Mediation in Excel - Sobel Test in Excel (Baron & Kenny 1986 Causal Steps Approach)
Data file: docs.google.com/spreadsheets/...
Written walkthrough: psychologyofbusiness.beehiiv....
*ChatGPT generated description.
This video shows mediation analysis using the Sobel test, which is a method for assessing whether there is a mediating variable between an independent variable (X) and a dependent variable (Y). In this analysis, Well-Being (X) is the independent variable, Job Performance (Y) is the dependent variable, and Job Engagement (M) is the mediating variable.
The article outlines the classic Baron and Kenny (1986) steps for mediation analysis, which involve assessing the relationships between X and Y, X and M, M and Y (while including X as a predictor), and determining whether full or partial mediation has occurred.
The Sobel test formula is provided as Sobel Statistic = a * b / SQRT(b² * Sa² + a² * Sb²), and it is used to test whether the indirect effect of X on Y through the mediator is significantly different from zero.
The article also introduces a calculator for performing the Sobel test in Excel, including the necessary regression analyses. The Sobel Statistic is calculated based on the coefficients and standard errors from the regression analyses.
The results of the analysis can indicate whether there is mediation (Job Engagement mediates the relationship between Well-Being and Job Performance) based on the p-value. Additionally, the article discusses the concept of full mediation (the X-Y relationship is no longer statistically significant) and partial mediation (the relationship is reduced).
The article provides references for further reading, including studies by MacKinnon et al. (2002) and Baron and Kenny (1986), as well as the work of Ruben Geert van den Berg.
In summary, the article explains how to conduct mediation analysis using the Sobel test, including the necessary calculations and steps, and discusses the significance of the results in terms of mediation and full or partial mediation.