TNAG-0730-FCO40-934-Representations-by-Reform-Club-of-Hong-Kong-1978 — Page 7

2

mean that the findings are probably biased, because 990 respondents represent only slightly more than half of the selected household samples. If the reason for the unsuccessful household visits and interviews were that peopla were simply

not interested in this issue, then the survey resulta could

hardly be approached with confidence.

The samoling method

4.

In paragraph 1 of "Technical Appendix", it is stated that the SRH Master Sampling Frame was a sample of 25,000 households "designed and drawn so as to be representative of all the households in the Colony". There is no explanation of how this

done, nor is there sufficient information on what the

sampling fraction was.

It is difficult to think of any means

whereby SRH could have obtained a complete sampling frame for

this purpose other than from the records of the Census and Statiati Department. But no approach was made by SHH to the department ·

for such records.

5.

Even assuming that the sampling method was satisfactory,

the survey could only produce valid results in aggreyata. It could not produce valid results broken down by small sections such as

age groups and districts.

This is an important point, as the

report makes some play with "the younger generation".

CONFIDENTIAL