Australia

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE CONFERENCE OF EUROPEAN STATISTICIANS Expert Meeting on Statistical Data Collection and Sources 22-24 May 2024, Geneva, Switzerland

19 April 2024

Australia’s Data Collection Modernisation

Jodie Stevenson (Australian Bureau of Statistics, Australia) [email protected] Abstract The presentation will outline Australia’s data collection modernisation program and the shift towards digital first survey collections. Australia’s data collection has evolved over many years with our primary mode being face to face interviewing. Like other jurisdictions this has resulted in declining response rates, challenges in meeting community expectations and increased cost. Australia is resetting our data collection capability through:

• Online collections that are accessible, secure and meet community expectations • Greater use of administrative data and initiatives such as our accounting software project • A focus on our workforce, providing greater certainty on our interviewer availability, considering the

geographic coverage for interviews • A contemporary contact centre, we are upgrading our telephone contact centre to support an increase in

calls and enable mode switching between face to face, online and telephone collections. The use of mixed modes and flexibility in mode switching is an area Australia is exploring. Australia’s data collection modernisation journey is underway. This presents an opportunity to address declining response rates, improve provider experience and produce quality statistics. However it does require new capability and new approaches and we will need to adapt our strategy to respond to new challenges over time.

AUSTRALIAN BUREAU OF STATISTICS Informing Australia’s important decisionsAUSTRALIAN BUREAU OF STATISTICS

Informing Australia’s important decisions

Towards a data ethics program for the ABS: Considering privacy, ethics and trust for our innovative data uses

Joanne Hillermann, Director, Privacy

Australian Bureau of Statistics

Data ethics are important to the ABS

221/03/20 24

Challenges

• Increasingly innovative data uses

• Cumulative impacts on privacy of large volumes of data and innovations

• Increasing use of AI

• Public expectations about data use

• Increasing security and cyber threats

• Resourcing pressures – need for more transparent data ethics processes and understanding citizen views about data uses

Issues

• Data breaches

• Other government failures negatively impacting public perceptions of data uses

• Lack of social acceptance for some innovative data uses

• Lack of transparency and clear public communications about some initiatives

Context: Privacy, ethics and trust in Australia

3

Privacy Clear legislative requirements

ABS Census and Statistics Act 1905

Privacy Act 1988

Australian Government Agencies Privacy Code

Ethics Principles and Guidelines

Artificial Intelligence Data Ethics Principles

National Statement on Ethics – guides human research

Trust Guides to actions and behaviours

Building Trust in the ABS and Data Use Strategy

Building Trust in the ABS and Data Use Strategy

4

Understand drivers of

public trust

Work with a whole-of-

government perspective

Take meaningful actions to

build trust

Timeline

ABS

initiatives

Whole of

Australian

Government

Initiatives

2015

Commenced

development

of large-scale

integrated

data assets

Australia's

AI Ethics

Framework

2019

2017

First Privacy

Impact

Assessment

for Integrated

data - PLIDA

2021

Integrated

people and

business data

– PLIDA and

BLADE

2023

First Private

sector data

integrated in

PLIDA

2024

Next

generation

integration -

NDDA/ANDII

operational

Australian Data

and Digital

Government

Strategy –

including Data

Ethics Framework

from July 2024

2023

Australian

Data

Strategy –

NDDA/ANDII

2021

PLIDA – Person Level Integrated Data Asset

BLADE – Business Level Analytical Data Environment

NDDA/ANDII – National Disability Data Asset/Australian National Data Integration Infrastructure

2019

Building Trust

in ABS and

Data Use

Strategy

Review of ABS privacy, ethics and trust practices

6

ABS has strong and mature privacy and trust practices, but existing ethics practices were only low to medium in maturity (2023).

Good data ethics practices:

▪ Foundationally concerned with the question of 'can vs should' for our data uses

▪ Aid accountability

▪ Foster responsible innovation

▪ Include both top-down (strategic) and bottom-up (operational) ethics considerations

▪ Allow levels of escalation for issues

▪ Specific to the ABS' needs

▪ Align with (but do not duplicate) existing governance/ ethics processes

▪ Effectively navigate the boundary between ethics and privacy

Case Study 1: Use of private sector electricity data for occupancy measures in the Census

7

The challenge:

Use electricity data (Smartmeter data) at the dwelling level to help identify households that are occupied on Census night

Outcome:

Significant privacy concerns were raised and Smartmeter data was not used at the dwelling level in the 2021 Census

Instead, it was used to derive a predicted count of occupied dwellings at an area level as an input to occupancy models

Key lesson:

Engaging early, with clear information about privacy impact and benefit, could have

reduced privacy concerns

Case Study 2: Ethics oversight capability for National Disability Data Asset (NDDA)

8

The challenge:

Develop the NDDA as a nation-wide, multi-sector linked data asset to help to better understand needs, services delivered and outcomes for people with disability.

Share data using new legislation (Data Availability and Transparency Act 2022) in accordance with the Australian Government Data Sharing Principles.

Establish controls to ensure the data is used appropriately and in consultation with the disability community.

Outcome:

Disability-informed Ethics Oversight Panel to consider projects with the disability community.

NDDA Charter sets out acceptable and unacceptable uses of the NDDA.

Co-governed by governments and the disability community via the NDDA Council. The Council will ensure the NDDA is used correctly and only for purposes that align with the disability community’s expectations and the Charter.

Key lesson:

Controls to ensure appropriate uses of data that are in line with population of interest views, are also desirable governance mechanisms for other integrated data assets.

Ethics work program for ABS from 2024-25

9

• Establish Data Ethics Advisory Committee

• Develop Data Ethics Principles and FrameworkStrategic

• Establish role of Chief Data Ethics Officer

• Develop internal data ethics policy

• Develop internal guidance and training materials

• Combine privacy and ethics assessments/ tools

• Develop new ways to assess citizen views about data uses

• Improve communications (with community)

• Learn from other NSOs

Operational

ABS is keen to learn from others

10

Questions for discussion about establishing a data ethics program:

▪ What are the highest priorities to do first?

▪ What mix of skills are required for the data ethics advisory committee?

▪ Does the committee assess individual projects or broad types of data

use?

▪ Possible scope for Data Ethics Framework – national or NSO, relating to

people data or broader?

▪ Accountability for ethics – Chief Ethics Officer role?

▪ What are the best ways of assessing community comfort with innovative

data uses?

▪ What hits and misses have you had with communications with the public?

AUSTRALIAN BUREAU OF STATISTICS Informing Australia’s important decisions

Thank you

The Acknowledgement of Country supplementary slides are applied to this PowerPoint on pages 2 and 3. You can choose from a ‘still’ or a ‘two-minute video. The two- minute video is most suited for presentations over half an hour. To play the video, please ensure your laptop is connected to the internet.

Note: An ABS National Acknowledgement of Country static slide is provided on this slide. However, when you are delivering a verbal speech always refer to the Traditional Custodians to find the Traditional Owners and Country you need to pay respects to and acknowledge.

‘Our Story. Our Future.’ was created by proud Wiradjuri, Wotjobaluk, Yuin and Gumbaynggirr artist Luke Penrith and Maluililgal people, Badu Island artist Naseli Tamwoy.

Limitations of SEEA waste accounts: conceptual, data collection and experiences from policy use

Peter Meadows, Centre of Environmental and Satellite Accounts, ABS

Waste in Australia (2018-19 waste accounts)

Total waste: 75.8 Megatonnes Total waste: 12.6 million African elephants

Limited time series available

Policy views

ABS view – Are Waste accounts suitable for domestic reporting? – Which products? – User and provider-based design principles – Should NSO’s be responsible for waste data collection?

Emerging policy views – Shift to circular economy

– Can waste accounting (and SEEA) adapt to report on new indicators? – Gaps

– Behavioral information – Return on recycling to the economy (Value add, life cycle benefits) – Impact on ecosystems

What defines waste & challenges?

– Traditional waste accounts have a limited view on destination of waste materials – Need for a ‘recycled products’ account

covering supply, use and inventories. – Intent to recycling does not match actual

recycling (could be useful metric but not used) – Measurement issues

– Surveys of waste generated cover ‘intended recycled materials and waste materials disposed

– Waste facility sites will have different stages of processing and measurement

Measured components

Australia’s response to data collection issues

– National standard for waste and resource recovery data and reporting – DCCEEW

– Defines waste products, collection methods and estimation needed for weight

Policy response

– Alignment with circular economy – Recycling Traceability – Product Stewardship – National Packaging Laws

– Alignment with Nature Positive Policy – Building a circular economy and reducing the impact of plastics on nature.

(https://www.dcceew.gov.au/about/news/stay-informed/communiques#environment-ministers-meeting)

= desire for a comprehensive circular economy (material flow account) – Addresses traceability of recycled materials (if applied correctly – PSUT)

Next steps?

– Circular economy? – Improvements on CSIRO’s model :

https://publications.csiro.au/publications/publication/PIcsiro:EP197565

– Physical data assets? – Integrate existing physical data (Agriculture, energy, construction, water, waste, trade).

– Waiting on government to decide.

‘Our Story. Our Future.’ was created by proud Wiradjuri, Wotjobaluk, Yuin and Gumbaynggirr artist Luke Penrith and Maluililgal people, Badu Island artist Naseli Tamwoy.

Working with blue carbon ecosystem accounts: value of coastal ecosystems in alleviating impacts of climate change Peter Meadows, Centre of Environmental and Satellite Accounts, ABS Based on work by Dastagir Mir, Centre of Environmental and Satellite Accounts , ABS (research leader)

The Challenge

– Coastal flooding resulting from storm surge is expected to be exacerbated with the increasing frequency and intensity of storms over this century.

– This poses a risk to coastal communities, infrastructure and economically valuable land within the coastal zone.

– Recent studies have found that coastal ecosystems like mangroves and saltmarshes could decrease flood depths and save millions of dollars in flood damages.

Designing Environmental Ecosystem Accounts from a user perspective

Some of the questions Environmental Ecosystem accounts can address are:

– How much of the ecosystem resources are left?

– Is the ecosystem health improving?

– What services do ecosystem provide?

– What are the impacts of development on coastal resources?

– What are the measures that can lead to a balanced discussion on trade offs?

– Can use these questions to help solve many data problems: climate change, economic growth through ecosystems, financing for nature.

Australia’s first National Ocean Account

Partnership between ABS and Environment ministry

Climate mitigation and resilience focus.

Included only ‘blue carbon’ ecosystems –

Mangroves Saltmarsh Seagrass

Coastal Protection Ecosystem Service

– Answering the challenge: What is the value and protection that coastal ecosystems provide for us?

– The National Ocean Account looked at measuring coastal protection services offered by mangroves and saltmarsh ecosystems.

– A coastal protection model was developed by the ABS to measure coastal protection ecosystem service.

– The aim of the model was to identify: – length of Australia’s coastline, – number of people and dwellings – that were likely to be offered coastal protection from storm surge through these ecosystems.

– Model used represents a first attempt to consolidate the available information on coastal protection services and is limited by data sources available at the national scale.

Coastal Protection Model

– Assumptions

– Saltmarsh and mangroves that met the following requirements were assumed to provide coastal protection services:

– Located within 200m of the coastline, and – had a defined belt width (perpendicular to the coast) of at least 40m for saltmarsh and at least

90m for mangroves.

– The protection afforded by these ecosystems was considered to be up to 1km from the coastline.

Coastal Protection Model

– Step 1: Identify saltmarsh and mangroves that are within 200m of the coastline.

Coastal Protection Model

– Step 2: Identify saltmarsh and mangroves that have a defined belt width (perpendicular to the coastline) of at least 40m for saltmarsh and at least 90m for mangroves.

Coastal Protection Model

– Step 3: Identify population and dwellings offered protection by saltmarsh and mangroves.

Coastal Protection Model

– Step 4: Identify length of coastline offered protection by saltmarsh and mangroves.

Results – social

– In 2021, 150 thousand dwellings were offered coastal protection by either saltmarsh, mangroves, or both.

– In 2021, 280 thousand people received coastal protection services from either saltmarsh, mangroves, or both.

Results – economic

Ecosystem Asset value contribution – Investment cost (2021) – $41.6 billion (USD$29.4 billion) for saltmarsh. – $196 billion (USD$139 billion) for mangroves.

Proxy Ecosystem Service value – Replacement cost per annum (2021 currency) – $0.4bn to $1.7bn (USD$0.3bn to $1.2bn) saltmarsh. – $2.0bn to $7.8bn (USD$1.4bn to $5.5bn) mangroves.

– Replacement cost method (replace with seawalls) – Or is it $72k (USD$51k) per hectare (see picture)

Lessons for adapting to climate change

– 40m of saltmarsh or 90m of mangroves can attenuate wave action and reduce coastal erosion (as well as other biodiversity and species nursery benefits)

– Alternative is to move settlements (very expensive) or build seawalls (expensive). We are looking closely at Indonesia and Philippines as comparisons.

– We still need to develop estimates where the area of ecosystems protects agricultural or other economic production activities from climate impacts.

– There is interest in incorporating these estimates or methods into disaster management.

Lessons from Australia

– For a first account, don’t be afraid to narrow the range of accounts needed

– Work with relevant agencies and peak bodies to clarify what needs to change and how they use data

– Work towards as close a solution as possible, but staying close to the SEEA standard

– Enjoy the discussions afterwards (our oceans accounts have won 4 awards so far…)

Submitted by the expert from Australia Informal document GRVA-18-11 18th GRVA, 22-26 January 2024 Provisional agenda item 6(b)

Proposal for the 05 series of amendments to UN Regulation No. 79 – Steering equipment

This document proposes the 05 series of amendments to United Nations Regulation No. 79 – Steering Equipment. The proposed amendments intend to facilitate a consistent approach across all Contracting Parties for Emergency Lane Keeping Systems (ELKS) covering vehicles of category M1 and N1. Currently the European Union mandates ELKS through EU Regulation 2021/646, with the proposed amendments closely aligning with its’ requirements. The modifications to the current text of the Regulation are marked in bold or strikethrough characters.

I. Proposal Amendments to UN Regulation 79

Paragraph 2.3.4.2., amend to read:

2.3.4.2. “Corrective Steering Function (CSF)" means a control function within an electronic control system whereby, for a limited duration, changes to the steering angle of one or more wheels and/or braking of individual wheels may result from the automatic evaluation of signals initiated on-board the vehicle, in order:

(a) To compensate a sudden, unexpected change in the side force of the vehicle, or;

(b) To improve the vehicle stability (e.g. side wind, differing adhesion road conditions "μ-split"), or;

(c) To correct lane departure. (e.g. to avoid crossing lane markings, leaving the road).

Insert a new paragraph 2.10., to read:

2.10 “Distance To Lane Marking (DTLM)” means the remaining lateral distance (perpendicular to the lane marking) between the inner side of the lane marking and most outer edge of the tyre before the subject vehicle crosses the inner side of the lane marking.

Paragraph 5.1.6.1.1., amend to read:

5.1.6.1.1. Every CSF intervention shall immediately be indicated to the driver by an optical warning signal which is displayed for at least 1 s or as long as the intervention exists, whichever is longer.

The optical signal may be the flashing of the failure warning signal specified in paragraph 5.1.6.1.7.1. When a flashing mode is used, a lighting phase shall be visible at the end of the intervention or later.

In the case of a CSF intervention which is controlled by an Electronic Stability Control (ESC) or a Vehicle Stability Function as specified in the relevant UN Regulation (i.e. UN Regulations Nos. 13, 13-H or 140), the ESC flashing telltale indicating the interventions of ESC may be used, as long as the intervention exists, as an alternative to the optical warning signal specified above.

Paragraph 5.1.6.1.3., amend to read:

5.1.6.1.3. The steering control effort necessary to override the directional control provided by the system shall not exceed 50 N in the whole range of CSF operations. Significant loss of steering support once overridden must not happen suddenly.

For CSF systems which do not act on the steering itself (e.g. differential braking type CSF), the steering input for overriding the CSF must not exceed 25 degrees.

Insert a new paragraph 5.1.6.1.5., to read:

5.1.6.1.5. The CSF must be active at least between 65 km/h and 130 km/h (or the maximum vehicle speed if it is below 130 km/h) and at all vehicle load conditions, unless deactivated as per paragraph 5.1.6.1.8.

Insert a new paragraph 5.1.6.1.6., to read:

5.1.6.1.6. Lane keep

The requirements of this paragraph only apply to CSF that meet the definition of subparagraph (c) of paragraph 2.3.4.2.

In the absence of conditions leading to deactivation or suppression of the system, the CSF must be able to prevent lane departure to a maximum DTLM of -0.3m, i.e. the outer edge of the tyre must not cross the inner side of the visible lane marking by any more than 0.3 m, for each scenario shown in the following table, as well as:

(a) for lateral departure velocities in the range of the 0.2 m/s to 0.5 m/s for vehicle speeds up 100 km/h and for lateral departure velocities in the range of 0.2 m/s to 0.3 m/s for vehicle speeds greater than 100 km/h and up to 130 km/h (or the maximum vehicle speed if it is below 130 km/h);

(b) on straight, flat and dry roads;

(c) for both solid line and dashed lane markings in line with one of those described in Annex 3 (Visible lane marking identification) of UN Regulation No. 130;

(d) with the lane markings being in good condition;

(e) in all illumination conditions without blinding of the sensors (e.g. direct blinding sunlight) and with activated passing-beam (dipped-beam) headlamps if necessary;

(f) in absence of weather conditions affecting the dynamic performance of the vehicle (e.g. no storm, not below 0 °C) or the visibility of lane markings (e.g. no fog).

No. Scenario description

1. Both solid and dashed lines – departure to the right side of the vehicle

2.

Both solid and dashed lines – departure to the left side of the vehicle

It is recognised that the performances required for the scenarios in this table may not be fully achieved in other conditions than those listed above. However, the system must not unreasonably switch the control strategy in these other conditions. This must be demonstrated in accordance with the testing requirements.

The lane keep capability must be tested in accordance with the relevant vehicle test(s) specified in Annex 8 to this Regulation.

Insert a new paragraph 5.1.6.1.7 (and subparagraphs), to read:

5.1.6.1.7. CSF failure warning

A warning must be provided when there is a failure detected in the CSF that prevents the system from fulfilling the requirements of this Regulation.

5.1.6.1.7.1. The failure warning must be a constant optical warning signal.

5.1.6.1.7.2. There must not be an appreciable time interval between each CSF self-check (an integrated function that checks for a system failure on a continuous basis at least while the system is active), and subsequently there must not be a delay in illuminating the warning signal, in the case of an electrically detectable failure.

5.1.6.1.7.3. Upon detection of any non-electrical failure condition (e.g. sensor misalignment), the warning signal as defined in paragraph 5.1.6.1.7.1. must be activated.

5.1.6.1.7.4. If the vehicle is equipped with a means to deactivate the CSF a warning must be given when the system is deactivated according to paragraph 5.1.6.1.8. This must be a constant optical warning signal. The failure warning signal specified in paragraph 5.1.6.1.7.1. may be used for this purpose.

Insert a new paragraph 5.1.6.1.8. (and subparagraphs), to read:

5.1.6.1.8. CSF deactivation

5.1.6.1.8.1. Manual deactivation

When a vehicle is equipped with a means to manually deactivate the CSF function, either partially or fully, the following conditions must apply as appropriate:

5.1.6.1.8.1.1. The full CSF function must be automatically and fully reinstated upon each activation of the vehicle master control switch.

5.1.6.1.8.1.2. The manual deactivation of the full CSF must not be possible with less than two deliberate actions, e.g. press and hold on a button, or select and confirm on menu option.

5.1.6.1.8.1.3. The manual deactivation capability must be tested in accordance with the relevant vehicle test(s) specified in Annex 8 to this Regulation.

5.1.6.1.8.2. Automatic deactivation

If the vehicle is equipped with a means to automatically deactivate the CSF function, either partially or fully, for instance in situations such as off-road use, being towed, a trailer being hitched to the vehicle or the electronic stability control (ESC) being deactivated, the following conditions must apply as appropriate:

5.1.6.1.8.2.1. For testing, the vehicle manufacturer must provide a list of situations and corresponding criteria where the CSF function is automatically deactivated which must be annexed to the test report.

5.1.6.1.8.2.2. The CSF function must be automatically and fully reactivated as soon as the conditions that led to the automatic deactivation are not present anymore.

5.1.6.1.8.3. A constant optical warning signal must inform the driver that the CSF function has been deactivated. The failure warning signal specified in paragraph 5.1.6.1.7.1. above may be used for this purpose.

Insert a new paragraph 5.1.6.1.9. (and subparagraphs), to read:

5.1.6.1.9. Automatic suppression

5.1.6.1.9.1 For driver intended manoeuvres

For testing, the manufacturer must provide a documentation package which gives access to the basic design and logic of the system for detection of likely driver intended manoeuvres and automatic suppression of the CSF. This package must include a list of parameters detected and a basic description of the method used to decide that the system should be suppressed, including limit values where possible. For the CSF, the Technical Service must assess the documentation package to show that driver unintentional manoeuvres, within the scope of the lane keep test parameters (in particular lateral departure velocity), will not result in automatic suppression of the system.

5.1.6.1.9.2 Automatic suppression of the CSF is also permitted in situations when other driver assist or automated steering functions, (i.e. ACSF, ESF or automated lane keeping), are controlling the lateral movement of the vehicle or other safety related functions (i.e. that is capable of changing the dynamic behaviour of the vehicle such as AEBS, ESC, etc.) are intervening. These situations must be declared by the manufacturer.

Insert a new paragraph 12.4. (and subparagraphs), to read:

12.4. Transitional Provisions applicable to the 05 series of amendments:

12.4.1. As from the official date of entry into force of the 05 series of amendments, no Contracting Party applying this Regulation shall refuse to grant or refuse to accept type approvals under this Regulation as amended by the 05 series of amendments.

12.4.2. As from [1 XX 202X], Contracting Parties applying this Regulation shall not be obliged to accept type approvals to the preceding series of amendments, first issued on or after [1 XX 202X].

12.4.3. Until [1 XX 202Y], Contracting Parties applying this Regulation shall accept type approvals to the preceding series of amendments, first issued before [1 XX 202X].

12.4.4. As from [1 XX 202Y], Contracting Parties applying this Regulation shall not be obliged to accept type approvals issued to the preceding series of amendments to this Regulation.

12.4.5. Notwithstanding paragraph 12.4.2. and 12.4.4., Contracting Parties applying this Regulation shall continue to accept UN type approvals issued according to a preceding series of amendments to this Regulation, for vehicles which are not affected by the provisions introduced with the 05 series of amendments.

Paragraphs 12.4 and 12.4.1. (former), re-number as paragraphs 12.5. and 12.5.1., to read:

12.4. General transitional provisions

12.4.1. Contracting Parties applying this Regulation shall not refuse to grant UN Type approvals according to any preceding series of amendments to this regulation or extensions thereof.

Annex 8, paragraph 2., amend to read:

2. Testing conditions

The tests shall be performed on a flat, dry asphalt or concrete surface affording good adhesion, which may not contain any irregularities (e.g. large dips or cracks, manhole covers or reflective studs) within a lateral distance of 3.0 m to either side of the centre of the test lane and with a longitudinal distance of 30 m ahead of the subject vehicle from the point after the test is complete. The ambient temperature shall be between 0 °C and 45 °C.

The ambient illumination conditions shall be of at least 2,000 lux without blinding of the sensors (e.g. direct blinding sunlight) and with activated low beam head lamps if necessary.

At the request of the manufacturer and with the agreement of the Approved Testing Facility tests may be conducted under deviating test conditions (suboptimal conditions, e.g. on a not dry surface; below the specified minimum ambient temperature), whilst the performance requirements are still to be met.

Annex 8, insert a new paragraph 2.3.3., to read:

2.3.3 Where the CSF is equipped with a user-adjustable timing threshold, the test specified in paragraph 3.1.3. must be performed with the timing threshold set at its latest setting for system intervention. No alteration must be made once the test procedure has begun.

Annex 8, paragraph 3.1.1.1., amend to read:

3.1.1.1. The vehicle shall be driven with an activated CSF on a road with lane markings on each side of the lane. In case of a CSF whose interventions are solely based on the evaluation of the presence and location of lane boundaries, the vehicle shall be driven on a road delimited by the boundaries as declared by the manufacturer (e.g. road edge).

If requested by the vehicle manufacturer the vehicle can be driven to calibrate the sensor system up to a maximum of 100 km on a mixture of urban and rural roads with other traffic and roadside furniture.

The test conditions and the vehicle test speed shall be within the operating range of the system.

During the test, the duration of the CSF interventions and of the optical and acoustic or haptic warning signal, as relevant, shall be recorded.

In the case of paragraph 5.1.6.1.2.1. of this Regulation, the vehicle shall be driven such that it attempts to leave the lane and causes CSF intervention to be maintained for a period longer than 10s (for M1, N1) or 30s (for M2, M3, N2, N3). If such a test cannot be practically achieved due to e.g. the limitations of the test facilities, with the consent of the type approval authority this requirement may be fulfilled through the use of documentation.

The test requirements are fulfilled if:

(a) The acoustic or haptic warning, as relevant, is provided no later than 10s (for M1, N1) or 30s (for M2, M3, N2, N3) after the beginning of the intervention.

In the case of paragraph 5.1.6.1.2.2. of this Regulation, the vehicle shall be driven such that it attempts to leave the lane and causes at least three interventions of the system within a rolling interval of 180 s.

The test requirements are fulfilled if:

(a) An optical warning signal is provided for each intervention, as long as the intervention exists, and

(b) An acoustic or haptic warning signal, as relevant, is provided at the second and third intervention

and

(c) The acoustic or haptic warning signal, as relevant, at the third intervention is at least 10 s longer than the one at the second intervention.

Annex 8, paragraph 3.1.2.2., amend to read:

3.1.2.2. The test requirements are fulfilled if:

(a) The force applied by the driver on the steering control to override the intervention does not exceed 50 N.

(b) There is no sudden loss of significant steering support once CSF is overridden.

(c) For CSF that do not act on the steering itself (e.g. differential braking type CSF), the steering input for overriding the CSF does not exceed 25 degrees.

Annex 8, insert a new paragraph 3.1.3. (and subparagraphs), to read:

3.1.3. Lane keep test

3.1.3.1. The CSF must be tested for test scenarios No 1 and No 2 described in paragraph 5.1.6.1.6.

3.1.3.1.1. Tests for all scenarios must be performed with lateral velocities of 0.2 m/s and 0.5 m/s.

3.1.3.1.2. A test path must be driven which consists of an initial straight path parallel to the lane marking being tested, followed by a fixed radius curve to apply a known lateral velocity and yaw to the subject vehicle, followed again by a straight path without any force applied on the steering control (e.g. by removing the hands from the steering control).

3.1.3.1.3. The subject vehicle speed during the test up to the point of system intervention must be 67 km/h +/– 1 km/h.

The curve of fixed radius driven to apply the lateral velocity required must have a radius 1,200 m or more.

The lateral velocity required must be achieved to a tolerance of +/– 0.05 m/s.

The vehicle manufacturer must provide information describing the radius of the curve to be driven and the location when the closed loop path and/or speed control must be ended so as to ensure a free drifting in order not to interfere an automatic suppression according to paragraph 5.1.6.1.9.1.

3.1.3.2. The test requirements are fulfilled if the subject vehicle does not cross the lane marking by greater than a DTLM of -0.3 m.

3.1.3.3. In addition, the vehicle manufacturer must demonstrate to the satisfaction of the Technical Service that the requirements for the whole speed range and lateral departure velocity range are fulfilled. This may be achieved on the basis of appropriate documentation appended to the test report.

II. Justification 1. Australia is currently proposing to mandate the fitment of ELKS for vehicle categories equivalent to M1 and N1. A feasible option is to adopt Corrective Steering Function (CSF) requirements within UN Regulation No. 79 to cover Corrective Directional Control Function requirements as specified in EU Regulation 2021/646.

2. Australia does not consider the current requirements for CSF in UN Regulation No. 79 to be adequate and prefers adopting similar requirements to those in EU Regulation 2021/646 (Emergency Lane Keeping Systems). This is because the current UN Regulation does not provide sufficient stringency in the performance requirements for light vehicles, such as automatically and fully reinstating the CSF upon each activation of the vehicle master control switch etc.

3. Therefore, the expert from Australia proposes to amend the performance requirements of UN Regulation No. 79 to align with the requirements in EU Regulation 2021/646 for vehicles of category M1 and N1.

4. These amendments will provide a consistent approach across all Contracting Parties for the performance requirements and regulation of ELKS in vehicles of category M1 and N1.

5. Research published by the Monash University Accident Research Centre (MUARC), reported that for light vehicles, 11 per cent of casualty crashes and 42 per cent of fatal crashes in Australia (on sealed roads with speed limits of ≥ 70 km/h) between 2013 to 2019 involved unintentional lane departure. Unintentional lane departure crashes included single-vehicle and multi-vehicle head-on and sideswipe crashes. Unintentional lane departure crashes represented 55 per cent of all road fatalities involving light vehicles, with this number increasing to 72 per cent at highway speeds of >=100 km/h (Stuart et al. 2021).

6. The MUARC research demonstrated that ELKS is effective in reducing road trauma resulting from crashes involving unintentional lane departures. Results estimated a 9.09 per cent saving in total annul fatal crashes when 100 per cent of the light vehicle fleet is fitted with a lane keep assist system. This corresponded to a 11.9 per cent saving in total annual fatalities (Stuart et al. 2021).

III. References Stuart Newstead, Linda Watson, Laurie Budd. 2021. The Potential Benefits of Lane Keep

Assist Systems in Australian Light Vehicles. Melbourne: Monash University Accident Research Centre (MUARC). Accessed November 1, 2023.

Improving survey response by developing an effective contact and cooperation strategy

Yvette Kezilas, Conor Kelly & Leyton Wood

With support from the Project Team: Annelie de Villiers Bryony Cresswell Katherine Birrer

Improving our contact and cooperation strategy

29/05/20233

Three approaches proposed to support the ABS’ contact and cooperation strategy:

1. Tailored approaches 2. Survey incentives 3. Household induction

4

Common Barriers

• Literacy/language

• Fear/lack of trust

• Shared accommodation/ sharehouses

• Online challenges

– Access/ability

– Mobile only access

• Self-deselection/ relevance

• Mail/address issues

Migrant Population

Indigenous Population

Northern Territory

Rural Areas

Young People

Low SES

Aged Pensioners

(65+)

Other Pensioners (eg disability,

carers)

1. Tailored approaches

Best practice engagement strategies

• Simple language

• Translation/audio services

• Clear survey scope

• Offer mixed-modes

• Mobile accessibility

• Incentives

1. Tailored approaches

29/05/20235

Design principles: • Clear and transparent call to

action • Simple language • Clear survey scope • Easy to follow instructions • Offering multiple modes

Literature suggests financial incentives have a positive effect on survey response

Most research has looked at incentives for voluntary surveys

Cognitive testing conducted to explore attitudes towards incentives for compulsory ABS surveys

2. Survey incentives

29/05/20236

Incentives not expected for compulsory surveys (if less than 30 mins)

Incentives expected only when respondent burden is high (1hr+)

29/05/20237

2. Survey incentives

3. Household survey induction model

29/05/20238

Household selected

Survey- specific

information provided

Survey completed

Contact details

registered Current state

Household selected

Survey- agnostic

information provided

Survey- specific

information & Survey

completed

Contact details

registered Future state?

Participants were comfortable not being told upfront what the survey is specifically about

...BUT

…Participants were not comfortable with being told the time commitment for the survey only once they had entered the webform

29/05/20239

3. Household survey induction model

Thank you for your time Questions?

Keep in touch: [email protected]

IMPROVING SURVEY RESPONSE BY

DEVELOPING AN EFFECTIVE CONTACT

AND COOPERATION STRATEGY

Yvette Kezilas, Conor Kelly, Leyton Wood

Data Collection Design Centre

Australian Bureau of Statistics

June 2023

PAPER PRESENTED AT

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

Expert Meeting on Statistical Data Collection – Rethinking Data Collection

12-14 June 2023

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

2 | P a g e

1. ABSTRACT

Collection of high-quality survey data requires effective contact strategies with people to

communicate why they have been selected in a survey, and to gain their cooperation to participate.

With statistical organisations around the world facing cost pressures and declining response rates,

having effective contact and cooperation strategies in data collection has never been more

important to maintain statistical quality and trust.

A strategic priority for the Australian Bureau of Statistics (ABS) is the production of high quality

statistics. This paper explores the effectiveness of three approaches designed to support and

advance the ABS’s contact and cooperation strategy for household and person surveys.

1. Developing a tailored approach strategy to engage key demographic groups. These include

those with a high non-response propensity, or groups that are critical to have represented in

the data based on the survey topic (e.g. those with low or high socio-economic status for

income and expenditure surveys).

2. Offering financial or non-financial incentives for completing a government survey.

3. Collecting contact details through a household survey induction model, whereby the

respondent is initially introduced to the survey program, rather than to a specific survey.

We designed the three approaches to work either independently, or in combination, depending on

the collection requirements.

For each approach, we carried out extensive literature reviews and employed human centred design

consultation methods. This was followed by prototype development and cognitive testing with

members of the public.

The opportunities and limitations of the approaches are outlined. We also suggest how and when

they should be deployed to a statistical agency’s contact and cooperation strategy, to improve

survey response.

2. INTRODUCTION

Household statistics produced by the Australian Bureau of Statistics (ABS) play a critical role in

informing the nation’s most important decisions. To produce these statistics, the ABS is reliant on

collecting survey data from households across the country. Collecting both high quality and timely

survey data requires both making effective contact with those who live at the selected dwellings and

obtaining their cooperation to start and complete an ABS survey.

For the ABS, making effective contact with respondents is challenging for various reasons.

Household survey samples are selected from an address-based register, with no additional contact

information (e.g. name or phone number) available (ABS, 2020). This means that the ABS is reliant

on postal mail and field interviewers to approach selected households and engage them in the

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

3 | P a g e

survey process. Without the ability to address the approach letters to the name of the resident/s,

often respondents will throw the letters out assuming them to be junk mail. Australia is a vast

country, making rural or remote areas, which are disproportionately represented by key

demographic groups, difficult and costly to survey in person.

Once contact has been made, seeking willing cooperation to complete an ABS survey is also

challenging. While most ABS surveys are compulsory, there is limited public awareness of the ABS’

collection of household data outside of the Census. This limited awareness, coupled with the rise in

exposure to scams (ABS, 2023), means that our survey requests are increasingly being dismissed as

fraudulent. Furthermore, many of our household surveys are time-consuming to complete (e.g. over

1 hour), cognitively burdensome (e.g. questions requiring detailed recall of spending), and/or

sensitive in nature (e.g. include questions about mental health). These surveys also tend to cover a

specific topic (e.g., mental health, labour force, income), so households will often self-deselect from

the survey as they assume their data is not relevant to the statistics produced (e.g., retirees often

self-deselect from the labour force survey).

Considering these challenges, having an effective contact and cooperation strategy is paramount to

the ABS producing quality and timely statistics. The aim of this paper and accompanying

presentation is to explore the effectiveness of three approaches designed to support and advance

the ABS’s contact and cooperation strategy. These approaches are:

1. Developing a tailored approach strategy to engage key demographic groups. These include

those with a high non-response propensity, or groups that are critical to have represented in

the data based on the survey topic (e.g. those with low or high socio-economic status for

income and expenditure surveys).

2. Offering incentives for completing a government survey

3. Collecting contact details through a household survey induction model, whereby the

respondent is initially introduced to the survey program, rather than to a specific survey.

These approaches were designed to work either independently, or in combination with one another,

depending on the collection requirements.

3. METHOD

The project consisted of two distinct phases – the ‘discovery phase’, which included deep-dives

through literature reviews and extensive consultation with experts across the ABS, and the

‘prototype and test phase’, which included prototype development and cognitive testing with

members of the public.

For cognitive testing, 16 members of the Australian public were recruited through Askable

recruitment service (Askable Pty Ltd, n.d). To capture a variety of demographics, filtering questions

were added in the Askable advertisement, followed by purposeful selection from those who replied.

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

4 | P a g e

Of the 16 participants, 12 were involved in the testing of ‘tailored approaches’ (see Section 4) and

the remaining 4 were involved in the testing of the ‘household survey induction model’ (see Section

6 below). All 16 participants were asked questions about incentives (Section 5).

4. TAILORED APPROACH TO ENGAGE KEY DEMOGRAPHIC GROUPS

There are two ways in which a contact and cooperation strategy can be tailored to engage key

demographic groups. You can create a single contact and cooperation strategy that is optimised to

engage multiple demographic groups using contemporary, best practice human-centred design

methods. Alternatively, you can create multiple contact and cooperation strategies that are tailored

to meet the specific needs of each key demographic group. Whether a single strategy or multiple

strategies should be adopted is largely dependent on the commonalities and differences between

the demographic groups in terms of barriers to survey response, and best practice engagement

techniques. Consideration also needs to be given to the cost and operational complexity associated

with managing and maintaining the strategy/ies.

To this end, a review of internal and external research was conducted to understand barriers, as well

as best practice engagement strategies, for key demographic groups. For this particular project,

demographic groups that are known to have a high non-response propensity, and/or are critical for

the production of accurate income and expenditure statistics were explored (see Figure 1).

The review found that while there were many differences amongst the groups, there were also

many commonalities, both in terms of barriers and strategies that can address them (see Figure 1).

Given these commonalities, the project team decided to pursue a single strategy because of the

potential benefit it would efficiently deliver to a broad range of demographic groups. Updated

approach letters were prototyped based on this decision, and on the review of the commonalities

between groups. The approach letters were positively received during cognitive testing, and now

form the basis for the standard suite of approach letters for ABS household surveys. Future research

will seek to further optimise this single approach strategy.

Figure 1. Barriers and best practice engagement strategies for key demographic groups for income

and expenditure statistics.

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

5 | P a g e

5. INCENTIVES FOR COMPLETING A SURVEY

Research has shown that incentives are an effective tool to motivate survey response (Church, 1993;

Edwards et al., 2005; Mercer et al., 2015; Ryu Wong, 2022). As a result, statistical organisations

around the world are increasingly turning to incentives to boost declining response rates, and

mitigate the risk of non-response bias (Crossley & Winter, 2015). To date, most studies looking at the

effectiveness of incentives have been conducted outside of Australia, with surveys that are

voluntary. It is possible that the effectiveness of incentives may be received differently in Australia,

and for non-voluntary surveys (most ABS surveys are compulsory).

The project team conducted cognitive testing with 16 members of the Australian public to explore

possible uses for incentives in ABS surveys. The interview questions focused on financial incentives

(e.g. a gift card) as the literature suggests this to be the most effective incentive type (compared to

non-financial incentives, such as tote bags or coffee mugs).

One key finding from cognitive testing was that participants did not expect to receive an incentive

for completing a compulsory ABS survey. They viewed completing the survey as something they “just

have to do” as part of their civic duties, with some expressing that financial incentives are

unnecessary or even wasteful. That said, participants also expected that a compulsory survey would

take between 5 to 30 minutes. Once informed that it may take 1 to 1.5 hours to complete the

survey, the majority of participants either expected or would appreciate compensation for what they

saw as an ‘unreasonable’ time commitment. This finding hints at a possible interaction between

compulsion (or at least civic obligation) and survey length on the perceived appropriateness of

financial incentives for government surveys. While incentives may be considered inappropriate for

shorter surveys, they could be an important tool for communicating value and reciprocity for surveys

that require a greater time commitment.

6. HOUSEHOLD SURVEY INDUCTION MODEL

The ABS’ current approach strategy involves sending letters to selected households explaining their

role in a specific survey. These letters include what they can expect at each stage of their

participation, including mode options, the sorts of questions they’ll be asked, and any additional

activities required to fulfil their obligation (e.g. completing an expenditure diary). This information is

provided upfront to help set respondents’ expectations on the survey process.

There are two challenges with the ABS’ current approach strategy: 1. there is some evidence that

providing too much information upfront can reduce response due to the perceived burden of the

task (Data Collection Design Centre, 2021), and 2. there is a growing organisational need for

flexibility in updating survey topics as new and/or urgent data needs arise. To address these two

challenges, we explored the concept of a ‘household survey induction model’, whereby generic

approach materials are provided to respondents at initial contact, with survey-specific information

provided at the time that they require it (e.g. upon starting the survey).

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

6 | P a g e

Cognitive testing found that participants were generally comfortable with not receiving survey

specific information at the first contact, and made assumptions that the survey would ask about a

range of topics relevant to their own lives. However, when participants were informed upon starting

the survey that it would take approximately 1.5 hours to complete, they expressed frustration with

not having been told this information upfront so they could plan when to do it. Participants said

knowing the length of the survey would not deter them from completing it, and would help them to

decide the most appropriate time for them to sit down to complete it.

7. CONCLUSION

Designing an effective contact and cooperation strategy is critical to the collection of high quality

survey data. In this paper we explored three approaches that have potential to advance the ABS’

contact and cooperation strategy. We found that small design updates to our materials, taking into

account commonalities amongst key demographic groups, can increase overall engagement with the

survey process. Furthermore, we found that a different approach may be warranted depending on

the time commitment and effort required to complete a survey. For surveys that are in line with

what is considered a reasonable time commitment for a compulsory survey (5 to 30 minutes),

approaching households through a household survey induction model may be appropriate.

However, for surveys beyond what the public deems a reasonable time commitment, ensuring

optimal engagement from respondents may require being upfront about the survey requirements

(particularly around time commitment), and/or providing incentives to motivate response.

8. ACKNOWLEDGEMENTS

We would like to thank Annelie De Villiers, Bryony Cresswell and Katherine Birrer for their assistance

with the design, test and analysis of the three approaches explored, as well as their unique and

thoughtful contributions to discussions throughout the project.

9. REFERENCES

Askable Pty Ltd (2022, July 13) Askable recruitment service. www.askable.com

Australian Bureau of Statistics. (2020). 13.2 million Australians exposed to scams. 13.2 million Australians exposed to scams | Australian Bureau of Statistics (abs.gov.au)

Australian Bureau of Statistics. (2023). ABS Address Register, Users’ Guide. ABS Address Register,

Users' Guide | Australian Bureau of Statistics Church, A. H. (1993). Estimating the effect of incentives on mail survey response rates: A meta-

analysis. Public opinion quarterly, 57(1), 62-79 Crossley, T. F., & Winter, J. K. (2015). Asking households about expenditures: what have we

learned?. In Improving the measurement of consumer expenditures (pp. 23-50). University of Chicago Press

IMPROVING SURVEY RESPONSE BY DEVELOPING AN EFFECTIVE CONTACT AND COOPERATION STRATEGIES

7 | P a g e

Data Collection Design Centre (2021). Survey of Income and Housing (SIH) “documents card” style

leaflet experiment, Australian Bureau of Statistics National Data Acquisition Division. Edwards, P., Cooper, R., Roberts, I., & Frost, C. (2005). Meta-analysis of randomised trials of

monetary incentives and response to mailed questionnaires. Journal of Epidemiology & Community Health, 59(11), 987-99

Finsterwalder, J., & Tombs, A. G. (2021). Infusing tribal reciprocity into service research: towards an

integrated and dynamic view of repayment, retaliation, and restorative justice for regenerative service ecosystem wellbeing. Service Business, 15(4), 563-586.

Mercer, A., Caporaso, A., Cantor, D., Townsend, R. (2015) How Much Gets You How Much?

Monetary Incentives and Response Rates in Household Surveys, Public Opinion Quarterly, Volume 79, Issue 1, Spring 2015, Pp. 105–129, https://doi.org/10.1093/poq/nfu

Ryu Wong, E. (2022) Incentives in Longitudinal Studies. Economical and Social Research Council.

Centre For Longitudinal Stud

Designing respondent communications for enterprise surveys Trials and tales from Australia

Spurti D’souza, Caitlin Huppert, Jodie Kline

Behavioural insights and human-centred design techniques were used to develop and evaluate:

Respondents at the core of service design

No statistically significant differences

Authoritative messaging in survey invitation materials

Control: ‘Please complete your…’ Treatment: ‘Notice of Selection’

Control: ‘Complete your survey online’ Treatment: ‘Complete your obligation online’

Control: ‘If you’d like to speak with us or need help participating in the survey, please call …’ Treatment: ‘Contact us if this business is unable to meet its reporting obligations. To speak to our team, please call ...’

Control: Legal authority statement placed lower Treatment: Legal authority statement placed higher

Reminder card

ncontrol group = 1,562 ntreatment group = 500

+ 4.71 percentage points in response rate at business-as-usual due date

+ 7.58 percentage points of treatment group had true self-initiated response

Clear and specific respondent-intentions

Simple instruction for using the card

A proactive reminder tool

Control = 5% response rate

Treatment = 13% response rate

Statistically significant difference

Late Reminder SMS

Moderate tone of urgency

Two clear options

Firm tone of urgency

Referencing the legal authority of the ABS to collect this information

Clear single call to action

‘We missed you’ emails

7

Plain English

Firm call to action

Two clear options

Simple legal authority statement, in an FAQ format.

No results available at this time

Thank You Letters

Plain English

Reinforcing the purpose and importance of data collection

Returning value to respondents with access to published data

Tailored to the respondent

Keep in touch [email protected]

DESIGNING RESPONDENT COMMUNICATIONS

FOR ENTERPRISE SURVEYS: TRIALS AND TALES

FROM AUSTRALIA

Spurti D’souza, Caitlin Huppert, Jodie Kline

Data Collection Design Centre

Australian Bureau of Statistics

June 2023

PAPER PRESENTED AT

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

Expert Meeting on Statistical Data Collection – Rethinking Data Collection

12-14 June 2023

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 2 of 12 - www.abs.gov.au

1.0 ABSTRACT

Behavioural insights and human-centred design underpin respondent communications the

Australian Bureau of Statistics (ABS) designs and develops. This approach is guided by the ABS’

enterprise-level strategic priorities, including to enhance organisational capability, resilience, and

adaptability, and to reduce burden on respondents. In doing so, our service offering improves

respondent experience and supports respondent engagement with our survey program.

This submission showcases modes and features of communications trialled in enterprise surveys

during 2022-23, including the drivers for delivery, the design process, and the outcomes of initial

implementation.

Specifically, we share learnings associated with:

1. Authoritative messaging in survey invitation materials for new-to-sample respondents.

2. Proactive reminder card to promote self-initiated response.

3. Late reminder SMS.

4. ‘We missed you’ email sent following an unanswered outbound call.

5. ‘Thank you’ paper letter.

The trials demonstrate benefits of placing the respondent at the core of service design, as well as

raise critical questions about how to balance respondent preference and statistical integrity.

2.0 INTRODUCTION

Respondents are central to the success of the Australian Bureau of Statistics’ (ABS) survey program.

Annually, the ABS runs more than thirty statistical collections. Without the willing cooperation of

businesses and organisations, generation of our national statistics would be compromised. We

couple this knowledge with recognition that our surveys need to be prioritised alongside other

business activities, in a context where labour and supply pressures are acute (ABS, 2022, June). With

this in mind, we strive to support respondent engagement through delivery of a very good

respondent experience.

Traditionally, we focused our energies on growing internal understanding of human behaviour

broadly, and theories of survey participation more specifically. Through a team of experts in the

fields of behavioural economics, sociology, psychology, and data science we applied behavioural

insights to the design of our services and products. Our knowledge of social exchange theory and

cognitive heuristics underpinned the development of respondent engagement strategies. Designs or

prototypes were often presented to respondents for feedback, via consultation late in the

development process, but respondents themselves were not routinely central to the design process.

More recently we have shifted away from solely relying on internal expertise, to also involve our

respondents in a respondent-centred development approach. Specifically, our design work now

draws primarily on human-centred design tools, coupled with application of behavioural insights.

Respondent perspective is now at the heart of the process.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 3 of 12 - www.abs.gov.au

Human-centred design (HCD) recognises the benefit associated with co-designing with users (in our

context, the survey respondents), encourages rapid prototyping and iterative testing, and obliges

designers to view problems from the perspective of users (Digital Victoria, 2022). HCD invites user

involvement in the design of products and services that meet their needs and deliver business value

(Smith, 2021; Digital Victoria, 2022). By using HCD, we have invested in research to engage with and

better understand the situation of the respondents.

Our application of HCD to enhance statistical data collection is demonstrated through case studies

and trials of five innovations associated with a large-scale project to redesign all economic survey

correspondence (Meld Studios, 2022 April; Data Collection Design Centre, 2023 May):

1. Authoritative messaging in survey invitation materials for new-to-sample respondents.

2. Proactive reminder card to promote self-initiated response.

3. Late reminder SMS.

4. ‘We missed you’ email sent following an unanswered outbound call.

5. ‘Thank you’ paper letter.

Development and implementation of these innovations delivered benefit to the survey program, as

well as exposed opportunities to mature ABS’ use of HCD and highlighted the complexities

associated with balancing respondent preference and statistical integrity.

3.0 INTERVENTIONS

3.1 Authoritative messaging in survey invitation materials for new-to-sample

respondents

Research question: Can increased authoritative messaging improve new-to-sample respondents’

survey participation?

Most ABS surveys are compulsory to complete; however, new-to-sample respondents often

comment that this is unclear, in turn, delaying their survey participation. This motivated us to

investigate whether increased authoritative language improves the effectiveness of the data

acquisition process with new respondents.

We conducted concept generation and user testing with respondents to explore variations in firmer

and authoritative language. The research concluded that respondents expected correspondence to

include a clear statement about whether completing an ABS survey is compulsory and they

preferred a plain English answer to this question over a detailed legal explanation. These HCD

techniques helped us understand the respondents’ cognitive processes, and understand that clear

and authoritative messaging can inform and influence their survey participation.

To determine how the strengthened authoritative messaging impacts respondents’ behaviours, in

2022 two small-scale randomised control trials were conducted over two cycles of a quarterly survey

(Data Collection Design Centre, 2022 October; Data Collection Design Centre, 2022 December). The

experimental and control groups contained 100-200 respondents each. The compliance model of

survey response as articulated by Snijkers, Berkenbosch and Luppes (2007), including Groves,

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 4 of 12 - www.abs.gov.au

Cialdini and Couper’s (1992) authority compliance principles was referenced and applied. Messaging

in approach materials for new-to-sample units was modified to emphasise the legal authority for

statistical data collection, and compulsion messaging was strengthened. Changes included removing

the word ‘survey’ and using language such as ‘obligation’. For example, ‘Complete your survey online’

was replaced with ‘Complete your obligation online’. Additionally, ‘Notice of Selection’ was the

header used for the experimental approach materials. In sum, only the text (no other design

elements, such as formatting) was changed between the control and experimental conditions.

The strengthened authoritative messaging in approach letters and emails did not make a statistically

significant difference to response rates or respondent engagement, as measured in the trials.

Findings suggest that the experimental compliance messaging had neither a detrimental impact on

timely survey participation, nor did it appear to deliver benefit compared to existing materials. There

are several ways to interpret the results:

• The compliance messaging tested was not strong enough or clear enough to impact

response and engagement by due date.

• The compliance principle of ‘liking’ coupled with soft use of the ‘authority’ compliance

principle embedded in the existing materials is as effective at driving response as firm

compliance messaging alone.

• Optimisation of other elements in correspondence are more likely to drive improvements in

response and engagement.

Subsequent prototype and user testing has moved away from focussing on the compulsion

messaging alone, and a shift towards a holistic review of the whole suite of correspondence and

escalating the authoritative tone in correspondence sent after the due date (see Willimack &

Snijkers, 2013 for discussion of escalation). This has tested well with respondents in usability testing,

and we continue to work closely with respondents and our policy and legal team to develop and test

variations of authoritative messaging as part of a project aiming to enhance the end-to-end

respondent journey.

3.2 Proactive reminder card to promote self-initiated response

Research question: Can a proactive reminder tool, a ‘reminder card’, improve survey participation

including timely response?

Obtaining timely response, a complete and submitted survey form by the respondents’ due date, is

an ongoing and growing challenge we face across many collections. Concurrently, from the

respondents’ perspective, they have competing priorities of which an ABS survey is simply one of

many. Respondents requested tailored tools to help them submit their survey on time. Goal-setting

theory (Locke & Latham, 2002), the INSPIRE framework (Faulkner et al, 2018), and a study to

enhance influenza vaccination rates (Milkman et al, 2011) posited that a goal needs to be set by the

acting individual, and their intention needs to be clear and specific.

Inspired by these researchers and by positive reaction from respondents to proactive reminder

materials, we randomly selected 500 respondents in a biennial enterprise survey to receive a

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 5 of 12 - www.abs.gov.au

reminder postcard in the same envelope as their invitation letter (Data Collection Design Centre,

2023 January). The reminder card contained key pieces of information the respondent needed to

complete their survey online: business-as-usual (BAU) due date, blank space for them to handwrite

their intended submission date, alphanumeric code to access the web form, URL to complete the

survey, URL for frequently asked questions, and help line phone number (see Image 1).

Image 1. Reminder card

The treatment group (25.2%, n = 126) experienced a significantly greater response rate, by 4.71

percentage points (p = 0.0158, Power = 94.09%), by the BAU due date than units in the control group

(20.49%, n = 320). The significantly different response rate held through for the two weeks that

followed the due date. We saw 4.74 percentage points greater response rates one week after the

due date (p = 0.0215, Power = 92.13%), and 5.39 percentage points greater response rates two

weeks after the due date (p = 0.0127, Power = 95.08%).

Of the timely responders, most were returning units in both the control group (73.75%, n = 236) and

in the treatment group (77.78%, n = 98). This could indicate that:

i. familiarity with the survey is beneficial to the respondent’s ability and willingness to

complete the survey by the due date, and

ii. the Reminder 1 email and SMS, sent on the day before the BAU due date, continues

to significantly boost response rates across a wide range of ABS business surveys.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 6 of 12 - www.abs.gov.au

Of the units that responded by the due date, significantly more of the treatment group responded

without any contact attempts than those in the control group (+ 7.58 percentage points, p = 0.0252,

Power = 88.78%). That is, the reminder card significantly improved true self-initiated response,

which directly reduced our follow-up outbound call costs and efforts.

Further trials will help us understand which surveys or types of respondents benefit from the

reminder card. For example, this biennial survey contacts individuals who are rarely in other ABS

surveys. The infrequent contact with the ABS coupled with the design principles in the reminder card

may have contributed to its success. Additionally, this survey sample has a high proportion of

returning units (73.79%, n = 3,043). Perhaps familiarity with the survey and the novelty of the

reminder card helped shape the favourable outcome. It could even be a combination of these and

other factors, that only further trials will help us unpick.

3.3 Late reminder SMS

Research question: Can a late reminder via SMS, after the business-as-usual due date, influence

survey response rates?

The ABS is exploring whether a late SMS reminder is an efficient method to improve the response

process for enterprise collections. Sending an SMS is identified as an effective form of behavioural

intervention, such as improving reporting behaviours among Australian welfare recipients (Copley et

al, 2017: Dillon et al, 2021). Much of the existing literature has focused on social surveys as opposed

to enterprise surveys. Respondents involved in the review of our correspondence suite appreciated

the convenience of SMS communications, but noted that mobile phones are, in the main, personal

devices, and advised us to tread cautiously when it came to use of SMS (Data Collection Design

Centre, 2023 May). This is consistent with current research into mobile device use within Australia,

which suggests that employees with personal mobile devices have at some point used their device

for business related purposes (Deloitte Access Economics, 2019).

Much of the existing literature has focused on individual based surveys as opposed to business-

based surveys. As such, through a concurrent project we sought respondent perspectives on use of

SMS. We coupled this work with a randomised control trial (RCT) and several live trials with the goal

of contributing to the literature, building an understanding of how to use SMS communication

between a national statistical organisation and survey respondents, specifically in the enterprise

survey space (see Image 2). The RCT was conducted within a survey of the Australian building

industry, while the live trials were conducted across three cycles of two quarterly Australian

employment and wages surveys.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 7 of 12 - www.abs.gov.au

Image 2. Late reminder SMS

The RCT found that SMS as a reminder is an effective means to communicate with businesses once

the due date has passed. We observed statistically significant increases in both survey response rate

and inbound contacts made by the respondents who received the SMS. Including a 13% response

rate for the treatment group and a 5% response rate for the control group. Additionally, in the

inbound calls received to the ABS call centre, there was a 5% call rate for the treatment group and a

2% call rate for the control group. Of the calls received from the treatment group, 82% resulted in

their data being collected over the phone.

In the live trials, however, we could not directly correlate the response rate increase with the SMS

due to it not being conducted under an experimental design. The results of these trials support the

need for further investigation.

3.4 ‘We missed you’ email sent following an unanswered outbound call

Research question: Can a ‘We missed you’ email help convert unsuccessful outbound calls into

successful contacts?

The December quarter is riddled with seasonal impacts due to the summer and holiday season in

Australia. Our outbound calls are answered less frequently, and more voicemail messages are left.

However, it is unclear whether respondents know the missed calls are from the ABS. We explored

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 8 of 12 - www.abs.gov.au

whether an email tailored to this situation would improve respondent contact efficiency during this

season.

To help bridge the gap between the missed call and the successful contact desired, we developed

and tested a prototype, and then trialled a ‘We missed you’ email (see Image 3). Our call centre staff

send it to a respondent following an outbound missed call, an unsuccessful contact. Respondents

suggested the design concept during idea generation sessions and additional respondents enhanced

the design during prototype testing (Data Collection Design Centre, 2023 May).

Image 3. ‘We missed you’ email

For easy adoption by our call centre staff, it was trialled with all business surveys in the field from

December 2022 to January 2023. Although business survey respondents prefer electronic

communications, there are challenges with emails, telephone calls and text messages in the current

environment due to spam, scams and phishing experiences. Additionally, we experienced

operational challenges whereby the ‘We missed you’ emails were applied inconsistently between

call centre staff as it added to their administrative burden.

Although these emails yielded no significant impacts on our usual metrics (e.g. response rates,

successful calls to contact), they helped generate a small volume of returned calls and online

responses. Furthermore, the call centre staff who did use it provided positive anecdotal feedback. It

has since been rolled out for wider and more consistent use, and continued monitoring. We expect

to see its impact in less tangible aspects, such as the respondent-ABS relationship.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 9 of 12 - www.abs.gov.au

3.5 ‘Thank you’ paper letter

Research question: Will a ‘Thank you’ paper letter sent in between cycles help maintain a longitudinal

panel?

Many of our cyclical enterprise surveys face longitudinal panel maintenance issues. We are currently

exploring whether a ‘Thank you’ paper letter in between cycles will help us maintain the panel, and

in turn improve their future survey participation.

Past ABS research suggests that survey respondents who submit in their first cycle are likely to

continue submitting regularly in subsequent cycles (Data Collection Design Centre, 2022, August).

However, our analysis across multiple quarterly surveys for multiple cycles, shows that this pattern

becomes disrupted after the third cycle of the survey, with respondents’ submission behaviours

becoming irregular (Data Collection Design Centre, 2022, August).

Literature suggests that positive reinforcement of desired respondent behaviour will lead to

respondents repeating the behaviour in future survey cycles (Greenberg & Dillman, 2021). Past

cognitive testing indicated that respondents feel their relationship with the ABS is very one-sided.

This reduced their motivation to engage with the ABS or to spend time ensuring the data that they

submitted was accurate.

In a large HCD project (Meld Studios, 2022), respondents indicated they wanted their efforts to be

acknowledged. Additionally, a small ABS study conducted in 2022 suggested that respondents who

receive a ‘Thank you’ letter after submitting a survey in one cycle are more likely to submit again

during the next cycle, compared with those who did not receive the ‘Thank you’ letter (Data

Collection Design Centre, 2022). We are now extending this work with a quarterly survey.

The current longitudinal study follows new respondents split into treatment and control groups for

five cycles, analysing their submission habits. This study is in the second of five phases, and as

expected at this point, there is currently no statistically significant difference in response rates

between the groups. At this stage there is no evidence of a detrimental impact of the ‘Thank You’

letter on the respondent. However, we will continue to monitor this and will have further

information in late 2024.

4.0 CONCLUSION

In the context of declining response rates (Australian Bureau of Statistics, 2022 April 27; Luiten, Hox

& de Leeuw, 2020; The World Bank, 2020), actively engaging with respondents as assets is accepted

as the way forward for ABS data collection. As indicated in the Information Notice of this expert

meeting (UNECE, 17 January 2023), there has been a ‘paradigm shift that puts the respondents at

the centre of the data collection design and management’.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 10 of 12 - www.abs.gov.au

In the past we focused on the symptoms of barriers and looked to external literature, theory, and

operational data to help us understand behaviour. Respondent feedback received via various

pathways was considered and valued and respondents were consulted and involved in prototype

testing, but they were not central to the design of solutions. There was not a two-way conversation,

whereby respondents were not invited to help define the problem or involved in idea generation

and concept testing. Instead, they were often engaged late in the design process once prototypes

had been developed.

The five interventions discussed demonstrate the benefits of placing the respondent at the core of

service and product design. By working with them we can understand barriers to response as

perceived by respondents. The iterative approach enabled us to evolve and test ideas, ahead of

scaled implementation. As such, by the time deliverables land in the field – which not all did – they

are more likely to connect with respondent need and therefore yield the desired outcomes.

Recent respondent-focused design and evaluation work has also revealed the benefits of HCD as an

approach that enables us to balance respondent preference and statistical integrity, by involving

both respondents and ABS staff in the design process. Through this, our requirements are

communicated more effectively, and respondents are clearer about their role in the construction of

national statistics. We introduced several of the ‘friendlier’ and pro-active communications that

respondents requested, and successfully tested the more immediate and urgent messaging to

promote compliance late in the survey cycle. Each intervention has the potential to boost data

integrity and does so by adhering to respondent-specified modes, content, and timing contact.

5.0 REFERENCES

Australian Bureau of Statistics. (2022, April 27). Raising survey response rates by using machine

learning to predict gold providers. www.abs.gov.au/statistics/research/raising-survey-response-

rates-using-machine-learning-predict-gold-providers#abstract/

Australian Bureau of Statistics. (2022, June). Business conditions and sentiments.

https://www.abs.gov.au/statistics/economy/business-indicators/business-conditions-and-

sentiments/latest-release

Copley, S., Brewer, J., Turenko, A., Wilson, J., & Hiscox, M. (2017). Effective use of SMS: Timely

reminders to report on time. Australian Government, Canberra.

Data Collection Design Centre. (2022, August). Analysis of response behaviour in new respondents.

Australian Bureau of Statistics National Data Acquisition Division, Melbourne.

Data Collection Design Centre. (2022, June). Tailoring early contact for late and non-responders:

Monitoring the impact of previous cycle interventions. Australian Bureau of Statistics National Data

Acquisition Division, Geelong.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 11 of 12 - www.abs.gov.au

Data Collection Design Centre. (2022, October). Can reminding business survey respondents that

completion is compulsory improve survey engagement? Perhaps, but only by a bit. Australian Bureau

of Statistics National Data Acquisition Division, Geelong.

Data Collection Design Centre. (2022, December). Trial of compliance messaging in approach

materials, Pilot study no.2. Australian Bureau of Statistics National Data Acquisition Division,

Geelong.

Data Collection Design Centre (2023, May). Designing survey correspondence to improve

engagement with businesses. Australian Bureau of Statistics National Data Acquisition Division,

Melbourne.

Data Collection Design Centre (2023, January). Reminder card mailed with approach letter to

encourage timely response. Australian Bureau of Statistics National Data Acquisition Division,

Melbourne.

Deloitte Access Economics (2019). Mobile Nation 2019: The 5G future.

https://www2.deloitte.com/content/dam/Deloitte/au/Documents/Economics/deloitte-au-

economics-mobile-nation-2019-080419.pdf

Digital Victoria (2022, August). Human-centred design playbook. Department of Premier and Cabinet

Service Design Team, Melbourne.

Dillon, A., Glazerman, S., & Rosenbaum, M. (2021). Messaging to improve response rates:

Effectiveness of pre-survey SMS messages. Global Poverty Research Lab Working Paper, (21-106).

Faulkner, N., Borg, K., Bragge, P., Curtis, J., Ghafoori, E., Goodwin, D., Jorgensen, B. S., Jungbluth, L.,

et al. (2018). The INSPIRE framework: How public administrators can increase compliance with

written requests using behavioural techniques. Public Administration Review, 79(1), 125-135.

Greenberg, P., & Dillman, D. (2023). Mail communications and survey response: A test of social

exchange versus pre-suasion theory for improving response rates and data quality. Journal of Survey

Statistics and Methodology, 11(1), 1-22.

Groves, R. M., Cialdini, R. B., & Couper, M. P. (1992). Understanding the decision to participate in a

survey. Public opinion quarterly, 56(4), 475-495.

Locke, L. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task

motivation. American Psychologist, 57(9), 705-717.

Luiten, A., Hox, J., & de Leeuw, E. (2020). Survey nonresponse trends and fieldwork effort in the 21st

century: Results of an international study across countries and surveys. Journal of Official Statistics,

36(3), 469-487.

Meld Studios (2022, April). ABS client experience review insights report. For the Australian Bureau of

Statistics National Data Acquisition Division, Melbourne.

DESIGNING RESPONDENT COMMUNICATIONS FOR ENTERPRISE SURVEYS Trials and Tales from Australia June 2023

Page 12 of 12 - www.abs.gov.au

Milkman, K. L., Beshears, J., Choi, J. J., Laibson, D., & Madrian, B. C. (2011). Using implementation

intentions prompts to enhance influenza vaccination rates. Proceedings of the National Academy of

Sciences, 108(26), 10415-10420.

Smith, K. (2021). Human-centred design online workshop 23-24 November 2021. Meld Studios.

Snijkers, G., Berkenbosch, B., & Luppes, M. (2007, June 18-21). Understanding the decision to

participate in a business survey. 3rd ICES-III, Montreal, Quebec.

The World Bank. (2020, July 6). Survey of National Statistical Offices (NSOs) during COVID-19.

www.worldbank.org/en/research/brief/survey-of-international-statistical-offices-nsos-during-covid-

19/

United Nations Economic Commission for Europe (UNECE), Statistical Division (2023, January 17).

Notice No.1, email correspondence. UNECE conference of European statisticians expert meeting on

statistical data collection: Rethinking data collection Information.

Willimack, D., & Snijkers. G. (2013). The business context and its implications for the survey response

process. In Designing and Conducting Business Surveys (pp. 39 – 82). Wiley.