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+
New Zealand General Social Survey: 2012
Embargoed until 10:45am  –  13 August 2013
Data quality

Period-specific information
This section contains information that has changed since the last release.

General information
This section contains information about the data that does not change between releases.

Period-specific information

Reference period

The data collection for the New Zealand General Social Survey (NZGSS) 2012 took place over the 12 months from April 2012 to March 2013.

Accuracy of the data

The NZGSS 2012 personal questionnaire was answered by 8,462 individuals. Households were selected at random using a multistage sample design.

Response rate

The target response rate for the survey is 80 percent. The achieved response rate for 2012 was 78 percent. The response rate was calculated by dividing the weighted percentage of eligible individuals who responded by the estimated number of eligible individuals.

Imputation for item non-response

Occasionally, responses are not given to all questions and in these instances, we sometimes use imputation to ‘fill’ the gaps. Imputation is very selectively done to maintain the quality of the data. It allows more questionnaire responses to be included in the final dataset.

For NZGSS 2012, imputation was only carried out for age, personal income, and labour force status. Very little imputation was required (see the table below for a breakdown of imputations).

We used donor imputation, a hot-deck imputation method, which replaces missing values by data values from another record (a donor). This donor is identified by matching the records on other (complete) matching variables, and is the one that most closely matches the record with missing values.

Number and percentage of imputed records for NZGSS 2012
By type of imputation (Personal questionnaire)
Type of imputation Number of imputed records Percentage of imputation
Age 10 0.11
Income 267 3.02
Labour force status 27 0.30


Apart from the above imputations, we made no edits on NZGSS 2012 data.

An edit occurs when a respondent’s inconsistent answer is changed to a more accurate answer based on other information they have provided. The quality assurance process employed for this survey did not identify any responses that required editing.

Differences between NZGSS 2012 and 2010 questionnaires

The content and methodology of the 2012 survey was largely the same as the 2010 survey. However, in NZGSS 2008 and NZGSS 2010 the participation in study question was only asked of those who were dissatisfied with their knowledge and skills. In NZGSS 2012, we asked this question of all respondents.

Revisions to sample

To improve the accuracy of the NZGSS we made some changes to the NZGSS 2012 sample, and reissued weights for NZGSS 2010 and NZGSS 2008.

We found the NZGSS 2008 and NZGSS 2010 samples did not have the expected socio-economic dimensions. This was discovered through comparison with the NZ Deprivation Index and is a consequence of non-random sampling.

The effect of this on most published estimates from the NZGSS 2010 and NZGSS 2008 is not statistically significant and it is unlikely to affect aggregate released data. However, point estimates were consistently away from the expected results for lower socio-economic groups, including the Pacific population.

We re-selected a sample for the NZGSS 2012 to ensure the socio-economic bias is corrected. This means the sample has been selected from different primary sampling units (PSUs) than for NZGSS 2010 and NZGSS 2008.

We have also reissued the weights for NZGSS 2010 and NZGSS 2008 to improve small group analysis, which we see as an important aspect of the survey.

See the General information section for more information on the sample design. 

Change due to 22 February 2011 earthquake

After the February 2011 Canterbury earthquake, we updated the sampling frame of PSUs to exclude dwellings in the red zone. This was done before we selected the sample for NZGSS 2012.

Exploratory analysis for 2012

A large body of international evidence shows that self-reported life satisfaction is a credible approach to measuring the well-being of a population. Given the policy interest in maximising people’s well-being, there is a need to better understand what makes people more satisfied with their lives.

To inform this report, we used a logistic regression model to look at the aspects of life most strongly associated with overall life satisfaction. The advantage of using regression analysis is that it holds other factors constant, while looking at the association between the likelihood of feeling ‘satisfied’ or ‘very satisfied’ with life and the factor of interest.

Measures from NZGSS that we included in the regression model were: 

  • had excellent or very good general health
  • had more than enough or enough money to meet everyday needs
  • had not felt lonely in the last four weeks
  • had no major problems with the house or flat in which they live
  • undertook voluntary work
  • found it easy to express their identity in New Zealand
  • felt that police and government departments treat everyone fairly
  • felt they belonged to New Zealand
  • experienced discrimination in the last 12 months
  • felt satisfied about the conditions of facilities in their town, city, or rural area
  • had face-to-face contact with family in the last four weeks
  • felt safe walking alone in their neighbourhood at night.

The first four measures showed the strongest association with overall life satisfaction. However, we did not include all 500 NZGSS variables in these models, but looked at a select number of well-known social variables.

Once we knew what NZGSS variables we would look at for this release, we looked at which population characteristics had the strongest relationship with those variables, and therefore contained the greatest disparities among groups.

To do this we ran logistic regression models with each of the NZGSS measures as the dependent variable, and these important population group variables as the independent variables.  

  • sex
  • age
  • ethnicity
  • household income
  • labour force status
  • family type
  • migrant status
  • region
  • housing tenure
  • level of education.

There is a good deal of disparity between each of these models. No population groups had a consistent relationship across all NZGSS measures we looked at. Because of this we have presented each social outcome measure by the four or five population groups having the strongest relationships with that particular measure.

General information

Survey population

The target population for the NZGSS is the usually resident New Zealand population aged 15 years and over in private dwellings in the North Island, South Island, or Waiheke Island.

The target population includes:

  • New Zealand usual residents temporarily overseas
  • New Zealand usual residents temporarily staying elsewhere in New Zealand (including other permanent and temporary private dwellings, institutions, and non-private dwellings; and people who have no fixed abode, but are found in private dwellings on the household enumeration date)
  • people in the New Zealand armed forces if they reside in a private dwelling
  • young adults at boarding schools (young adults who fall into this category are not surveyed in the personal questionnaire, but are included as members of the household in the household questionnaire).

The target population excludes:

  • overseas visitors and international students who expect to be resident in New Zealand for less than 12 months
  • people living in non-private dwellings such as hotels, motels, boarding houses, hostels, and homes for the elderly
  • patients in hospitals, or residents of psychiatric and penal institutions
  • people living on offshore islands (excluding Waiheke Island)
  • members of the non-New Zealand armed forces and their dependants
  • non-New Zealand diplomats or diplomatic staff members and their dependants
  • New Zealand usual residents temporarily overseas who do not return within the survey period
  • New Zealand usual residents temporarily staying elsewhere in New Zealand (including other permanent and temporary private dwellings, institutions, and non-private dwellings; and people who have no fixed abode, but stay at private dwellings) who don't return within the survey period
  • New Zealand usual residents who live in remote areas that are costly or difficult to access.

Survey content

We use household and personal questionnaires to collect the data. One individual in the household completes the household questionnaire, which collects information about all the usually resident people in that household (eg family relationships and household income). One individual in the household aged 15 years or over is randomly selected to answer the personal questionnaire. Some questions in the personal questionnaire (such as the Economic Living Standard Index (ELSI) questions, including adequacy of income to meet everyday needs) are not asked of respondents who are under 18 years of age.

Interviews are conducted using computer-assisted personal interviews and last an average of 45 minutes.

Reliability of survey estimates

Two types of error are possible in estimates based on a sample survey: sampling error and non-sampling error. Sampling error can be measured and quantifies the variability that occurs by chance because a sample rather than an entire population is surveyed. Non-sampling errors are all errors that are not sampling errors. These errors are not quantifiable and include unintentional mistakes by respondents, variation in the respondent's and interviewer's interpretation of the questions asked, and errors in the recording and coding of data. Statistics New Zealand endeavours to minimise the impact of these errors by applying best survey practices and monitoring known indicators (eg, non-response).

Sampling errors are estimated using a jackknife method, which is based on the variation between estimates, and on taking 100 mutually exclusive subsamples from the whole sample. Sampling errors are quoted at the 95 percent confidence level. For example, if the estimated total number of people is 1,575,200, and the estimate is subject to a sampling error of plus or minus 35,500, or 2.3 percent (measured at the 95 percent confidence level), that shows there is a 95 percent chance the true total number of people lies between 1,539,700 and 1,610,700.

Smaller estimates, such as the total number of Pacific people (191,000), are subject to larger sampling errors than larger estimates. This estimate is subject to a sampling error of plus or minus 21,300, or 11.2 percent (measured at the 95 percent confidence level).

High-level checks of the ethnic groups indicated the samples are broadly representative of the population. However conclusions about groups with small population size, such as Pacific people, can potentially be vulnerable to unmeasured differences between the survey participants and the population.

An output with a relative sampling error of 30 percent to 49.9 percent should be viewed with caution (flagged in tables by an asterisk*), and an error of 50 percent or more should be considered unreliable (flagged by **).

Confidentiality and suppression

Tables with very few contributors are suppressed (‘S’). These cells have an estimated population of less than 1,000 and are deemed to be unreliable and a risk to respondents’ confidentiality.

Rounding and percentages

All percentages used in the text are calculated from weighted data and then rounded to two significant figures. The percentages in the tables are rounded to one decimal place. To improve the readability of the data, the calculation of percentages excludes residual categories (eg ‘don't know’ and ‘refused’) in the population base from which percentages are calculated.

Accuracy of the data

Sample design information

The NZGSS uses a three-stage sample selection method, similar to other Statistics New Zealand’s household surveys.

The first stage of the selection consists of selecting a total of 1,200 PSUs from the Household Survey Frame (HSF).  The HSF is the standard sampling frame Statistics NZ uses to select samples and to manage overlap control for all its household surveys. The HSF lists PSUs with attributes determined by data from the census. PSUs are then assigned to standard strata based on these attributes. 

The second stage of sample selection consists of selecting eligible dwellings within the selected PSUs.  In the third stage, we select one eligible individual within each selected dwelling. The eligible individual is chosen at random from the list of all eligible individuals in the dwelling.

The NZGSS is designed to provide estimates at a national level.


The survey has two sets of weights attached, one for the household and one for the person. The household weight is used to describe the attributes of a household; for example, how many households have dependent children who live outside that household. The person weight is used to describe the attributes of a person; for example, how many people are ‘very satisfied’ with their life overall.

More information

See New Zealand General Social Survey for more information.


While all care and diligence has been used in processing, analysing, and extracting data and information in this publication, Statistics NZ gives no warranty it is error-free and will not be liable for any loss or damage suffered by the use directly, or indirectly, of the information in this publication.


Our information releases are delivered electronically by third parties. Delivery may be delayed by circumstances outside our control. Statistics NZ does not accept responsibility for any such delay. 

Crown copyright©

Creative Commons logo.
This work is licensed under the Creative Commons Attribution 3.0 New Zealand licence. You are free to copy, distribute, and adapt the work, as long as you attribute the work to Statistics NZ and abide by the other licence terms. Please note you may not use any departmental or governmental emblem, logo, or coat of arms in any way that infringes any provision of the Flags, Emblems, and Names Protection Act 1981. Use the wording 'Statistics New Zealand' in your attribution, not the Statistics NZ logo. 

  • 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+