The Bradford Evaluation Framework

Levels of Evaluation

45 minutes

Monitoring

Purpose

Supports regular review of service delivery

Relates to the inputs, activities and outputs of the logic model only

Provides opportunity to identify challenges around data capture and implementation early, and put support in place

What it will tell us

Data quality and suitability

Performance against key indicators

Potential evaluability (see evaluability checklist)

What it won’t tell us

Whether the intervention is effective

Data

Basic delivery data

Method

Descriptive statistics

Suggested Outputs & Timing

Quarterly or termly reporting, but will depend on delivery patterns/length

What you need

Consent routinely collected

System for data capture

Agreed performance indicators

Implementation

Purpose

Establishes whether the intervention can be delivered as set out in Service Design

Relates to the inputs, activities and outputs of the logic model only

Informs future evaluations

What it will tell us

Fidelity to the model, which may include recruitment, reach/engagement, satisfaction

Possible reasons why the intervention is or is no’t performing as expected

What future evaluation might be possible (see evaluability checklist)

What it won’t tell us

Whether the intervention is effective

Data

Basic delivery data

As required:

Satisfaction questionnaires

Additional questionnaires

Interviews / focus groups* Observations*

Method

Descriptive statistics

*Qualitative analysis

Suggested Outputs & Timing

Annual or by contracted periods

What you need

As per monitoring

As required:

Access to intervention for observations

Access to staff and participants for interviews

Before and After

Purpose

Establishes whether participants self-report changes in outcomes after participation in a service

Relates to short term outcomes of the logic model

What it will tell us

An estimate of change in outcome(s) after participation in the service

This may provide Early Intervention Foundation (EIF) level 2 evidence

What it won’t tell us

Whether the service is effective

Data

Standard data set out in data requirements 

Outcome measures collected at the start and end of participation

Method

Basic statistical analyses comparing results before and after the service

Suggested Outputs & Timing

Report before the end of the contract period

What you need

As per monitoring;

Logic model and service design are suitably defined.

Use of reliable, valid and relevant outcome measures taken before and after participation in the service

Effectiveness

Purpose

Provides evidence of effectiveness of the service

Relates to short, medium and long-term outcomes of the logic model

What it will tell us

An estimate and interpretation of the strength and direction of the effect(s) of an intervention on an outcome

Assessment of certainty of the results

This may provide EIF level 3 evidence

What it won’t tell us

Effects on those not reached

Effects beyond the evaluation period

Effects on outcomes not measured

Data

Baseline & Outcome data

As required:

Additional questionnaires

Interviews / focus groups* Observations*

Comparison group

Method

Comparison of outcome(s) between participants and control group; may use randomised control trials or quasi-experimental methods

Suggested Outputs & Timing

Depends on scale of intervention, level of participation, and timing of outcome measures

What you need

Intervention EIF rating of 2

Sufficient number of participants

An identifiable control group

Background studies

Purpose

Provide context to evaluations and inform future evaluations

What it will tell us

Depends on study aim and design

What it won’t tell us

As required:

Questionnaires

Interviews / focus groups* Observations*

Research data

Data
Method

Various; depends on study aim and design

Suggested Outputs & Timing

Depends on study aim and design

What you need

As required:

Access to services for observations

Access to staff and participants for interviews

Willingness of service team to facilitate extra study