18 BASES FOR CURRICULUM MAKING 



size of farm, amount of capital, location, quality of soil, market- 

 ing facilities, etc. Therefore the final data used for this part of 



the study cover ratings of over one hun- 



Many separate dred fifty groups in several middle west- 

 groups were rated ern states. Many more groups than this 

 to overcome local number have been rated, but the data ful- 

 variations. filling the requirements already stated 



alone have been used in the final conclu- 

 sions. It may be interesting to note that, as a rule, each group 

 represents a particular community and that these communities 

 are widely scattered throughout a number of states. Separate 

 studies of the correlations for each group show wide variations 

 among the groups together with many significant and interest- ' 

 ing peculiarities. This would be expected by any one at all 

 familiar with farm life. 



Since communities and groups of farmers vary to such 

 great degrees, it is necessary to combine the results for all of 

 the groups in each relationship. This is done by finding the 

 median of all the correlations for each particular relationship. 

 The r's for each summation G, however, need not be computed. , 

 Since summation G's may appear in values from to 42, these 

 may be arranged as a scale and the frequencies of the various 

 values listed against this scale. The median of these frequencies 

 is easily found 10 by counting in to the mid-point and noting at 

 which summation G on the scale this appears. This summation 

 G will give the correlation desired. This correlation is the cen- 

 tral tendency of the relationships for all of the groups ranked. 

 In the present study one hundred and fifty or more summation 

 G's for each characteristic are listed from which such central 

 tendencies are taken. 



10. See Chapters III and IV in Thorndike's Mental & So- 

 cial Measurements. 



