The derivation of above simplified chart began with Table 56 on page 110 of the Evaluation of Weather Based "Smart' Irrigation controllers.

The N column definition is the number of each controller. The "95% Conf Interval" is the range over which 95% of the data can be found. The "95% confidence interval" is a standard statistical tool to test for significance.

The definition of "statistical significance" is "the quality of being worthy of attention." Statistical significance is shown below where both lines have equal averages. The red line has data points spread equally so that they all equal probably. Its average does NOT tell us anything about the central tendency of the data. So the average of the straight line is NOT statistically significant.

The curve has data points clustered around its average. So its average tells us VERY much about central tendency of its data. Its average is statistically significant.

In Table 56 above, only 2 companies had compute water savings was statistically significant or "worthy of attention": Accurate WeatherSet and ET Water. Together we had enough controllers that saved enough water that** All Sites** average data was also "worthy of attention." So while this 309-study compute water savings (or increases) for all 10 controllers, the averages for 8 of the controllers are NOT "worthy of attention."

The more complicated Table 56 was first simplified and sorted by water savings as shown below.

This data was then plot as shown below and shown on the homepage.