Question

Anyone who has studied statistics or research has heard the saying "Correlation does not imply causation." What factors must an analyst consider to decide whether the correlation is meaningful enough to investigate further?

Answer 1

An analyst must consider possible Confounding Variables to decide whether the correlation is meaningful enough to investigate further. Correlation between two variables simply indicates a relationship exists between the two variables. Causation is more specific: Causation implies that one variable actually causes the other. So, in analyzing the correlation studies, the researcher must consider possible Confounding Variables, which indirectly controls both dependent variable as well as dependent variable, thus creating spurious correlation between dependent and independent variables, which in reality is not there. For example, the high positive correlation between death by drowning in water and increase in sale of ice-cream is not a reality. The spurious correlation between these two variables is due to the Confounding Variable: hot climate. In hot climate people go for water sports and they get drowned. In hot summer, sale of ice-cream is high.