Try, Test and Learn Fund – Evaluation Activity Planning 2017
Stage / task
Proposed evaluation-related activities
Indicative
Potential Evaluation Unit role and notes
(Overview)
timing
TTL Fund evaluation Scoping PIR
January
EU to lead and advise? TTL team to prepare documentation and
provide support as appropriate
This task involves an
Procuring evaluation provider to deliver PIR
February
EU to lead and advise? TTL to undertake administrative and
evaluation of the TTL
documentation tasks as appropriate
Fund as a whole. This
Managing evaluation provider contract
February
EU responsible for contract management? TTL to provide input as
may take the form of a
onwards
needed
Post Implementation
Review.
Idea assessment
Evaluation content in assessment team training and supporting
January
EU to provide input and advice? TTL team to prepare
documentation (i.e. operationalising “potential for robust
documentation.
This stage involves the
evaluation of outcomes” and other evaluation issues in
assessment of the TTL
assessment criterion 1, see footnote over page).
Fund idea submissions.
Short evaluation presentation in assessment team
February
EU to prepare content and deliver? TTL team to provide TTL
It also involves the
induction/training session
specific content as needed.
collation of a shortlist of
recommended ideas to
Specialist evaluation input into assessments of ideas as
22 Feb – 10
EU to provide staffing assistance / input to assessment process
progress to co-
required? (some possible approach include: evaluation team
March
as appropriate.
development, to be
moderation of selection of ideas; small team of specialised
provided to the Inter-
assessors completing review against criteria 1 for all subs; more
Departmental
detailed evaluation readiness assessment of top scoring ideas?
Committee.
Consideration of shortlist from evaluation perspective – i.e
6 March – 10
EU to provide input to shortlisting as appropriate.
ensuring recommended ideas have a feasible evaluation
March
workload and a mix of complex and more straightforward
evaluation needs.
Co-development
General planning of evaluation-focused components of the co-
January –
TTL Team will be planning co-development more broadly. EU can
development approach
February
provide direct input or can advise through a workshop/meeting(s).
This stage involves
Specific planning of evaluation-focused co-development
13 March –
This could be conducted by EU staff? Or by contracted evaluation
collaborative activities
activities for each shortlisted ideas
17 March
consultants.
including idea submitters
and relevant experts to
Conducting evaluation-focused activities, such as program logic
17 March –
This could be conducted by EU staff? Or by contracted evaluation
develop and refine
or theory of change workshops with co-development partners.
14 April
consultants.
shortlisted ideas to
Evaluation scoping and planning for each project under co-
11 March –
EU to lead? Could include advice from Expert Advisory Panel
detailed project
development
14 April
members, and assistance from contracted evaluation consultants.
proposals. Project
May also include idea submitters if evaluation partners were
proposals will be
reviewed by the Minister,
included in idea submission team.
with a selection
(If required) procuring evaluation specialists to conduct
13 March –
EU to lead and advise? TTL to undertake administrative and
proceeding to funding
evaluation planning for projects during co-development period.
17 March
documentation tasks as appropriate
and delivery.
Funding
Procuring evaluation providers to deliver evaluations of funded
13 May to 26
EU to lead and advise? TTL to undertake administrative and
projects
May
documentation tasks as appropriate
1
This stage involves
selecting a delivery
partner to implement
selected ideas. This will
generally involve a direct
or restricted grant
process.
Delivery and
Managing evaluation contracts
June
EU responsible for contract management? TTL to provide input as
evaluation of
onwards
needed
projects
This stage involves the
delivery and evaluation
of selected projects.
Footnote (from Try, Test and Learn Fund Trance 1 Handbook)
Appropriateness for trial and evaluation
The idea will be assessed on its potential for a robust evaluation of its outcomes and how suitable it is to be run as a trial.
The Department will consider whether an idea offers a testable hypothesis that could be empirically evaluated, including whether it is suitable for best practice methods
including randomised controlled trials.
The Department will also consider whether the idea has the potential to be delivered to a sufficient number of participants for statistically significant analysis of program
outcomes. The appropriate number of participants for statistically significant analysis will vary by program and will be explored with Proposers of shortlisted ideas during the
co-development stage.
Ideas that are suitable for trial will need to have the potential to be scaled up or replicated and delivered to more individuals or communities if they are successful. They
should also have the potential to be discontinued after an initial trial period if they are not.
The Department will also consider whether the idea would be overly complex, impractical or expensive to implement and evaluate.
2

Monitoring and Evaluation of Policy Responses
Monitoring and evaluation will be the key to testing and learning from the policy responses
trialled under the Try, Test and Learn Fund. The diagram below sets out the implementation
activities that will involve consideration of monitoring and evaluation issues, and which
players will have a role.
Monitoring and evaluation activities
All policy responses will be evaluated using a combination of analysis of Priority Investment
Approach data and modelling, and supplementary data collection and evaluation. It is likely
that external evaluators will be engaged to evaluate policy responses in liaison with service
providers and the Government. The roles of these stakeholders are captured below.
Monitoring and evaluation stakeholder roles
Stakeholder
Roles
Government
• Assessing policy responses for evaluation readiness
• Engaging external evaluators (for most policy responses)
• Liaising with external evaluators and service providers
External evaluator
• Design monitoring and evaluation for relevant policy response
• Carry out ongoing monitoring and evaluation as appropriate
• Liaising with Government and service providers
Service provider
• Collecting and sharing data on policy responses
• Liaising with Government and external evaluators
s 22
Assessment of policy responses
Evaluation readiness will be one of the factors considered in the assessment and selection
of policy responses, which is discussed in the separate paper about the application process.
Monitoring and evaluation design
The monitoring and evaluation design for a policy response will define outcomes and
timeframes, data and data collection, and analytical methods to be used. Where they are
engaged, external evaluators will lead this design process.
Outcomes
3
link to page 4
The Try, Test and Learn Fund will support targeted trials of policy responses for defined
groups of several thousand people. While these trials may have significant impacts on the
target groups, it is unlikely that these impacts will be visible in the annual population-wide
valuations. Instead, the outcomes will be measured through tailored analysis and
evaluations conducted internally and by external, independent evaluators.
The outcomes used to measure success will be guided by the goals of the Priority
Investment Approach.
1 The specific outcomes will differ for each policy response and the
barriers to work that it seeks to address. They will involve questions such as:
• Are priority groups participating in work or education at increased levels?
• Are priority groups less reliant on welfare (e.g. receiving lower rates of payment)?
• Are priority groups finding work and exiting the welfare system sooner?
• Are the average future lifetime costs for priority groups increasing?
Sources and methods The Priority Investment Approach model and dataset will be the first data source considered
in the monitoring and evaluation design for a policy response. The questions above can be
tracked using indicators in this model and dataset, such as:
• Number of people on payment
• Payment transitions
• Earnings from work
• Payment rate (full rate or part rate due to income from work)
• Participation in education
• Average future lifetime costs (noting that the influence of policy changes on these
costs may not be reflected in the model for two or more years)
Outcomes will be measured by comparing indicators for people participating in the policy
response with a similar group of people not participating. This approach allows us to
measure the impact that the policy is having. The methods used to achieve this comparison
may include:
• Randomised controlled trials
• Quasi-experimental methods
• Difference-in-difference analysis
Importantly, data sources beyond the Priority Investment Approach model and dataset will
be used to ensure a complete picture of what’s working and how. Surveys, post-participation
monitoring, qualitative data and customer focus groups or interviews will be utilised where
appropriate to complement econometric, statistical and other quantitative techniques.
Outcomes that may be captured with this additional analysis include priority groups’ capacity
for work, increasing work skills and knowledge, wellbeing, social inclusion, financial stress
and family functioning.
1 Improve the lifetime wellbeing of people and families in Australia through an Investment Approach
that: increases the capacity of people to live independently of welfare; decreases the
Commonwealth’s long-term social security liabilities; and reduces the propensity for intergenerational
welfare dependence.
4
Monitoring and evaluation
In most cases, the external evaluator will carry out the bulk of the monitoring and evaluation
activities for a given policy response. Monitoring will involve continuous assessment that
generates detailed information on the progress of the policy response. Evaluation will involve
a systematic study of the effectiveness, efficiency and impact of the policy response. Both of
these activities will be used to inform decisions to continue, amend or conclude the
implementation of the policy response. Evaluation will also be used to gather lessons to
inform future policy design.
5
We are keen to brainstorm your key questions and issues about evaluation,
data and the Priority Investment Approach for use in our issues register
Evaluation and the Priority
Investment Approach
Your questions and issues
•
•
Evaluation using the
What
questions about the Priority Investment Approach
Priority Investment
need to be resolved to help your team to move forward?
Approach: using the
• What are the chief
issues and risks facing evaluation using
tools and administrative
the Priority Investment Approach, and evaluation of the TTL
data of the Approach to
Fund and its interventions?
help to evaluate
• Who wil be the key
stakeholders to resolve these issues,
programs
risks and questions?
•
Evaluation of the Try,
Test and Learn (TTL)
Our discussion questions
Fund:
•
Stakeholders: what role wil DEX play in evaluation under
o
Evaluating the TTL
the Priority Investment Approach? How wil other
Fund as an overal
stakeholders be involved?
measure
•
Evaluation processes: what wil we need to make
o
Evaluating the
evaluation
of TTL interventions and evaluation
using the
separate
Priority Investment Approach a success?
interventions
•
Access to data: what wil be the processes for access to the
funded under the
data and tools of the Priority Investment Approach? What
TTL Fund
data governance / management issues arise?
9
Try, Test and Learn Fund
s 22
The Government announced the $96 million Try, Test and Learn (TTL) Fund in the 2016-17
Budget. The TTL Fund will finance new or innovative policy responses that seek to support
the priority groups identified through the actuarial analysis, and the goals of the Priority
Investment Approach. The TTL Fund will also seek to foster collaborative and data-driven
policy making across government and with external stakeholders.
The TTL Fund will enable the development, implementation and tailoring of innovative policy
responses aimed at addressing individuals’ barriers to participation and supporting people
with the capacity to work to do so. The policy responses will be evaluated to assess their
efficacy, with the results being used to transform and inform our investment in existing
programs or make the case for investment in new policy responses. More information on
how policy responses will be evaluated is provided below.
s 22
Evaluation of Try, Test and Learn Policy Responses
Rigorous monitoring and evaluation of Try, Test and Learn Fund policy responses will be
very important. It will allow informed decisions to be made about the continuation of policy
responses financed by the Fund. It will also ensure that the Fund generates high quality
10
evidence on what works, how, and for whom. This evidence can then be used to change and
adapt existing programs and to make the case for new policy responses.
Try, Test and Learn Fund policy responses will be evaluated using a range of evaluation
approaches depending on the nature of each response. The first step for each policy
response will be working out how the Priority Investment Approach model and datasets can
contribute to its evaluation. Other evaluation methods will then be incorporated to support
the Priority Investment Approach analysis and create a complete picture of the effectiveness
of a particular policy response, and how and why it is effective or ineffective.
Priority Investment Approach model and datasets
The Priority Investment Approach broadly entails two tools which can be used in the
evaluation of policy responses:
• The actuarial model that generates the Priority Investment Approach annual
valuations. This model simulates people’s future life trajectories and interactions with
the payment system.
• The longitudinal administrative dataset which contributed to the development of the
actuarial model. This is a quarterly longitudinal social security dataset based on
administrative data from the welfare payment system, and currently spans the period
2001 to 2015.
The capacity of these tools to evaluate the impacts of Try, Test and Learn Fund policy
responses will continue to increase as they are developed over the coming years.
Longitudinal administrative dataset
The longitudinal dataset of social security administrative data has the potential to be used to
evaluate policy responses in several ways. Any of the indicators in the dataset which were
relevant to the impacts of a given policy response could be analysed to support the
response’s evaluation. The indicators in the dataset include, for example, payment receipt,
payment rate, earnings, birth of children, caring responsibilities, experience of domestic
violence and disability or mental health conditions.
A range of methods could be used to analyse the indicators described above, including:
• Quasi-experimental studies of participant impacts: longitudinal analysis of
participants or trial locations compared to analysis of control groups or locations.
Control groups or locations would be matched to the participant group or location.
• Randomised control trials: random assignment of individuals or communities to
participation in a policy response. Longitudinal analysis of these participating
individuals or communities with control groups. Control groups would be matched to
the participant group or locations.
Actuarial model
The other Priority Investment Approach tool available to support the evaluation of policy
responses is the actuarial model. The model can be used to understand the long term
11
impact of decisions made today and in the future at a fairly high 'system' level. The baseline
valuation model provides a good foundation for accurately projecting future lifetime costs at
the whole population level and for particular subgroups.
However, like all statistical models, the ability of the actuarial model to differentiate average
future lifetime costs between different people is limited by the extent of the modelling
variables included in the model. That is, the groups considered by the model can only be as
specific as the characteristics provided by the modelling variables. As more modelling
variables are included, the model will be able to provide future lifetime costs for more refined
groups.
The baseline model used a limited number of modelling variables to determine people's
future lifetime costs and pathways. These modelling variables included welfare class, age,
gender, partner status, number of dependent children, highest level of education attained,
duration in the welfare system, age entered welfare system, and so on. The 2016 valuation
will include additional modelling variables, such as earnings indicator and DSP medical
condition type.
Other evaluation methods
Other evaluation methods and data sources will support evaluation with the actuarial model
and dataset to ensure a complete picture of both the impact of policy responses, and also
how and why they are effective or ineffective. Supporting analysis will use both existing data
sources (e.g. Data Exchange, program or service data) and customised data collection such
as surveys, interviews or action research methods.
Using a rich, mixed-methods approach to evaluation will ensure that we generate detailed
understanding of the impact Try, Test and Learn Fund responses are having on individuals
and communities and have robust evidence on which to base future policy decisions.
12
Measuring success –Try, Test and Learn Fund outcomes and evaluation
The Try, Test and Learn Fund will support targeted trials of policy responses for defined priority groups of
several thousand people. While these trials may have significant impacts on the target groups, it is unlikely
these will be visible in the annual population wide valuations.
Instead, the outcomes of Try, Test and Learn Fund policy responses will be measured in tailored analysis and
evaluations conducted internal y and by external independent evaluators.
The measures of success will differ for each individual Try, Test and Learn Fund response and the barriers to
work they are addressing. Further, the approach for the evaluation of policy responses will be designed with
stakeholders as part of the overall consultative design process for the Fund.
Overall, the success of the Try, Test and Learn Fund will be tied to achieving the goals of the Priority
Investment Approach, which entail improving the lifetime wellbeing of people and families in Australia by:
• Increasing the capacity of people to live independently of welfare
• Decreasing the Commonwealth’s long-term social security liabilities
• Reducing the propensity for intergenerational welfare dependence.
The success of the Try, Test and Learn Fund will be measured against the following outcomes:
• Are priority groups participating in work or education at increased levels?
• Are priority groups less reliant on welfare, for example are they receiving part rather than full rates
of payments?
• Are priority groups finding work and exiting the welfare system sooner?
• Are the average future lifetime costs for priority groups decreasing?
Indicators of these outcomes can be tracked in the Priority Investment Approach model and dataset. The key
indicators will be:
• Number of people on payment
• Payment transitions
• Earnings from work
• Payment rate – full rate or part rate due to income from work
• Participation in education
• Average future lifetime costs (the model may take two or more years to reflect policy changes in
future costs).
For individual priority groups and responses there will be additional specific outcomes.
For example, for young carers a key outcome will be the proportions of young carers who are participating in
training or education, and who leave the system once their caring role ceases rather than transitioning onto
an unemployment payment.
In the long-term, we will also be able to track the outcomes of people’s children to measure whether the risk
of intergenerational welfare dependence has changed.
The Priority Investment Approach model and dataset will be the first port of call for measurement of the
outcomes.
We will measure outcomes by comparing indicators for people participating in a new policy or support to a
similar group of people not participating. This allows us to see and measure the impact the policy is having.
13
In some cases we will use the Priority Investment Approach data to conduct Randomised Control Trial
designs. In other cases we will find control groups to match the group or location that has already been
select to participate in a policy trial.
Importantly, we will also use other data sources and evaluation methods to ensure a complete picture of
what’s working and how. Surveys, post-participation monitoring and qualitative data and customer focus
groups or interviews will be utilised where appropriate to complement econometric, statistical and other
quantitative techniques.
Outcomes that may be captured with additional analysis include priority groups’ capacity for work,
increasing work skills and knowledge, wellbeing, social inclusion, financial stress and family functioning.
This will support the Priority Investment Approach’s important goal of improving people’s lives, not just
reducing payment recipient numbers or costs.
14
s 22
Monitoring and evaluation design
Evaluation across the TTL Fund policy cycle
Evaluation of TTL Fund policy responses will involve a systematic study of the effectiveness,
efficiency and impact of the policy response. As outlined in the sections above, evaluation will be
considered at all stages of the TTL Fund process as summarised below:
● during the
idea selection process, evaluation readiness will be used as an assessment
criterion to ensure al ideas that are selected have the potential to be effectively and
rigorously evaluated
● during the
pre-market co-development stage, adjustments may be made to an initial idea to
increase the efficiency and effectiveness of evaluation or allow for sophisticated
methodologies such as randomised controlled trials. The general approach to evaluation
would be established at this time
● during the
selection process service providers will be sought who have the capacity to
support evaluation during the implementation of a program or project
6 It is important to note that while the agility of funding is a key strength of the TTL Fund, it wil be critical to
give service providers security to deliver initiatives, particularly larger initiatives, confidently. Public messaging
wil consider this balance between agility and security careful y.
7 As outlined earlier in this paper, the suitability of a program being trial ed and ceased would be considered in
the assessment process. For example, a program that involved large investments up front (e.g. in funding,
delivery partnerships/relationships, participants’ time etc.) may be less suitable for the TTL Fund due to the
difficulty in ceasing these programs if they are not working effectively.
15
● during the
co-development stage, DSS will work with the service provider and any external
evaluation providers to develop and plan for the evaluation and conduct any baseline data
collection required
● during the
implementation stage, data from monitoring and evaluation activities will be
used to assess performance against outcomes and KPIs and inform decision making at stop-
go points
●
after the implementation stage:
o where feasible, post-participation evaluation would be undertaken to ascertain the long-
term effect of an initiative. This may also include supporting evaluation continuity if a trial
is scaled up outside the TTL Fund
o final evaluation results will be publically shared and added to a knowledge store of
evidence gathered via the Try, Test and Learn Fund policy responses. Final evaluation
results may also be used by Commonwealth agencies, service providers or other
stakeholders to seek other funding sources to continue a policy trialled through the Try,
Test and Learn Fund.
Overarching measures of success
The outcomes used to measure success wil be guided by the goals of the Priority Investment
Approach. The specific outcomes will differ for each policy response and the barriers to work that it
seeks to address. They wil involve questions such as:
● are priority groups participating in work or education at increased levels?
● are priority groups less reliant on welfare (e.g. receiving lower rates of payment)?
● are priority groups finding work and exiting the welfare system sooner?
● are the average future lifetime costs for priority groups decreasing?
Evaluation data sources and methods
While TTL Fund policy responses may have significant impacts on the target groups, it is unlikely that
these impacts will be visible in annual Priority Investment Approach population-wide valuations.
Instead, the outcomes wil be measured through tailored analysis and evaluations conducted
internally and by external, independent evaluators.
The Priority Investment Approach model and dataset will be the first data source considered in the
monitoring and evaluation design for a policy response. The questions above may be able to be
tracked using indicators in this model and dataset, such as the number of people on payment,
payment transitions, or average future lifetime costs (noting that the influence of policy changes on
these costs may not be reflected in the model for two or more years).
Outcomes will be measured by comparing indicators for people participating in the policy response
with a similar group of people not participating. This approach allows us to measure the impact that
the policy is having. The methods used to achieve this comparison may include:
16
link to page 14
● randomised controlled trials
8
● quasi-experimental methods
● difference-in-difference analysis.
Importantly, data sources beyond the Priority Investment Approach model and dataset wil be used
to ensure a complete picture of what’s working and how. Surveys, post-participation monitoring,
qualitative data and customer focus groups or interviews will be utilised where appropriate to
complement econometric, statistical and other quantitative techniques. Outcomes that may be
captured with this additional analysis include priority groups’ capacity for work, increasing work
skills and knowledge, wellbeing, social inclusion, financial stress and family functioning.
The Priority Investment Approach Taskforce is considering the best approach to securing human
research ethics clearance for evaluations conducted under the Try, Test and Learn Fund, and could
consider using an existing ethics committee such as the AIFS Ethics Committee or the Australian
Health Ethics Committee.
s 22
17
Document Outline
- 2. Evaluation for MO meeting on TTL Fund 27 Sept 2016 (2).pdf
- 3. Investment Approach Issues Register - Issues by responsibility.pdf
- Issues summary
- Issues with detail
- Issues by responsibility
- 4. Australian Priority Investment Approach Evaluation Responsibilities.pdf
- Australian Priority Investment Approach: Evaluation Responsibilities
- Purpose and scope
- Roles and responsibilities
- Coordination
- 5. PowerPoint Presentation.pdf
- 8. Try, Test and Learn Fund Design.pdf
- Try, Test and Learn Fund Design
- Overview
- Objectives and structure
- Design principles
- Design of the TTL Fund
- Ideas generation phase
- Submission and assessment process
- The processes supporting the submission and assessment of policy response ideas will strive for accessibility by supporting input from diverse proposers and minimising red tape. The primary mechanism for proposing ideas will be online written submissi...
- Online written submissions
- Policy Hack
- Implementation and performance reporting
- Outcomes-focused funding agreements
- Decision-making framework for funding decisions
- Monitoring and evaluation design
- Next steps