Stats NZ has a new website.

For new releases go to

As we transition to our new site, you'll still find some Stats NZ information here on this archive site.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
Information about the survey

Survey background

 The New Zealand Government has a range of initiatives aimed at increasing New Zealand's economic growth rate above the Organisation for Economic Co-operation and Development (OECD) average and sustaining this higher growth performance over a number of years. For New Zealand’s economic performance to be measured against these initiatives, a wide range of data on a variety of measures needs to be collected.

Because of the wide range of data needed, Statistics New Zealand has developed an integrated, modular survey – the Business Operations Survey – as a way of collecting the required information while minimising the reporting load for New Zealand businesses. The survey has been designed to include up to three ‘modules’ and has been run annually since 2005.

The main objective of the survey is to collect information on the operations of New Zealand businesses in order to quantify business behaviour, capacity, and performance. In addition, each module in the survey has its own specific objectives. The modules included in the Business Operations Survey 2009 and their objectives are listed below.

Module A: Business performance module

The objective of this module is to provide a longitudinal series of information relating to business performance.

Module B: Innovation module

The objectives of this module are to:

  • provide information on the innovation characteristics of New Zealand private sector businesses that contributes to policy development which aids innovation
  • understand the dynamics of innovative businesses.

The innovation module runs every two years, and replaced Statistics NZ’s Innovation Survey, last run in 2003.

This module has been designed in accordance with OECD guidelines to develop understanding of the contribution of innovation to the New Zealand economy by measuring the following aspects:

  • levels of firm innovation
  • how and why firms collaborate with other firms and institutions in order to innovate
  • factors affecting the ability of firms to innovate
  • outcomes of innovation for firms, including its affect on exports.

Module C: Business practices module

This module collects data on a range of practices, some of which were collected in the Business Operations Survey 2005. In addition, questions relating to recent financing arrangements were also included to gain an understanding of the current situation.

Target population

The target population for the Business Operations Survey 2009 was live enterprise units on Statistics NZ’s Business Frame that at the population selection date:

  • were economically significant enterprises (those that have an annual GST turnover figure of greater than $30,000)
  • had six or more employees
  • had been operating for one or more years
  • were classified to the Australian and New Zealand Standard Industrial Classification – New Zealand Version 2006 (ANZSIC06) codes listed as ‘in scope’ in List 1 (see below)
  • were private enterprises as defined by New Zealand Institutional Sector 1996 Classification (NZISC96) (see ‘List 2: NZISC96 codes’).

An enterprise is defined as a business or service entity operating in New Zealand, such as a company, partnership, trust, government department or agency, state-owned enterprise, university, or self-employed individual.

The final estimated population size for the survey was 36,348 enterprises.

List 1: ANZSIC06 codes

In scope
ANZSIC06 code – description

A – Agriculture, forestry, and fishing
B – Mining and quarrying
C – Manufacturing
D – Electricity, gas, water, and waste services
E – Construction
F – Wholesale trade
G – Retail trade
H – Accommodation, cafes, and restaurants
I – Transport and storage
J – Information media and telecommunications
K – Financial and insurance services
L – Rental, hiring, and real estate services
M – Professional, scientific, and technical services
N – Administrative and support services
P – Education and training
Q – Health care and social assistance
R91 – Sport and recreation activities
R92 – Gambling activities
S94 – Repair and maintenance.

Out of scope
O – Public administration and safety
R89 – Heritage activities
R90 – Creative and performing arts activities
S95 – Personal and other services
S96 – Private household employing staff and undifferentiated goods and service producing activities of households for own use

List 2: NZISC96 codes

In scope
NZISC96 code – description
1111 – Private corporate producer enterprises
1121 – Private non-corporate producer enterprises
1211 – Producer boards
1311 – Central government enterprises
2211 – Private registered banks
2221 – Private other broad money (M3) depository organisations
2291 – Private other depository organisations nec
2311 – Private other financial organisations excluding insurance and pension funds
2411 – Private insurance and pension funds.

Out of scope
1321 – Local government enterprises
21 – Central bank
2212, 2213, 2222, 2223, 2292, 2293, 2312, 2313, 2412, 2413 – Central and local government financial intermediaries
3 – General government
4 – Private non-profit organisations serving households
5 – Households
6 – Rest of world

Sample design

The survey has been designed to produce aggregate statistics at a national level. This design does not allow statistics to be produced at a regional level.

Measurement errors

The Business Operations Survey 2009 results are subject to measurement errors, including both non-sample and sample errors. These errors should be considered when analysing the results from the survey.

Non-sample errors

Non-sample errors include mistakes respondents make when completing questionnaires, variation in the respondents’ interpretation of the questions asked, and errors made during the processing of the data. In addition, the survey applied imputation methodologies to cope with non-respondents. Statistics NZ adopts procedures to minimise these types of error, but they may still occur and are not quantifiable.

Given the nature of the data collected, there are limitations on the level of accuracy that the survey can give. Many respondents do not keep separate accounts of their innovation expenditure, or records may not be kept in the form required for the survey and estimation may be required. Even though the questionnaire had detailed descriptions of what should and should not be included as innovation, there may still be differences in interpretation of what constitutes innovation and the nature of any cooperative arrangements with other businesses involved in the innovation process.

Sample errors

The estimates in this report are based on a sample of businesses. While this sample has been has been chosen to be representative of the overall business population, somewhat different figures might have been obtained if a complete census of the entire business population had been taken using the same questionnaire and processing methods etc. Because the estimates are based on a sample of businesses, all estimates have a sampling error associated with them. The variability of a survey estimate, due to the random nature of the sample selection process, is measured by its sampling errors.

Most of the tables of this release are percentages of the total number of businesses in New Zealand in each size and industry category. The absolute sampling errors for the businesses population are presented in the following table. The table should only be used on the overall estimates that are percentages of all businesses. 

Table 14.01

Sample errors for Business Operations Survey 2009
 Size of estimate  Sampling error for total population Sampling error for innovators
1  0.4 0.6
2  0.6 0.8
3  0.7 1
5  1.0 1.2
10  1.3  1.7
20  1.8 2.3
30  2.0  2.6
 50  2.2  2.8
 70  2.0  2.6
 80  1.8  2.3
 90  1.3  1.7
 95  1.0  1.2
 97  0.7  1
 98  0.6  0.8
 99  0.4  0.6
The sampling errors provided above are measured at the 95 percent confidence level.

How to use sample errors:

For example, the estimated number of businesses with export sales in 2009 is 18 percent. This estimate is subject to a relative sampling error of approximately plus or minus 1.55. This means that 95 percent of the possible samples of the same size will produce an estimate between: 18 - 1.55 and 18 + 1.55, that is, between 16.45 and 19.55.

Sampling errors vary from estimate to estimate, and with population breakdown and size. Similar tables of sampling errors at a size and industry level for specific variables can be provided on request: email

Response rate

The Business Operations Survey 2009 had an 80 percent response rate target. The survey achieved an actual response rate of 82.4 percent, which represented 5,603 businesses.

Non-response and imputation

Unit non-responses

Unit (or complete) non-response occurs when units in the sample do not return the questionnaire. Non-response is accounted for by adjusting the weights of the responding units.

Item non-responses

Item (or partial) non-response is when units return the questionnaire, but some questions are not answered. Here, imputation was carried out on the unanswered questions, but not if the respondent answered less than 60 percent of the questionnaire.

Imputation cells and merging

Units were assigned to imputation cells for the calculation and assignment of imputation factors. Imputation cells were based on industry and rolling mean employment. If an imputation cell did not have enough respondents within the cell, then it was merged before imputation, following a list of merging preferences until a sufficient number of responses were achieved.

Imputation of numeric variables

The imputation methods used were weighted mean imputation and donor imputation.

Using the weighted mean method, a weighted mean was calculated from linked responding units for each numeric line code within each imputation cell. Non-responding units were then imputed with the weighted mean for their imputation cell. Weighted mean imputation was used to impute totals.

Donor imputation randomly selected a donor from within each imputation cell. The non-respondent was then imputed with the value(s) from the donor. Donor imputation was used to impute components and percentages so that the distribution was maintained.

Imputation of categoric questions

For categoric imputation, the method used was nearest neighbour imputation, which involved finding a donor with the most similar responses across all categoric variables.


Influential responses were excluded from the imputation factor calculations for numeric variables. Three kinds of unlinking were used:

  • automatic exclusion – due to logic, that is, unit was non-response (unit or item), specially treated or not required to answer that question
  • automatic unlinking – due to influence, that is, units with undesirable influence on imputation factor calculations for a variable were automatically detected and unlinked for that variable (with the ability to manually decline this). The checks were carried out at the imputation cell level or merged imputation cell level and were done separately for each variable
  • manual unlinking – due to influence, that is, additional units with undesirable influence on imputation factor calculations that were not automatically detected could be unlinked.

Special treatment

Special treatment candidates were identified as outliers . If a unit was specially treated then its final weight was set to 1 and it was unlinked for all imputation factor calculations. If a unit was not specially treated then its final weight was its adjusted weight.

  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+
  • Share this page to Facebook
  • Share this page to Twitter
  • Share this page to Google+