Correlation does not Imply Causation when Calculating Trend Lines

Analysis

So, just because you got a high R-squared on a trend does not mean you can do predictions with your trend line. It just means that the data trends in the same direction. You can get yourself, and your business, in trouble if you use a trend line for forecast and there is no causation between the sets of data. For example, there is a high correlation between ice cream sales and the number of lifeguard assists at the beach.

Prevent Drownings

Stop the Sale of Ice Cream

With that logic we should stop the sales of ice cream at the beach. Obviously there is a high correlation between ice cream sale and lifeguard assists, but that is because both occur during the heat of summer when ice cream sale is high and more people get in the water.

So what do you do? One technique is to use some of the Post-It note techniques I’ll be blogging about to create causal diagrams. Causal diagrams can help you understand cause and effect relationships.

Learn how to check the R-squared or “goodness of fit” for a trend.

Learn how to create a trendline in a chart or TREND data in a worksheet.

Learn how to extend a TREND line to create a forecast.

Learn how to show or hide a TREND line under the check box control.

 

Share the power...