I refer to your editorials headlined, 'Number discrepancy' (October 20) and 'Odd numbers' (October 25). Your first editorial questioned the findings of an official survey to gauge the number of mainlanders eligible for right of abode, citing a Hong Kong Human Rights Monitor report to support your argument. Critics of the survey conducted by the Hong Kong SAR Government tended to ignore the fact that it was a large-scale territory-wide survey for which a scientific sampling method was employed. Data collection was performed professionally and followed stringently the methodology adopted in modern surveys. Altogether 46,000 household members were interviewed. The respondents formed an unbiased sample which was fully representative of the Hong Kong community. Since publication of its results, the survey including its methodology has been scrutinised by many parties. Following discussions, and with detailed explanations given by the Commissioner for Census and Statistics on various occasions, there is now wide acceptance of its reliability. In the government survey two interview methods were employed in parallel - the conventional 'Direct Questioning Method' (DQM) and the 'Randomised Response Technique' (RRT). The DQM was applied to half of the household members in the sample and the RRT to the other half. The DQM findings were discarded because this method failed to elicit accurate data. As reported by the interviewers, the respondents were inhibited, felt uneasy, or gave perfunctory answers when directly questioned, thus rendering this method inappropriate. Respondents to the RRT went through a random process which determined whether they were to answer the sensitive question or an unrelated question on taxi patronage. There were no inhibitions and thus no disincentives to impede them from giving the true answers. Reliance on RRT is not unique to Hong Kong. For more than three decades this survey method has been accepted in many statistically-advanced countries for studying highly sensitive social issues. The survey conducted by Pam Baker and Co. covers a sample size of less than 5,000. All respondents are mainlanders who have come to Hong Kong at a specific time for the purpose of seeking right of abode. In technical terms, this is a 'self-selected' sample which is by definition biased. The survey organiser is the legal representative of the respondents who claim to be eligible for right of abode in Hong Kong. The eligibility for right of abode is the subject of a case currently being heard before the Court of Final Appeal. It would not be appropriate for me to comment further. Turning to your second editorial of October 25, the cessation of the survey started by the mainland authorities was decided immediately after the Standing Committee of the National People's Congress (NPCSC) had interpreted the relevant provisions of the Basic Law. Nothing ever suggested that there might be any significant discrepancies between the SAR Government's survey and the mainland survey, if the latter was continued. But the resource requirements for the survey to be carried out in the initially selected cities, and indeed when extended to other relevant places in the mainland, would have been enormous. The Public Security Bureau considered it unnecessary to continue with the survey since the NPCSC had already interpreted the Basic Law provisions relevant to the eligibility of right of abode. TIMOTHY TONG for Secretary for Security