南京大虐殺・従軍慰安婦強制連行の嘘

　 ところで、北村教授の『「南京事件」の探求』（文春文庫）にも指摘されていますが、スマイス調査の内、the Agricultural Surveyについては奇妙な推計方法が採用されています。
　 前文によれば
In the Agricultural Survey two investigators were sent out across each of the six hsien armed with passes from three organizations. They were instructed to follow a main road one way and then to return zigzagging across the main road in the form of a figure eight in order to cover the areas back from the main roads. On this circuit they were to secure a village schedule from every third village on their route and in that village fill out one farm schedule for every tenth farm family that had returned to the village.
　 ３集落毎に、１０軒に１軒のサンプル調査を行っています。
　 これは良く知られている事柄です。
In the Agricultural Survey the average per farm family studied was worked out by hsien and then that average was multiplied by the total number of farm families in each hsien, as given by Professor Buck in his Land Utilization in China. The grand total was secured by adding up the hsien totals and any grand averages were then computed from these totals, thereby giving grand averages weighted according to the number of farm families in each hsien.
　 しかし、その調査結果をこの様に処理していることについては、余り詳しく知られていないかも知れません。
　 具体的な手順は本文から、
(1)サンプル調査で不在率を算定
(2)1932年Buck調査の総数に不在率を掛けて人口を推定
(3)更に、推定人口に死亡率を掛けて死亡者数を推定
というロジックを採用しているものと読み取れますが、この推計方法が成り立つのは母集団に大きな変化が無い＝近郊農村地帯の世帯分布が1932年調査時点から変化していないと見られる場合だけであり、社会統計の専門家であるスマイスがそれを知らないはずはありません。(例えば、集落が丸ごと疎開したようなケースはカバーできません。)
　 しかし果たして、中国軍の清野作戦で大打撃を受けた進軍ルートの農村地帯が、1932年と同じ世帯分布を保っていたなどという仮定に、常識的に考えた上で合理性があるでしょうか？
　 THE WAR AND PERSONS MIGRATION FROM THE FARMSでは農村地帯の避難民について正確な数がつかめないことを次のとおり認めています。
Investigators' reports show that 133,000 members of farm families resident in March (11 per cent of the estimated original members of those families) had left and had not returned. It must be remembered that possibly three times as many persons in entire families were still away; but we cannot accurately consider them because of inadequate information.