1921.] 



Agricultural Statistics. 



641 



to which I have referred above arises. It will hardly be denied 

 that there would be a constant bias in the returns from growers. 

 Farmers as a class have an inherited tendency — bred of long 

 generations of tenancy— to under-state the results of their 

 farming. This unconscious mental bias, for which there is no 

 reason under modern conditions, may be dying out, but there 

 still remains the possibility of conscious bias. It is clearly in 

 the inte]-est of producers that any published returns of the year's 

 harvest should not over-estimate the crops. The shorter the yield 

 the better the price, may not always hold good in relation to 

 products which are subject to world competition, but nevertheless 

 it is an economic axiom which no producer is likely to overlook. 

 If he is asked to make a return of quantities which he does not 

 himself precisely know at the time of the return, it is ob^'ious 

 that he will take care to err on the side of caution. The cumula- 

 tive effect of these tendencies upon over 400.000 I'eturns would 

 be substantial and the whole results would be invalidated. 



Statistics of crops based on the quantitative results of sample 

 plots are in theory not open to similar objection, but in practice 

 it would be extremely difficult to arrange a system which would 

 give accurate arithmetical results. Not only would the plots 

 have to be so distiibuted as to represent all conditions of soil 

 and situation for each crop, but the number of plots representing 

 each variety of soil and situation would have to be proportionate 

 to the whole area on which similar conditions prevailed. The 

 yield of wheat given on heavy clay plots, for example, would 

 have to be weighted by the total area of heavy clay land on 

 which wheat was grown that year — and so on. Differences in 

 manui'ing could to some extent be allowed for. but it would be 

 difficult to make proper allowance for differences in cultivation. 

 The sample plots would naturally tend to be above the average 

 in cultivation, and especially in cleanliness. 



From the collecting officers are obtained estimates of the 

 natural weight of grain and of the yield of straw. They have 

 also to furnish monthly reports on the appearance of the crops, 

 and the general agTicultural conditions, and to make forecasts 

 of the probable yield of crops during the later periods of their 

 growth . 



Parliament has always been indifferent in regard to the 

 collection of statistics. If this country does possess a system 

 of economic statistics of which on the whole it has no reason 

 to be ashamed, the fact is attributable to the persistent efforts 

 of a few individuals, such as Sir Robert Giffen, Sir Alfred 



E 



