What executives get wrong about statistics: Moving from statistical significance to effect sizes and practical impact

Statistical significance functions as an arbiter of sorts for data analysis purporting to show a relationship between two or more variables. Unfortunately, in far too many situations, statistical significance may lead decision-makers relying on data and analytics to improve business decisions astray, particularly in the context of big data. In this article, I outline reasons why executives should develop a healthy discernment when they see the phrase "statistically significant" in media outlets, internal analyses, consulting reports, and other sources. To overcome the limitations of focusing on statistical significance, I propose executives shift their attention toward the effect size reported from a statistical model. While not without limitation, effect sizes are more useful to decision-makers, highlight the practical implication of analyses, and help in quantifying the uncertainty inherent to working with data.