Research MethodBack

This "Opinion survey on white collars' mental well-being at work" was a random telephone survey conducted by telephone interviewers under close supervision. To minimize sampling bias, the following sampling technique was adopted:

The target population of this survey was Cantonese-speaking white collars of Hong Kong aged between 18 and 35. When telephone contact was successfully established with a target household, one white collar aged between 18 and 35 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.

This survey was conducted between 17 and 21 July 2003. A total of 515 qualified white collars were successfully interviewed. As shown from the calculation below, the effective response rate of this survey was 77.1% (Table 1), and the standard sampling error for percentages based on this sample was less than 2.2 percentage points. In other words, the sampling error for all percentages was less than plus/minus 4.5 percentage points at 95% confidence level.

^ Including "household-level refusal" (32 cases) and "known respondent refusal" (1 case).
* Figure obtained by prorata.

To guarantee that there were sufficient number of cases for sub-group analysis, quota sampling was adopted for this survey. A minimum quota of 160 cases was set for each of the three target age groups, namely, 18-25, 26-30 and 31-35. To increase the representativeness of the survey results, all raw data collected have been weighted according to the natural distribution of eligible respondents interviewed by age before the timeline cut at 2034 on 18 July 2003, i.e. before quota sampling came into operation. With this statistical adjustment in place, the figures reported in this report hereafter could then be generalized to represent the views of the target population, i.e. Cantonese speaking white-collars of age 18-35, subject to the sampling error mentioned in the previous section. It is also noteworthy that, due to the weighting effect, some individual figures reported might not round up to a hundred percent in the tables.

As shown in Table 2, among the 5,938 telephone numbers sampled for the survey, 1,917 were confirmed to be ineligible, among them 104 were fax or data lines, 616 were invalid telephone numbers, 13 were call-forwarding numbers, while another 105 were non-residential numbers. Besides, 97 of them were invalidated due to special technological reasons, while 982 cases were voided because target respondents were unavailable at the numbers provided.

Meanwhile, a total of 2,486 telephone numbers were invalidated before the research team could confirm their eligibility. Among them 46 were busy lines and 1,213 were no-answer calls after making a maximum of 5 times' recalls. Thirty cases were diverted to answering devices while another 498 were blocked. Moreover, 115 cases were treated as unsuccessful because of language problems, while 316 interviews were terminated before the screening question. Besides, 268 cases were voided for other problems.

On the other hand, 1,020 cases failed to complete the interview. Among them 32 were rejected at the household level, another 1 rejected the interview immediately after their eligibility was confirmed, 953 were unfinished cases with appointment dates beyond the end of fieldwork period. Besides, 6 cases were incomplete due to unexpected termination of interviews, 28 were classified as miscellaneous due to other non-contact problems, and the remaining 515 were successful cases (Table 2).