252 



ANDERSON 



to think about them analytically. Greater precision is no great benefit unless 

 it leads to a sharper analysis. 



Let us therefore examine pattern data on as simple a level as possible. In 



/ 



' ■ 



Length of B 



Length of B 



Fig. 1-4.- — Fig. 1. Pattern data reduced to their mathematically simplest form. 

 The two hnes differ in (1) length, (2) position on the grid, (3) slope. — Fig. 2. A 

 simple example of a machine which could produce pattern data such as those in 

 fig. 1. The large square represents sheets of plastic which are being cut by a right- 

 angled cutting bar, to produce small squares of plastic in which A equals B. Further 

 explanation in the text. — Fig. 3. A plotted against B when the machine in fig. 2 is 

 in proper adjustment. — Fig. 4. A plotted against B for five samples each from two 

 days' run when the machine was not being properly set up by the supervisor. 

 Further discussion in the text. Note that straight lines drawn through the two series 

 of dots would produce a figure essentially like the theoretical one of fig. 1. 



mathematical terms the simplest patterns I can imagine are the two straight 

 lines in fig. 1. They differ in their length, ths angle of thsir slops, and their 

 position on the descartian grid. To make the problem easier to think about, 



