Respondent Load Strategy

Action	Implementation date
Increase the use of administrative data	Ongoing
Review the survey approval process	October 2007
Assess the respondent load impact of new surveys	March 2008
Research over-sampling of Māori, and provide alternatives	March 2008
Develop a respondent advocacy position	September 2008
Develop sample rotation for business surveys	June 2009

Strategy

Over the last five years, we have improved processes and implemented a number of measures to minimise respondent load. We made significant gains by using administrative-sourced data, such as tax records, to reduce the amount of direct surveying.

The first option to consider before introducing a new survey is whether administrative sources could be used to supplement or replace direct data collection. We should also assess other survey data that we already hold, or that others across the Official Statistics System have collected.

It is also important to ensure that when we liaise with other agencies about using their data, any extra requirements will not result in additional respondent load.

When direct surveying is the only option, there are still many methodological decisions that can help minimise respondent load. These include improved sample rotation techniques for business surveys and better data modelling.

The over-sampling of Māori is a major issue for social surveying. We must investigate alternatives to alleviate this.

To support much of this work and to keep a consistent focus on respondent load, we will create an advocacy position. This will ensure that we address respondent issues and can give objective advice to senior managers.

The initiatives associated with approving new collections include strengthening the way in which this assessment is made and shifting it to the front of the approval process. Survey approval documentation must also be explicit about the load imposed, balanced against the value of the collection.

Back to top

Implementing actions

Increase the use of administrative data

We will continue to investigate making more use of administrative data to reduce load on business respondents. However, the data from administrative sources needs to be of a high enough quality to be used as a replacement for survey data. Potential future uses of administrative data include:

• electronic transaction data being used for monthly retail trade statistics and/or for measuring tourist spending
• data from payroll systems and the Linked Employer-Employee Dataset (LEED) being used in the quarterly employment statistics
• tax data to replace/substitute income variables in some social surveys
• scanner data to supplement/replace some Household Economic Survey data
• further investigation of GST as a source of data.

Review the survey approval process

We will review the current survey approval process to ensure that respondent-load analysis is consistent and is balanced against the output value of the data produced.

Assess the respondent load impact of new surveys

Assessment of respondent load will help the Government Statistician and the Minister of Statistics make decisions that balance the load implications of new surveys with benefits. All new Statistics NZ surveys that require ministerial approval will include a respondent-load assessment that contains:

• Improved measures of the load imposed by the survey – this may include measures of time taken, obtained through cognitive and postal pre-testing, and measures of the impact on subpopulations within the sample (for example, Māori, industry sectors, etc).
• An assessment of the new survey’s impact on load target parameters – this will require a general assessment of the proposed sample and/or population in the design phase.

We will raise any issues identified during the assessment of a new survey's impact on the load targets with the Government Statistician and identify these issues to the Minister of Statistics. This improved process should ensure that respondent load is more transparent to decision makers considering a new survey.

Back to top

Research over-sampling of Māori, and provide alternatives

Currently there is no consensus among official statistics producers on the degree to which Māori need to be over-sampled in social surveys. For instance, the health sector has adopted a target of 'equal explanatory power for Māori’, in recognition that Māori statistical needs have equal status with those of the total New Zealand population.

As a first step towards facilitating a consensus, we will undertake a programme of policy and statistical research that looks at information needs for Māori, the resulting load on the population, and the extent to which alternatives to sampling (such as the use of small area techniques and administrative data) are practical. Much of this work will also be applicable to Pacific and Asian peoples.

Develop a respondent advocacy position

We will create a respondent advocacy position in 2008 as part of the Integrated Data Collection division. The role will be clearly separated from the operational processes associated with other collection and respondent management activity to ensure that objective advice can be given to the Government Statistician. The purpose of this role is to:

• ensure that Statistics NZ policies relating to respondent load are followed
• coordinate the assessment of requests for relief and respondent-load complaints
• assist the Government Statistician and senior management to address respondent-load concerns with external groups
• provide senior management with timely and objective advice on respondent-load matters, and give input into internal and external reports
• provide advice and input into respondent-load management practices within the Official Statistics System.

Develop sample rotation for business surveys

We are developing methodologies to move businesses into and out of surveys over time. Rotating businesses will not decrease the number of businesses being surveyed but would spread the load more equitably across the business community.

Due to the small size of New Zealand’s economy, it may not be possible to rotate large businesses, as generally their responses are critical for accurate and meaningful data. However, it may be possible to rotate small businesses. We will need to assess how practicable this is, so that it does not affect the quality of data.

We will develop a sample rotation methodology, to ensure that the load is spread equitably and expectations around how long businesses remain in a survey are transparent.