Research Design返回

The public survey comprised two groups of samples, namely, a general sample and a booster sample, with target sample sizes of 1,000+ and 300+ respectively. The target population of the former group was Cantonese-speaking population of Hong Kong aged 15 or above, whereas that of the latter group was Cantonese-speaking full-time students aged 15 or above. Such a design facilitated the comparison of opinions held by the two groups of people, while assuring the representativeness of both samples.

For the study of expert opinion, subjects were conveniently sampled from the following categories -

Group 1: Environmental groups
Group 2: Arts and cultural groups
Group 3: Professional bodies
Group 4: Academic bodies
Group 5: Principals of secondary schools
Group 6: The media
Group 7: Analysts and bankers
Group 8: Voluntary participants

It should be noted that Group 8 of the expert study was only added in the middle of the qualitative study, to cater for some additional contacts provided voluntarily to POP by a respondent in Group 2. All contacts from Groups 1 to 7, amounting to 621, were provided to POP by the Sino Group which commissioned the study, while Group 8 comprised 12 fax numbers supplied by the Ink Society of Hong Kong. After screening out the invalid numbers (e.g. international telephone numbers), a total of 619 valid numbers remained for the qualitative study.

The representative survey of the general public was conducted by telephone interviewers under close supervision. To minimize sampling bias, telephone numbers were first drawn randomly from the residential telephone directories as "seed numbers", from which another set of numbers was generated using the "plus/minus one/two" method, in order to capture the unlisted numbers. Duplicated numbers were then filtered, and the remaining numbers were mixed in random order to produce the final telephone sample. When telephone contact was successfully established with a target household, the target popluation was selected. If more than one subject had been available, selection was made using the "next birthday rule" which selected the person who had his/her birthday next from all those present.

The data collected in the general public survey have been adjusted according to the age and gender distributions of the Hong Kong population as reported in the 2001 Population Census. All analyses on the general sample in this report are based on the weighted data. However, no such treatment has been applied to the booster sample since no government statistical data regarding students' demographic profile is available.

2. Contact information of the general public survey

Telephone interviews were conducted during the period of 22-26 September 2004. A total of 1,009 Cantonese-speaking Hong Kong citizens aged 15 or above were successfully interviewed. The overall response rate of this survey was 61.4% (Table 1), and the standard sampling error for percentages based on this sample was less than 1.6 percentage points. In other words, the sampling error for all percentages was less than plus/minus 3 percentage points at 95% confidence level.

Table 1 Calculation of response rate (general sample)

Response rate
= [ Successful cases / (Successful cases + Incomplete cases* + Refusal cases by eligible respondents^) ]
= 1,009 / ( 1,009 + 616 + 18 )
= 61.4%

* Including "partial interview" and "interview terminated before the screening question"
^ Including "household-level refusal" and "known respondent refusal"

As shown in Table 2, among the 8,270 telephone numbers sampled for the survey, 2,925 were confirmed to be ineligible, among them 437 were fax or data lines, 1,735 were invalid telephone numbers, 69 were call-forwarding numbers, while another 439 were non-residential numbers. Besides, 146 of them were invalidated due to special technological reasons, while 99 cases were voided because target respondents were unavailable at the numbers provided.

Meanwhile, a total of 2,737 telephone numbers were invalidated before the research team could confirm their eligibility. Among them 153 were busy lines and 1,617 were no-answer calls after making a maximum of 5 times' recalls. Sixty-one cases were diverted to answering devices while another 128 were blocked. Moreover, 243 cases were treated as unsuccessful because of language problems, while 529 interviews were terminated before the screening question. Six more cases were voided for other problems.

On the other hand, 1,599 cases failed to complete the interview after confirming their eligibility. Among them 12 were rejected at the household level, another 6 rejected the interview immediately after their eligibility was confirmed, 1,488 were unfinished cases with appointment dates beyond the end of fieldwork period. Besides, 87 cases were incomplete due to unexpected termination of interviews, 6 were classified as miscellaneous due to other non-contact problems, and the remaining 1,009 were successful cases (Table 2).

Table 2 Detailed breakdown of contact information of the survey (general sample)

	Frequency	Percentage
Respondents' ineligibility confirmed	2,925	35.4
Fax/ data line	437	5.3
Invalid number	1,735	21.0
Call-forwarding/ mobile/ pager number	69	0.8
Non-residential number	439	5.3
Special technological difficulties	146	1.8
No eligible respondents	99	1.2

Respondents' eligibility not confirmed	2,737	33.1
Line busy	153	1.9
No answer	1,617	19.6
Answering device	61	0.7
Call-blocking	128	1.5
Language problem	243	2.9
Interview terminated before the screening question	529	6.4
Others	6	0.1

Respondents' eligibility confirmed, but failed to complete the interview	1,599	19.4
Household-level refusal	12	0.1
Known respondent refusal	6	0.1
Appointment date beyond the fieldwork period	1,488	18.0
Partial interview	87	1.1
Miscellaneous	6	0.1

Successful cases	1,009	12.2

Total	8,270	100.0

3. Contact information of the booster sample

Telephone interviews were conducted during the period of 26-29 September 2004. A total of 311 Cantonese-speaking full-time students aged 15 or above were successfully interviewed. The overall response rate of this survey was 84.0% (Table 3), and the standard sampling error for percentages based on this sample was less than 2.8 percentage points. In other words, the sampling error for all percentages was less than plus/minus 6 percentage points at 95% confidence level.

Table 3 Calculation of effective response rate (booster sample)

Effective response rate
= [ Successful cases / (Successful cases + Incomplete cases* + Refusal cases by eligible respondents^ + Refusal cases by prorated-eligible respondents#) ]
= 311 / { 311+2+5+146 [(311+2+5) / 311+2+5+568] }
= 84.0%

* Including "partial interview" only.
^ Including "household-level refusal" and "known respondent refusal"

As shown in Table 4, among the 2,972 telephone numbers sampled for the survey, 1,183 were confirmed to be ineligible, among them 87 were fax or data lines, 411 were invalid telephone numbers, 13 were call-forwarding numbers, while another 63 were non-residential numbers. Besides, 41 of them were invalidated due to special technological reasons, while 568 cases were voided because target respondents were unavailable at the numbers provided.

Meanwhile, a total of 994 telephone numbers were invalidated before the research team could confirm their eligibility. Among them 163 were busy lines and 572 were no-answer calls after making a maximum of 5 times' recalls. Forty-two cases were diverted to answering devices while another 35 were blocked. Moreover, 36 cases were treated as unsuccessful because of language problems, while 146 interviews were terminated before the screening question.

On the other hand, 484 cases failed to complete the interview after confirming their eligibility. Among them 4 were rejected at the household level, another one rejected the interview immediately after their eligibility was confirmed, 476 were unfinished cases with appointment dates beyond the end of fieldwork period. Besides, 2 cases were incomplete due to unexpected termination of interviews, one were classified as miscellaneous due to other non-contact problems, and the remaining 311 were successful cases (Table 4).

Table 4 Detailed breakdown of contact information of the survey (booster sample)

	Frequency	Percentage
Respondents' ineligibility confirmed	1,183	39.7
Fax/ data line	87	2.9
Invalid number	411	13.8
Call-forwarding/ mobile/ pager number	13	0.4
Non-residential number	63	2.1
Special technological difficulties	41	1.4
No eligible respondents	568	19.1

Respondents' eligibility not confirmed	994	33.4
Line busy	163	5.5
No answer	572	19.2
Answering device	42	1.4
Call-blocking	35	1.2
Language problem	36	1.2
Interview terminated before the screening question	146	4.9

Respondents' eligibility confirmed, but failed to complete the interview	484	16.2
Household-level refusal	4	0.1
Known respondent refusal	1	0.0
Appointment date beyond the fieldwork period	476	16.0
Partial interview	2	0.1
Miscellaneous	1	0.0

Successful cases	311	10.5

Total	2,972	100.0

The contact information of the two samples are summarised as follow:

Table 5 Summary of the contact information of the public survey

Date of survey	:	General Sample (GS) - 22 to 26 September 2004
		Booster Sample (BS) - 26 to 29 September 2004
Sample size	:	GS - 1,009 successful cases; BS - 311 successful cases
Response rate	:	GS - 61.4%; BS - 84.0%
Standard error	:	GS - Less than 1.6%; BS - Less than 2.8%
Target population	:	GS -Cantonese-speaking population aged 15 or above.
		BS -Full-time students aged 15 or above.
Survey method	:	GS/BS - Telephone survey with interviewers.
Sampling method	:	GS/BS - Telephone numbers were selected randomly from telephone directories and mixed with additional numbers generated by the computer.
Survey design	:	GS/BS - The questionnaire and fieldwork were designed and conducted independently by the POP Team.
Weighting method	:	GS - The data reported have been adjusted according to the age and gender distributions of the HK population as reported in the 2001 Population Census.BS - No statistical weighting has been applied.

4. Contact information of the study of experts

As explained before, the list of experts consisted of 633 telephone numbers divided into 8 separate groups. After filtering off invalid numbers, a total of 619 valid numbers remained, and they were distributed as follows:

Table 6 Details of the contact lists

	Number of raw numbers		Number of valid numbers
Group 1	4	)	4	)
Group 2	49	)	44	)
Group 3	9	)	9	)
Group 4	15	)	15	)
Group 5	487	)	479	)
Group 6	17	)	17	)
Group 7	40	) 621	40	) 608
Group 8	12	11
Overall	633	619

All respondents were given the liberty to choose from 3 survey modes: answering questions directly on the telephone, completeing a paper questionniare and return it by fax, or filling in the online survey. Of the 186 successful cases, a landslide majority of 155 cases opted for the telephone mode, 31 returned the paper questionnaire via fax, whilst nobody chose the online mode. The number of successful cases registered by each mode is summaried as follows:

Table 7 Number of successful cases registered by different modes

	Valid tel. numbers	Telephone interviews	Returned by fax	Total cases
Group 1	4	0	0	0
Group 2	44	9	4	13
Group 3	9	0	0	0
Group 4	15	2	0	2
Group 5	479	120	23	143
Group 6	17	6	1	7
Group 7	40	18	1	19
Group 8	11	0	2	2
Overall	619	155	31	186

Interviews were conducted during the period of 23 September - 6 October 2004. A total of 186 subjects were successfully interviewed, after the research team has exhausted all numbers on the contact lists, and made a maximum of 5 attempts before dropping a number. The overall response rate of this part of the study was 56.0% (Table 8). However, because this part of the study is not meant to be representative, there is not much point in discussing the error margin. Views collected are only meant for qualitative discussions.

Table 8 Calculation of response rate

Response rate
= [ Successful cases / (Successful cases + Incomplete cases* + Refusal cases by eligible respondents) ]
= 186 / ( 186 +10 + 136 )
= 56.0%

* Including "partial interview" and "interview terminated before the screening question"

Across the groups, "media workers" (Group 6) and "analysts and bankers" (Group 7) seemed to be the most enthusiastic, with the respective response rates of 78% and 76% (Table 9). Details of the response rate of each group are shown in the following table:

Table 9 Response rate of each group

	Valid numbers	Successful cases (A)	Partial cases (B)	Refusals (C)	Response rate*
Group 1	4	0	0	2	0.0%
Group 2	44	13	1	14	46.4%
Group 3	9	0	1	6	0.0%
Group 4	15	2	0	5	28.6%
Group 5	479	143	7	93	58.8%
Group 6	17	7	0	2	77.8%
Group 7	40	19	1	5	76.0%
Group 8	11	2	0	9	18.2%
Overall	619	186	10	136	56.0%

* Response rate is calculated as the number of successful cases (A) divided by the sum of successful cases (A) plus refusals from confirmed eligible respondents (C) and incomplete cases (B).

As shown in Table 10, among the 619 telephone numbers sampled for the survey, 56 were confirmed to be ineligible, among them 17 were fax or data lines, 5 were invalid telephone numbers, while another one was non-office number. Besides, 10 of them were invalidated due to special technological reasons, while 23 cases were voided because target respondents were unavailable at the numbers provided.

Meanwhile, a total of 153 telephone numbers were invalidated before the research team could confirm their eligibility. Among them 21 were busy lines and 101 were no-answer calls after making a maximum of 5 times' recalls. Nine cases were diverted to answering devices while another 13 were blocked. Moreover, one case was treated as unsuccessful because of language problems, while 8 interviews were terminated before the screening question.

On the other hand, 224 cases failed to complete the interview after confirming their eligibility. Among them 136 rejected the interview immediately after their eligibility was confirmed, 85 were unfinished cases with appointment dates beyond the end of fieldwork period. Besides, 2 cases were incomplete due to unexpected termination of interviews, another one has passed away, and the remaining 186 were successful cases (Table 10).

Table 10 Detailed breakdown of contact information of the survey

	Frequency	Percentage
Respondents' ineligibility confirmed	56	9.0
Fax/ data line	17	2.7
Invalid number	5	0.8
Non-office number	1	0.2
Special technological difficulties	10	1.6
No eligible respondents	23	3.7

Respondents' eligibility not confirmed	153	24.8
Line busy	21	3.4
No answer	101	16.3
Answering device	9	1.5
Call-blocking	13	1.5
Language problem	1	0.2
Interview terminated before the screening question	8	1.3

Respondents' eligibility confirmed, but failed to complete the interview	224	36.3
Known respondent refusal	136	22.0
Appointment date beyond the fieldwork period	85	13.7
Partial interview	2	0.4
Respondent dead	1	0.2

Successful cases	186	30.0

Total	619	100.0

Since the majority of cases in this survey came from secondary school principals, when interpreting the overall figures of this report, special caution should be taken to allow for the domination of opinions coming from this group, even though we are looking at them qualitatively.

The contact information of this part of the study is summarised as follow:

Table 11 Summary of the contact information of the expert survey

Date of interviews	:	23 September to 6 October 2004
Sample size	:	186 successful cases
Response rate	:	56.0%
Target population	:	Experts from environmental groups, arts/cultural groups, professional bodies, academic bodies, secondary school principals, the media, analysts and bankers, and voluntary participants referred by respondents.
Interview method	:	Mixed modes of telephone interviews, fax and online questionnaire surveys
Selection method	:	Most contacts were supplied by the Sino Group, supplemented by a few provided voluntarily by the Ink Society of Hong Kong.
Survey design	:	The questionnaire and fieldwork were designed and conducted independently by the POP Team.