Research DesignBack


1. Survey method
 
Data was collected through a random telephone survey conducted by interviewers under close supervision.
 
2. Target population and sample size
 
The target population of this survey was Cantonese-speaking residents in Hong Kong of age 18 or above. The total sample size of this survey was set at 1,000+ successful interviews with the qualified respondents.
 
3. Sampling method
 
At the first level of sampling, telephone numbers were first be drawn randomly from residential telephone directories to become "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 be filtered, and the remaining numbers mixed in random order to produce the final telephone sample. At the second level of sampling, when telephone contact was successfully established with a target household, one person qualified was selected from all those present using the "next birthday" rule (i.e. select the member whose birthday is nearest to the interview date).
 
4. Fieldwork operation
 
Between 16 April and 4 May 2004, POP deployed a pool of well-trained telephone interviewers to conduct the interviews. All data were collected by interviewers using a Computer Assisted Telephone Interview (CATI) system, which allowed real-time data capture and consolidation. To ensure data quality, voice recording, screen capturing and real-time camera surveillance, on top of on-site supervision, were used to monitor interviewers' performance.
 
5. Response rate calculation
 
Response rate is calculated as the number of successful cases divided by the sum of successful cases plus effective refusals from confirmed eligible respondents and incomplete cases. Each target telephone number was attempted 5 times before the number was dropped as "busy", "no answer" or "answering device".
 
POP's standard formula for calculated response rate is as follows:

   Response rate
= [ Successful cases / (Successful cases + Incomplete cases* + Refusal cases by eligible respondents^) ]
* Including "partial interview" and "interview terminated before the screening question"
^ Including "household-level refusal" and "known respondent refusal"
Detailed contact information of the survey has been given in the frequency tables.
 
6. Application of significance test
 
A "difference-of-proportions" test was employed whenever applicable to check for significant changes. Figures marked with double asterisks (**) denoted that the result has been tested to be statistically significant at p=0.01 level, whereas those with single asterisk (*) denoted statistical significance at p=0.05 level.