, then it is vital for the measurement of the variable in question. Therefore, measurement refers to a process or simply a method of quantifying observations of empirical studies according to appropriate and precise methods and procedures. For example, sociologists may be interested in measuring the features and characteristics of groups or individuals. Other variables which can be measured are attitudes, perceptions as well as opinions of an individual or group of people. In addition, the behavior of people or groups can also be quantified in empirical terms.
It is of importance to note that variables are different; hence, the same approach cannot be used to measure all variables. It is due to variables are categorical and their quality vary as well as type. On the other hand, the variables also change in their intensity or amount, whereby one category may differ from other categories in one way or another. For instance, numerical variables differ with one another in terms of amount or quantity. The levels of measurements also vary with each other. For instance, nominal level is a categorical variable which varies in type or quality, however, they lack rank or order of their attributes. Examples of nominal level variables include religion, race, and gender. The ordinal level is the categorical variables which have attributes to be ranked or ordered. The ranking makes certain attributes to represent more or less of quality than others. Examples of ordinal variables are social status or class, educational level as well as the Likert-Scale type variables.
The interval of variables may not only be in rank, but the difference cannot be quantified. For example, IQ scores. Furthermore, other numerical variables may be considered as the ratio variables since they possess the true zero that represents the total quality absence to be measured. It means that a meaningful fraction (or ratio) can be constructed with a ratio variable. For example, weight is a ratio variable.
Understanding the levels of measurements assists the researcher to make a wise decision on how to interpret data. In addition, it enables researchers to decide what statistical analysis is appropriate to the assigned values.
According to Allen and Yen (1979), validity can be improved by making sure that the goals and objectives of the study are clearly defined and match the assessment measures designed.
