Last edited on 27th April, 2004 by gs/ gs . - Host home page:http://www.statlab.uni-heidelberg.de

This example comes from a survey of sitona weevils on alfalfa stands of varying ages. The survey covered six farmers' properties, with varying numbers of stands per farm. The numbers of sitona weevil larvae per square metre are given for each stand in the data set. (Sitona weevil larval numbers per m2 with age of alfalfa stand for each farm. Data by courtesy of Dr Steve Goldson, MAFTech, Lincoln, New Zealand.) The survey objective was to confirm a previously observed phenomenon, that weevil populations decline as the alfalfa stand gets older. In the data set this trend occurs within five out of the six farms. However, a simple regression model is inadequate to model the data since there are differences in larval numbers between the farms, presumably due to differences in management or other characteristics. This suggests parallel lines are required, one per farm.

Source: D.J. Saville, G. R. Wood: *Statistical Methods: The Geometric Approach*. Ch. 17. New York:Springer 1991.