Introduction
In a statistical analysis of research data, different variables can be used. They include various categories which are important for a researcher. Thus, in healthcare research, such categories as age, gender, height, weight, etc. can be applied. As a rule, several types of variables are used in statistical analysis to provide background for the research questions. Some variables demand fixed answers, the others depend on the intentions of a researcher.
Description of Categorical Variables
All of the provided databases contain categorical variables. For this discussion, categories such as education and marital status can be used. The first category, which is education, may include variables such as school, college, or university education. Variables can be changed depending on the purpose of the investigation. Thus, if education information is very important, a number of variables can be increased and they can be more detailed. As for the category of marital status, it consists of such variables as married, single, divorced, or widowed. I believe that these variables were properly selected as categorical. I would probably make the same classification if I had to use them in research.
Analysis of Categorical Variables’ Application
The use of categorical variables in diverse studies are the issues of separate researchers and scholars are still not unanimous in their application. Some suggest that categorical variables are popular among social scientists, for example, consumer psychologists (Iacobucci, 2012). Although they are frequently used, the issue of suitable methods of “incorporating categorical variables” is probably one of the problems in mediation analysis (Iacobucci, 2012, 583). Particular treatment of categorical variables is suggested by Rhemtulla, Brosseau-Liard, & Savalei (2012). They claim that “Any categorical variable is nonnormal by virtue of being discrete rather than continuous. Categorical variables are likely to produce nonzero kurtosis estimates, depending on the frequency of the middle categories, and category asymmetry will further lead to nonzero skewness” (Rhemtula et al., 2012, 354). In case the medical research is addressed, descriptive statistics is often applied to characterize the patient sample. Terman, Shields, Hume, & Silbergleit agree that in their research of the influence of age and chronic medical conditions on neurological outcomes “continuous variables were described using medians and interquartile ranges” (2012, 170). At the same time, categorical variables were considered “frequencies and percentages within each group (Terman et al., 2012, 170).
In the databases under analysis, the variables of education and marital status were selected as categorical. This is a suitable choice for the research because the variables in these categories do not follow any logical order. Continuous variables do not fit the categories of marital status and education in this research since they do not contain any numbers which are necessary for continuous variables.
However, there are other possibilities for categorical variables in different studies. Thus, Nicolaides, Syngelaki, Ashoor, Birdir, & Touzet in the research of noninvasive prenatal testing for fetal trisomies presented the descriptive data “in median and interquartile range for continuous variables and in numbers and percentages for categorical variables” (2012, 374.e2).
Conclusion
On the whole, the choice of variables and their types should be carefully considered by a researcher. Previous literary reviews can be helpful here. Thus, concerning categorical variables, Garrido, Abad, & Pondosa suggest studying their estimators (2016). It should be kept in mi that the selected variables and their types predetermines the results of any research, its reliability and validity.
References
Garrido, L.E., Abad, F.J., & Pondosa, V. (2016). Are fit indices really fit to estimate the number of factors with categorical variables? Some cautionary findings via Monte Carlo Simulation. Psychological Methods, 21(1), 93-111. Web.
Iacobucci, D. (2012). Mediation analysis and categorical variables: The final frontier. Journal of Consumer Psychology, 22, 582-594. Web.
Nicolaides, K.H., Syngelaki, A., Ashoor, G., Birdir, C., & Touzet, G. (2012). Noninvasive prenatal testing for fetal trisomies in a routinely screened first-trimester population. American Journal of Obstetrics and Gynecology, 207(5), 374.e1-374.e6. =
Rhemtulla, M., Brosseau-Liard, P.E., & Savalei, V. (2012). When can categorical variables be treated as continuous? A comparison of robust continuous and categorical SEM estimation methods under suboptimal conditions. Psychological Methods, 17(3), 354-373. Web.
Terman, S.W., Shields, T.A., Hume, B., & Silbergleit, R. (2015). The influence of age and chronic medical conditions on neurological outcomes in out of hospital cardiac arrest. Resuscitation, 89, 169-176.