+ Conceptualizing Influence and Impact in Development Research Katie Wright.

+

Conceptualizing Influence and Impact in Development Research

Katie Wright

+Outline and Key Questions 1. How has research

influence been defined in development research?

2. What are the complexities?

3. Are we interested in someone using the research or a desired outcome of change (e.g. fewer poor people)?

4. How can policy influence be captured?

5. What are the methodological complexities?

6. Conclusions

+

1. How has Research Influence been Conceptualised in Development Research?

+2. Conceptualising Influence in Development Research

(i) Outputs produced by research?

(ii) Impact processes (how research outputs are used)?

(iii) Impacts per se (initial consequence of research use in various decision arenas)?

(iv) Outcomes (the subsequent consequences of changes in decision arenas)?

(Davies et al, 2005)

+

3. Defining Research Impact/ Influence

Various terms used inter-changeably – uptake/use versus outcomes/change

Influence what and whom? – i.e. what modalities and audiences – a change in thinking or doing?

Different models of understanding research influence have implicit assumptions (Linear/Incremental/Interactive)

+4. Too linear?

+5. Still too linear?

6. Why do some ideas start epidemics?

Content and Processes ‘sticky messages’

Connectors andChampions

‘the law of the few’

Context and Opportunism‘the power of context’

Molas Gallart + Tang?

+7. A ‘layered’ model of policy influence in agriculture (Sumner and Rulmi, 2009)

+

8. How can Research Influence be Captured? Methodological Complexities

+9. How to assess research influence/impact?

How to track subtle changes (e.g. changes to people’s behaviours, attitudes and understanding of social issues)?

This type of impact is a lot more common than direct impact with ideas ‘creeping‘ into policy deliberation and is very difficult to measure (Davies et al 2005, Coe at el 2002).

+ 10. How to assess those who have been influenced?

Where do you look for impacts? Who are the users?

May be completely different from those anticipated by the project designers!

E.g. When looking at policy makers it may be difficult to discern who the key people are….

(Davies et al 2005, Ryan & Garrett 2003)

+ 11. How to decide when to look for influence/impacts?

It can be difficult to decide when it is best to evaluate impact.

Most monitoring and evaluation happens soon after project completion - may not find evidence of impact. (Davies et al 2005, Carden 2004, Ryan & Garrett 2003)

If an evaluation is carried out shortly after project completion & discovers little impact, does this mean that there will be no impact?

+12. How to assess attribution or contribution or the counter-factual?

How do you assess the contribution of research to a certain outcome? Was it the key driver?

Different factors support ongoing change – how to evaluate the input of only one of these factors?

Impacts of individual projects/ collective?

What would have happened if the research had not been carried out, how do you measure this? (Carden 2004, White 2006)

+13. Should assessment be qualitative, quantitative, or both?

How to balance qualitative descriptions with quantitative assessments, considering both perceptual and objective judgements? (Davies et al 2005)

+14. At what level of influence should investigation of influence take place?

What is the focus: Institutional, programmatic, thematic or project level? (Ryan & Garrett 2003)

+15.What is a good outcome?

In order to see if poverty and inequality has been reduced, we need to devise appropriate measures (Meinzen-Dick et al. 2004: 1)

+16. Conclusions

Research impact and influence have been defined in a multiplicity of ways.

Assessing impact of our research requires understanding the comparative strengths and weaknesses of various conceptual models (all with underlying assumptions) which seek to capture different types of research impact and the process by which this occurs.

Assessing types of impact this may involve use of different methodologies some which are more linear than others including work by IDRC (Earl et al 2001, Hovland 2007 Carden 2005), ESRC (Molas-Gallart & Tang 2007; Wooding et al 2007), IFPRI and others.

+17. Conclusions Tracking impact may include different

combinations of approaches. (e.g. backwards-tracking, post-research tracing, or case studies).

Developing impact indicators across different domains of impact with attention to indirect and unintended impacts.

Understanding the processes through which research influence occurs requires consideration of the factors that enhance research/user interaction or ‘take-up’.

Critical to this is sensitivity to context as well as understanding of how different stakeholders ‘absorb’ ideas and the politics of whose ideas ‘count’.