Proxy Variables: A Review of Their Use in Research
Proxy variables are variables that researchers use to represent a concept or phenomenon when direct measures are not available. This review article examines the use of proxy variables in the research literature, provides an overview of the types of proxy variables, and discusses the advantages and disadvantages of their use. In addition, the article provides examples of proxy variables across different disciplines and offers implications for researchers considering their use.
Proxy variables are used in a variety of disciplines, including education, economics, psychology, sociology, and public health. In education, proxy variables are used to measure educational attainment, such as the number of years of schooling or the highest educational degree obtained. In economics, proxies are used to measure income, such as the number of years of work experience or occupation. In psychology, proxy variables are used to measure psychological traits, such as the number of friends a person has or the number of books they read. In sociology, proxies are used to measure social status, such as the number of memberships to clubs or associations a person has. In public health, proxies are used to measure health status, such as the number of health-related behaviors a person engages in or the number of chronic diseases they have.
The advantages of using proxy variables are that they can provide reliable and valid measures of a concept or phenomenon when direct measures are not available. For example, when direct measures of income are not available, researchers can use proxy variables such as work experience or occupation to measure income. In addition, proxy variables can be used to measure concepts or phenomena that are difficult to measure directly, such as psychological traits or social status.
The disadvantages of using proxy variables are that they may not accurately measure the concept or phenomenon they are intended to measure. For example, work experience may not accurately measure income, as some individuals may have long work experience but relatively low incomes. In addition, the use of proxy variables may lead to measurement bias, as certain groups may be more or less likely to report certain proxy variables. For example, individuals with lower educational attainment may be less likely to report their educational degree.
This review of proxy variables provides implications for researchers considering their use. First, researchers should be aware of the advantages and disadvantages of using proxy variables, and consider whether the use of a proxy variable is appropriate for their research question. Second, researchers should consider potential sources of measurement bias when selecting a proxy variable. Third, researchers should consider the reliability and validity of the proxy variable they select. Finally, researchers should consider the implications of the results of their study when using proxy variables.
In conclusion, proxy variables are a useful tool for researchers when direct measures are not available. However, researchers should be aware of the advantages and disadvantages of using proxy variables, as well as potential sources of measurement bias, before selecting a proxy variable for their study.
References
Holland, P. W. (1986). Statistics and causal inference. Journal of the American Statistical Association, 81(396), 945-960.
Kline, R. B. (2015). Principles and practice of structural equation modeling. Guilford Press.
Krantz, D. H., Luce, R. D., Suppes, P., & Tversky, A. (1971). Foundations of measurement. Academic Press.
Robins, J. M., & Greenland, S. (1992). Identifiability and exchangeability for direct and indirect effects. Epidemiology, 3(3), 143-155.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.