The Ashburton Guardian. Magna est Veritas et Prævalebit. WEDNESDAY, APRIL 3, 1889. MISLEADING STATISTICS.
Although proverbially exceedingly dry, considered as ordinary literature, statistics are, nevertheless, exceedingly useful, and indeed indispensable to the' business man, no less than to the politicoeconomist or the journalistic scribbler. But it is also proverbial that " figures can be made to prove anything," and, we would add, especially when the figures themselves are more or less unreliable. Now we do not wish to imply blame to this colony's painstaking statistician, Mr W. F. B. Brown, when we say that the elaborate tables of statistics issued from his office are in many respects very unsatisfactory — the fault is not carelessness of compilation, but that of radical defects in the system, which, however, are so patent that reform is absolutely necessary. As regards the tables of popula- , tion of our principal cities, it has already been pointed out in several quarters that a method which takes no account of the population of their immediate suburbs and which estimates their relative numbers upon the basis of certain arbitrary municipal boundaries is wholly misleading and unsatisfactory — because giving absolutely incorreot information at least so far as all but an initiated few (and those few of course within the colony itself) are concerned. Then again as regards the vital statistics us they are termed (the tables which give proportions of deaths and of cases of different diseases) not only does the same radical error affect their statistical correctness, but the very list of places for which the figures are given needs revision, several centres of population, having as large, or larger, populations than those which are included in the returns, being left out of consideration. Yet again, in relation to immigration and emigration, we are by no means satisfied that anything like exactitude is obtained as regards the numbers of those leaving the colony. The number of arrivals is easily checked ; not bo that of the departures, but we think that it would be quite possible to devise a method of attaining • more accurate figures than are now obtained. But in no direction is there more need for reform than in connection with the agricultural statistics. (Jn the appearance of those for this year, which showed that the estimated average yield per acre of wheat was smaller this year than last year, we at once said that there must be something wrong, either in this year's figures or in those for last year. We ourselves were able to give instances showing that the yield had been underestimated, and argued that the discrepancy as to the results of the two years was probably explainable bp the fact that underestimates had this year been very general. That this view of the case was a correct one appears to be demonstrated by information furnished to an exchange by a Wellington correspondent who tele graphed a day or two ago as follows. — Attention has been called by Bome of | the collectors of agricultural statistics to ! the fact ikut these are obtained too early in the season to be trustworthy, and that they consequently possess little , practical value. At present they are taken before the farmers have been ab'e to thresh out any of the grain, and consequently the statements made as to the yield per acre are mere guesses, often of the wildest and most inaccurate nature. Information has been received subse quently from grain growers expressing regret at having unintentionally misled the collectors, as they found their yield was enormously greater, owing to the favorable character of the season, than they had ever dreamed of getting. One man estimated a large area of wheat at 82 bushels to the acre. He got 45. Another put his oats at 40, and got 70, and so on. Experience proves that this year in particular the actual 'grain yield has been generally from 20 to .50 per cent more than the growers estimated before threshing. A good^-deal of difficulty has been experienced by the collectors of agricultural statistics in inducing farmers and others to make a full and true declaration of live stock in their possession. Some said that they were afraid some new tax would bo imposed on them if they gave a full list. On being warned that they would be liable to a penalty for refusing, they gave a statement which jn several pases was subsequently discovered to be a long way short of the truth," It may therefore, we think, be taken as pro?en that the amount of grain grown this year is considerably larger than the totals given in the agricultural statistics, and hence that if these annual statistics are to be a trustworthy guide s.ome means must be taken to arrive at greater accuracy. * It has been suggested that this would be attained by taking the returns at a later date, when farmers have had time to make at least an experimental threshing, and whether by this means or by some other, certain it is that reform is needed if the statistics are to be of any value. In point of fact it seems to us that the JRegistrar- General would do well to go in for a thorough revision of his methods and a general overhaul of the entire lepftrtment under bil charge,