Sessional_Paper_1931 — Page 99

93

(c) Preparatory Work at the Census Office.

There are two distinct methods of taking a Census of the population which have been designated respectively as (i) the Householder Method and (ii) the Canvasser Method. In the first the occupier of each dwelling is the person held primarily responsible for furnishing a written record of the particulars required from each of the inmates of the dwelling occupied by him. In the second method the original entries are made by the enumerator as the result of enquiries instituted by him.

There are many advantages in the Householder Method, namely, (i) the liability of error is diminished, as all statements rest on the householder's own authority, and erroneous entries due to hurry or carelessness in taking down the answers cannot occur; (ii) the time of the enumerator is saved and the cost is thereby reduced; (ii) the labour of writing out the particulars is divided; and (iv) such division of labour makes it possible for the collection of particulars to be synchronous for all parts of the country.

The chief objection to the Householder Method is that the accuracy of a census carried out through its agency is limited by the intelligence of the population, and their educated capacity to grasp the meaning of the blank forms and to fill them up properly. It is clear that in Hong Kong this educated capacity is not possessed by a large proportion of the Chinese population.

The chief advantage of the alternative Canvasser Method is that difficulties due to the ignorance or illiteracy of the population are to a large extent obviated. But to make this method a success the enumerators appointed must be of high intelligence and should have not only a clear knowledge of what is required and how it should be recorded, but should also have the ability readily to obtain the desired information by means. of questions well chosen and carefully asked.

The chief disadvantage of the Canvasser Method lies in the fact that the enumeration, owing to the time taken in obtaining the desired information, is not simultaneous for the whole population and thus it becomes difficult to avoid omis- sions and duplications. To reduce as much as possible the objectionable features of this method, special expedients have been adopted in several countries, e.g. India and Ceylon, by having a preliminary canvass followed by a final revision on Census Night. This method was employed in Hong Kong on this occasion.

In previous censuses in Hong Kong the Householder Method was mainly adopted with a post-census addition of the Canvasser Method. The enumerators, who were few in number, were instructed to give, when delivering the forms, as full an explanation of the purport of the form as might be necessary and to help the public, but so far as can be seen little or nothing was done prior to the Census date. The system then employed, therefore, was one which depended for its accuracy on the intelligence and educated capacity of the people, while the population dealt with was largely illiterate and quite uninterested in, if not hostile to, the purposes of the census. An attempt was made to fill in the gaps left by the illiterate by having the schedules filled in by the enumerators when collecting them, but this, of course, was done after the date of the census. As the forms might not be collected for several days after the Census Night, a serious error must have crept in due to defective recollection of the facts as they existed on Census day.

To avoid this source of error, while at the same time retaining synchronous enumeration, the method of having a preliminary census followed by a rapid revision on the morning after Census Night was followed, the intention being that the enumerator could, during the preliminary period which commenced on February 19th (17 days before Census Night), enter at his leisure particulars of the persons likely to be in his "block" on the Census date and would have as little writing as possible to do on making the final revision and collecting the forms on the morning after the Census date.