1921.] 



Agricultural Statistics. 



639 



probably better than in any other country. It has, indeed, 



formed the model on which many other countries have founded, 

 or revised, their methods. Starting as we did, before most 

 other countries, profiting by the lessons of long experience, 

 and with so small an agricultural area to deal with as com- 

 pared with most other countries, it would indeed be disappoint- 

 ing if we did not occupy a foremost place in this respect. 



The basis of our agricultural statistics is the return obtained 

 in June each year from every occupier of an agricultural hold- 

 ing of over one acre. This gives the acreage under each 

 crop and the numbers of live stock of different classes on the 

 holding at the date of the return. 



Prior to 1917, this return was voluntary, but it was made 

 compulsory by the Corn Production Act of 1917. There 

 appears to be an impression in the minds of some that this 

 alteration has affected the consecutiyeness of the annual 

 returns. This can only arise from unfamiliarity with statistical 

 work on a large scale. Under the voluntary system there were 

 some refusals, or failures, to make a return. The proportion 

 varied in different districts, but over the whole country only 

 about 3 per cent, of the total number of occupiers failed to fill 

 up the form or to give the collecting officer information to 

 enable him to do so. I have actually been asked whether the 

 absence of this 3 per cent, implied that the totals published 

 for the whole country were deficient to this extent, as though 

 it were possible for any reasonably honest person to put his 

 name to a statement purporting to give a total which he knew 

 to be untrue. It is hardly necessary to say that it was part of 

 the collecting officer's duty to furnish an estimate of the crops 

 and live stock on any holding for which he failed to get an 

 actual return. The fields and live stock of a farm are, broadly 

 speaking, visible to anyone, and with the aid of the rate book 

 it is not very difficult to make a fair estimate within a reason- 

 able margin of error. The presence of 3 per cent, of such 

 estimated returns could not seriously affect the validity of the 

 totals. 



It is also urged that a compulsory return is more accurate 

 than a voluntary one. This is probably so in some degree, 

 though the additional accuracy may easily be over-estimated. 

 The important point is the absence of deliberate bias. If the 

 persons making the returns have an interest, or think they 

 have, in under-statement or over- statement of the figures, it 

 may be confidently assumed that the returns as a whole will 



