Both surveys were conducted by Research New Zealand (formerly BRC Marketing and Social Research). This section, written by Research NZ, describes the survey methodology.

Sampling - employers survey

In 2005, the sampling frame for the survey of employers was all business enterprises (with known contact details) with five or more employees, operating in the nine specific ANZSIC groupings listed in Table 1 below. A total sample of n=1,100 employers with five or more employees were interviewed for the 2005 survey.

In consultation with the Department, this sample size was replicated in the 2008 survey with the inclusion of two extra ANZSIC groupings (Government, Administration and Defence, which was not included in the 2005 survey sample frame, and the Financial and Insurance sector, which has since been 'split-out' from the Property and Business Services sector as a separate ANZSIC grouping). In addition to the expanded number of ANZSIC groupings, the sample frame for the 2008 employers survey was expanded to also include enterprises with fewer than five full-time employees (FTEs).

The sample for the employers survey was purchased from apnfinda Limited, which is an Auckland-based company specialising in contact databases. This organisation holds New Zealand's largest, most up-to-date and comprehensive business database (a result of purchasing and consolidating the UBD and Wise business databases) that is available for commercial purchase.

In total, apndata (a business unit of apnfinda Limited) provided Research New Zealand with the contact details of senior managers and business owners of 7,000 randomly selected businesses from throughout New Zealand, each of which employed one or more full-time staff. A quota scheme was used when sampling businesses identified as belonging to different ANZIC Groupings to ensure that the final achieved sample would be relatively representative of the distribution of such of enterprises in Statistics New Zealand's Business Frame, while still over-sampling specific industry groups for the purposes of analysis and reporting.

Sampling - employees survey

The target population for the survey of employees was all New Zealand residents, aged 15 years and over, who were currently working in paid employment in April and May 2008.

The primary sampling unit for the survey was all households with telephones that are listed in the White Pages of the New Zealand telephone directories. The secondary sampling unit was individuals aged 15 years and over living in these households who were currently working for wages or salary for 10 hours or greater per week at the time of surveying.

In order to achieve an optimal response rate and complete the telephone interviewing efficiently, an initial sample was drawn from the General and Maori Electoral Rolls. This provided the opportunity to send pre-notification letters to the households of potential respondents to the survey and to target potential respondent households by ethnicity (i.e. Maori, Pacific peoples and Asian peoples, Others).

For the purposes of effectively targeting potential respondent households by ethnicity, the Electoral Rolls were divided into three mutually exclusive lists, in the following order of priority:

• Those individuals on the Electoral Rolls who identify as being of Maori descent.
• All individuals residing in census mesh blocks known to have high proportions (30 percent of the mesh block total, or greater) of people identifying as being Pacific and/or Asian peoples in the 2006 Census.
• All other individuals in the Electoral Rolls not selected for one of the above two lists.

A sample of individual electors was randomly drawn from each of the three lists using a random number generator. The resulting lists were then sent to Telecom for tele-matching against the selected electors' mailing and residential addresses as detailed in the Electoral Rolls.

A total of 17,100 electors' details were provided to Telecom for matching, of which Telecom was able to match 8,395 records (a 49.1 percent match rate). Of the 8,395 matched records, 658 were identified as being unlisted numbers, leaving a final pool of 7,737 potential households to be approached to participate in the survey.

Questionnaire design

The draft questionnaires for both surveys were adapted from those used for the 2005 surveys, in consultation with Department of Labour staff, and ‘fine-tuned’ to better reflect the objectives of the intended evaluation of the Work-Life Balance Programme and the future review of the Employment Relations (Flexible Working Arrangements) Amendment Act 2007.

Prior to the commencement of the main surveying, a pilot survey with a sample of n=53 employers was carried out between 28 and 29 April, as well as a pilot survey with a sample of n=19 employees on 21 April in order to ensure that the data was captured in an appropriate manner and to the Department’s satisfaction.

Copies of the final questionnaires for both surveys are available on request.

Survey process

Prior to contacting the sampled employers and employees by telephone, a pre-notification letter was sent to their respective business or household to inform them that Research New Zealand was conducting the research on the Department’s behalf, and that they might be contacted to participate in a ten to 12 minute confidential interview.

As a matter of course, the letter also explained the purpose of the research; that it would be conducted in accordance with the Code of Practice of the Market Research Society of New Zealand (MRSNZ) and the Privacy Act 1993; and that Research New Zealand would report the research findings in a way that would ensure that individual respondents and businesses would remain anonymous.

The survey of employers was conducted as a CATI-based (Computer Assisted Telephone Interview) telephone survey of n=1,100 employers. The interviewing for the employers survey took place between 28 April and 28 May 2008. The employers survey achieved a response rate of 29.3 percent using the Statistics New Zealand’s response rate calculation (a participation rate of 39.9 percent was achieved, wherein the named respondent in the apnfinda sample was spoken to and consented to be interviewed). The average interview duration for the employers survey was 12.6 minutes.

The interviewing for the employees survey was undertaken between 21 April and 1 June. When a household was contacted, the interviewer asked to speak to a person normally living in the household, aged 15 years or older and who works 10 hours or more per week in paid employment, including self-employment. The employees survey achieved a response rate of 22.5 percent using the Statistics New Zealand’s response rate calculation (a participation rate of 33.3 percent, wherein it was determined that an eligible respondent was in the household and that person subsequently consented to be interviewed). The average interview duration for the employees survey was 14.4 minutes.

Post-stratification and margins of error

Prior to analysis and reporting, the achieved samples for both surveys were weighted to be representative of their respective populations.

In the case of the survey of New Zealand employers (Table 2), the achieved sample was weighted to be proportionally representative of the number of enterprises, estimated by Statistics New Zealand to have one or more employees in 2007 in each of the surveyed ANZIC Groupings.

The maximum margin of error, at the 95 percent confidence level, for the weighted achieved sample of n=1,100 employers is ± 3.1 percent, after allowing for a post-stratification design effect coefficient (Deft) of 1.047.

In the case of the survey of New Zealand employees, the achieved sample was weighted to be proportionally representative of the number of employed New Zealanders aged 15 years and older based upon their carer status and prioritised ethnicity, This weighting scheme was based upon 2006 Census data (Table 3).

The maximum margin of error, at the 95 percent confidence level, for the weighted achieved sample of n=1,004 employees is ± 3.2 percent, after allowing for a post-stratification design effect coefficient (Deft) of 1.034.