Occupations in Shortage in New Zealand: 2006
Previous Section | Table of Contents
Appendix 1: Methodology for the Survey of Employers who have Recently Advertised
The SERA is a short telephone survey of employers who have recently advertised job vacancies. The purpose of the survey is to identify occupations in which employers are having difficulty filling vacancies.
Research New Zealand was contracted by the Department of Labour to carry out SERA 2006 from May to October 2006. A total of 3,534 successful interviews were completed with employers (or recruitment agencies). These employers were attempting to fill a total of 5,486 vacancies[8]. The response rate for the survey was 64%.
Earlier large sample SERA surveys were conducted in-house by the Department in 2003 (February to December) and 2005 (May to August), with no equivalent survey being conducted in 2004.
The survey sample was drawn primarily from the Department's Job Vacancy Monitor[9] (JVM) but was also supplemented from various internet job boards and specialist websites when there was difficulty reaching sample quotas for particular occupations from the JVM. The SERA 2006 sample was stratified on the following basis:
- A total of 2,500 completed interviews covering three major occupation groups: professionals, technicians and associate professionals, and trades workers. The sample size for each individual 5-digit occupation was determined by the share of employment of the 5-digit occupation from total employment in the three major groups. A cap of 40 interviews was placed on large occupations, and a minimum of 10 interviews was sought for some small occupations that were thought to be of particular interest.
- A total of 1,000 completed interviews covering all six other major occupation groups using a similar methodology to that described above.
Employers were approached six to ten weeks after advertising and were asked whether they had filled their vacancy, and the number of suitable applicants (including whether they had the right to work in New Zealand at the time they were interviewed). Employers were phoned back if they had not yet filled the position on offer, but they expected to do so within ten weeks of advertising.
Vacancies that were not filled with a suitable candidate within ten weeks of advertising were deemed to not be filled (even if there was some chance of the vacancy being filled after the ten week cut-off point). It is possible that this business rule may impact on some occupations more than others (e.g. highly paid positions, which are typically advertised through recruitment agencies, may take longer to finalise appointments).
If the appointee to any position did not have the qualifications and experience to be regarded by the employer as a member of the occupation being advertised (e.g. a book-keeper being appointed to an accountant vacancy), then the vacancy was deemed to not be filled.
Weighting
Fill rates presented at the NZSCO 1-digit, 2-digit and 3-digit level were weighted to compensate for any under or over sampling of individual occupations in the survey. Weights were calculated for each occupation[10] with a view to make the vacancy counts from the SERA sample reflective of the total newspaper vacancy counts from the JVM. Data from the 2006, 2005, and 2003 SERA surveys were each weighted separately based on vacancy levels in the JVM at a similar time to the survey being carried out. As the JVM does not currently include web-based vacancies or vacancies published in community papers, the weighted fill rates are only reflective of newspaper vacancy counts rather than total vacancy counts.
The SERA surveys carried out in 2003 and 2005 both excluded a few 5-digit occupations from the technician and associate professional occupational group (e.g. sales representatives and technical representatives). As these occupations account for a reasonable proportion of advertised newspaper vacancies within this NZSCO group, the weighted fill rates for the major group from 2003 and 2005 are not comparable with that for 2006 (which did include these occupations). For this reason, fill rates are not shown in the report for 2003 and 2005 for the NZSCO groups 3, 33, and 331.
Outliers
Individual employers who were advertising 15 or more vacancies for an individual occupation were excluded from calculations so as to not unduly bias fill rates. It can be noted that 72% of employers were advertising only one vacancy and 16% were advertising two vacancies, with only 0.7% of employers advertising 15 or more vacancies.
Comparability with previously published figures
Fill rates presented in this report from the 2003 and 2005 SERA surveys may differ from those published previously for two reasons. Firstly, the weighting system used in this report for the grouped categories is different to that applied previously. Figures presented in last years report at the NZSCO 1-digit, 2-digit and 3-digit level were incorrectly weighted based on employment share. Fill rates for 5-digit occupations are not affected as they are not weighted.
Secondly, a change was made in the methodology concerning outliers. Figures published previously were calculated by removing employers who advertised 20 or more vacancies. However, the threshold for counting an employer as an outlier was changed in the current study so that employers seeking to fill 15 or more positions were excluded. Fill rates for 2003 and 2005 were recalculated using this refined methodology.
Sample Error
The fill rates presented in this report are subject to sampling error, which must be considered when interpreting the results. The larger the sample size that a fill rate is calculated from (in this case, the number of vacancies included in the SERA), the smaller the sample error will be. For example, the maximum margin of error at a confidence level of 95% for various sample sizes is as follows:
| Sample size | Margin of error |
|---|---|
| 20 | ±21.9% |
| 40 | ±15.5% |
| 60 | ±12.5% |
| 80 | ±11.0% |
| 100 | ±9.8% |
That is, if the number of vacancies sampled was 40 and the fill rate from the survey was 50%, you could say that you were 95% confident that the actual fill rate for employers lies between 34.5% and 65.5% (i.e. 50% ±15.5%).
Questionnaire
The SERA 2006 questionnaire is available upon request.
For further information contact: info@dol.govt.nz
[8] Short-term vacancies (i.e. for periods of less than three months) were excluded from the survey.
[9] The JVM programme began in November 2002 to gather and analyse detailed information on job vacancy advertisements. The JVM involves the monthly analysis of advertised job vacancies in 25 daily newspapers and two internet IT job boards.
[10] For the NZSCO major groups 2, 3 and 7 (which had the largest sample sizes in SERA 2006), weights were calculated at the 5-digit level, while for the other major occupational groups, weights were calculated at the 3-digit level.