Correlations versus Causation
In this chapter, we introduce the concept of correlation versus causation. We show how to estimate a simple multivariate regression model using as an example the relationship between body mass index and wages. We outline how models estimated by ordinary least squares will be unbiased if certain conditions are met that are unlikely to be the case in most real-world scenarios. Next, we outline the different types of bias that you are likely to encounter when estimating. Finally, we introduce an instrumental approach as one method for estimating causal relationships.
KeywordsCorrelation Causation Endogeneity Omitted variable bias Reverse causality Instrumental variable
References and Further Reading
- Bernal, J. L., Cummins, S., & Gasparrini, A. (2017). Interrupted time series regression for the evaluation of public health interventions: A tutorial. International Journal of Epidemiology, 46(1), 348–355.Google Scholar
- Sargan, J. D. (1958). The estimation of economic relationships using instrumental variables. Econometrica, 26(3): 393–415. JSTOR 1907619. https://doi.org/10.2307/1907619
- Stock, J. H., & Yogo, M. (2002). Testing for weak instruments in linear IV regression.Google Scholar
- University of Essex. Institute for Social and Economic Research, NatCen Social Research, Kantar Public. (2016). Understanding society: Waves 1-6, 2009–2015 [data collection], 8th ed. UK Data Service. SN: 6614, https://doi.org/10.5255/UKDA-SN-6614-9