REVIEW
19
shall at the same time be simple enough to be included in a census questionnaire and searching enough to uncover at least two states of mind that in which the refugee once decided to leave his own country, and that in which he now decides not to go back there.
The same limitation caused the omission of some topics often included in censuses. We have already seen why no direct question was included on race or nationality, and why a woman could not be asked how many children she had ever borne. To explain why no question was asked on religion is not so easy if one starts with the premise, common to many countries, that each individual has an affiliation to some faith or sect which excludes him from all others. To a new arrival from Europe or India it may seem incredible that this premise is not accepted here, but the fact is that many Chinese do not ordinarily accept it, and receive the proposition that one must 'belong' to one religion and one only with the same amused incredulity as a man would who, having stocked his medicine chest with aspirin, cascara and iodine, is told that henceforward he must use one of the three exclusively for all ailments and throw away the other two. There would therefore be a large proportion of the population who would answer 'Yes' to every religion they had heard of and 'Don't know' to the rest, making the answers not only very hard to tabulate but quite useless when tabulated. It was therefore decided, as in the last two censuses, to omit the question and leave it to each religious organization to calculate how many members it possesses.
.
い
*
*
*
Having settled the date and the topics, the next matter to be decided was how the results were to be processed. In all previous censuses hand sorting had been used, but even in 1931 (when the population was only 849,751) this method was able to produce only simple lists, with no cross-tabulation except for sex, and consequent waste of a wealth of valuable information. With a population in excess of three million even simple lists could not be satisfactorily made by hand sorting and mechanical sorting with punched cards was used.
The difficulty would obviously be getting the cards punched. Limited use of punch-card systems was already being made in three departments, but the total 1959 establishment of 15 punch