an analysis of influence of evaluation in the context of socio-economic development plan (sedp)...
TRANSCRIPT
An analysis of influence of evaluation in the context of Socio-Economic Development Plan (SEDP) implementation in Vietnam - The case of Hai lang district, Quang tri province
Ha Minh TriHCMC, 2 April 2015
Outline (1/2)
1. Introduction 1.1 The problem statement1.2 The research context1.3 The purpose of the study1.4 Overview of research steps
2. Literature review2.1 Literature review on evaluation use & influence2.2 Conceptual framework2.3 Research questions
3. Case study setting (Hai Lang district)
2
Outline (2/2)
4. Research methodology4.1 Research design4.2 Research instruments & data collection4.3 Hypotheses4.4 Limitations
5. Analysis and findings5.1 Data analysis5.2 Findings5.3 Construct validity and reliability
6. Research and policy implications6.1 Contributions to theory6.2 Implications for future research & policy
3
1. Introduction
4
1.1 The problem statement
• Importance of accountability and learning.• Accountability = challenge for public sector management.• Monitoring and evaluation (M&E) findings underused to serve
SEDP targets and objectives.• Empirical studies on evaluation influence: none in the public
sector Vietnam.
5
1.2 Research context – Monitoring & Evaluation in Vietnam
• Inspection system still limited to cope with state management requirements.
• M&E system for Poverty Reduction Strategy Paper (PRSP) and SEDP at central level switching from traditional to evidence-based. This shift is at a slower speed at district and commune levels.
6
1.3 Purpose of the study
- To identify and explain factors which may affect evaluation influence in the public sector of Hai Lang.
7
1.4 Overview of research steps
Developing specific research questions + hypotheses
Conducting literature review: to contextualise the research
Exploring research setting: M&E in Vietnam
Developing a conceptual framework based on Theory of Evaluation Influence
Research design & methodology: Mixed Methods Explanatory Design: (1) QUAN methodology uses EFA, CFA and Structural Path Model and (2) QUAL study is used to explain QUAN results.
Developing research instruments & collecting data
Empirical research & data analysis, presenting findings, analysis & discussion (with QUAL analysis enhancing QUAN findings)
Conclusions & implications
Start with problem statement
8
2. Literature review
9
2.1 Literature review – Main findings on evaluation use (1/6)
• Evaluation that was underused in the US in late 1960s stimulated research on evaluation use.
• Literature on evaluation use emerged since 1970s.• Five most common types of evaluation use: instrumental,
conceptual, persuasive, process and imposed use.
10
2.1 Factors affecting evaluation use (2/6)
11
7 factors under evaluation impl.
6 factors under decision/policy setting
9 factors under stakeholder involvement
Evaluation quality Information needs Involvement and information needs
Credibility Decision characteristics Involvement and credibility
Relevance Political climate Involvement and relevance
Communication quality
Competing information Involvement and communication quality
Findings Personal characteristics Involvement and findings
Timeliness Commitment /receptiveness to evaluation
Involvement and commitment/receptiveness to evaluation
Evaluator competence
Involvement and personal characteristics
Involvement & decision characteristics
Direct stakeholder involvement
2.1 Main findings on evaluation influence (3/6)
• Evaluation influence appeared since 2000s as scope of evaluation use was limited.
• Evaluation influence evolved with 2 remarkable models/theories: (1) Integrated model of influence (Kirkhart, 2000), and (2) Theory of Evaluation Influence (Mark and Henry, 2004)
12
2.1 Mark and Henry’s Theory of Evaluation Influence (2004)
Evaluation context*· Expertise· Communication· Instruction· Time· Resources· Role flexibility
Decision/policy setting*· Administrative
support· Micro politics· Culture· Information needs· Impetus· Skills
Evaluation activitiesEvaluation inputs
Attributes of:· Stakeholder
selection and participation
· Evaluation planning and design
· Data collection and analysis
· Developing conclusions and recommendations
· Report generation· Information
dissemination
Evaluation “Outputs”
Knowledge attributes*· Responsiveness· Credibility· Sophistication· Communication· Timeliness
General mechanisms
· Elaboration· Heuristics· Priming· Salience· Skill acquisition· Persuation· Justification· Minority-opinion· Policy consideration· Standard setting· Policy discussion and
deliberation· Coalition formation
Intermediate and long-term outcomes
Cognitive/affective· Salience· Opinion valence· Descriptive norms· Agenda setting
Motivational· Personal goals· Social reward· Incentives· Market forces
Behavioural· Individual practice· Collaborative practice· Programme
continuation. Termination or expansion
· Policy adoption
Social betterment
Contingencies in the environment:· Competing processes· Facilitating factors· Inhibiting factors
2.1 Where is this dissertation positioned? (6/6)
Evaluation use (since 1970s)
Instrumental use
Conceptual use
Persuasive use
Process use
Imposed use
Evaluation influence (since 2000s)
Integrated model of influence (Kirkhart, 2000)
Theory of Evaluation Influence (Mark and Henry, 2004)
Evaluation influence in Vietnam public sector
14
This work
2.2 Conceptual framework (1/2)
Evaluation Inputs Evaluation Activities Evaluation Influence
Decision/Policy setting•Evaluation plans & methods
• Involvement of programme holders in evaluation processes
• Generation of evaluation reports
• Dissemination of evaluation reports/findings
• Evaluation influence at individual level
Contingencies in environment•Facilitating factors: e.g. willingness to cooperate, etc.•Inhibiting factors: e.g. wrong time, time pressure, etc.
Source: Developed from Mark & Henry (2004).
15
Evaluation context•Lessons learned/ best practices in evaluation•Evaluation partners•Evaluation capacity
2.2 Conceptual framework (2/2)
• Theory-based evaluation approach was used to elaborate the conceptual framework to establish a theory of change of evaluation influence framework.
• This theory of change was used as a guide to develop a research instrument.
16
2.3 Research questions
(1) What are the factors which may affect evaluation influence at district and commune levels of Hai Lang district as perceived by the staff involved?
(2) How do the identified factors impact on evaluation influence at district and commune levels of Hai Lang district?
17
Two central questions guided the study:
3. Case study setting
18
3. Case study setting (1/2)
19
3. Case study setting (2/2)
• Consisting of 21 communes and township.• A typical rural, in-land district, not a border, remote or
mountainous district.• Not a district with a high percentage of ethnic minorities.• Supported by a Finnish-funded Rural Development
Programme for 13 years (1997-2009) in 3 phases. Phase I (1997-2000) for 14/21 communes and township. Phases II & III for all communes and townships.
• Programme’s main objectives include (1) improving local livelihoods, and (2) strengthening local capacity (participatory planning, M&E, public administration reform, etc.)
20
4. Research methodology
21
4.1 Research design
• The study applied a mixed methods explanatory design• Methodology: QUAN survey + QUAL study
22
QuantitativeData
collectionand analysis
Follow up with
QualitativeData
collection and analysis
Interpretation
Source: Creswell (2012).
4.2 Research instruments & data collection (1/2)
For quantitative survey•Used a 39-item questionnaire measured by a 7-point Likert scales for a sample of 275 participants•Measured 8 dimensions:
(1) Lessons learned/best practices in evaluation(2) Programme partners(3) Evaluation capacity(4) Evaluation plans and methods(5) Involvement of programme holders in evaluation processes(6) Generation of evaluation reports(7) Dissemination of evaluation reports/information to stakeholders(8) Evaluation influence
23
4.2 Research instruments & data collection (2/2)
For qualitative study•Used semi-structured interviews and document review.•Interviewed 55 purposefully selected participants with 12 different characteristics (see next slide).•Documents for review include inspection reports, project/programme proposals, progress and evaluation reports.
24
25
Group code Characteristics # of individuals
Group 1 With longer exposure to Programme 5Group 2 With shorter exposure to Programme 4Group 3 District 4Group 4 Commune 4Group 5 Socio-economic sector 5Group 6 Agriculture, Natural Resources and
Environment sector5
Group 7 Health and Education sector 4Group 8 Administration and Justice sector 5Group 9 Culture and Information sector 5Group 10 College 5Group 11 Manager 4Group 12 Staff 5
Total 55
4.3 HypothesesH1: Involvement of evaluation partners (Factor 1) is positively associated with evaluation influence.H2: Evaluation capacity (Factor 2) is positively associated with evaluation influence.H3: Evaluation plans and methods (Factor 3) is positively associated with evaluation influence.H4: Generation of evaluation report for management (Factor 4) is positively associated with evaluation influence.H5: Generation of evaluation report for partnership (Factor 5) is positively associated with evaluation Influence.H11: Path model (see figure next slide) holds for groups with longer and shorter exposure to Finnish-funded programmes (incl. 5 factors and 5 controls).
26
4.4 Limitations
1. Participants’ challenges to recall effects of evaluation may affect explanations of quantitative results.
2. Lack of a comprehensive database of documents and limited accessibility to documents might influence the explanations and confirmations of quantitative results.
3. Bias in interpretation of qualitative results due to fact that author was previously working in the Programme (1 year).
27
5. Analysis & findings
28
5.1 Quantitative data analysisSteps of analysis
Data entered into computer
Data screening & examination: missing values, outliers, normality
Data analysis: Exploratory factor analysis (EFA) using SPSS
Data analysis: Confirmatory factor analysis (CFA) (Reliability, validity analysis) using Amos
Data analysis: Structural path model testing, interpretation of results using Amos
29
Research question 1: What are the factors which may affect evaluation influence at district and commune levels of Hai Lang district as perceived by the staff involved?
5.2 Quantitative findings (1/3)
• 5 factors resulted from EFA process: (1) Evaluation partners(2) Evaluation capacity(3) Evaluation plans & methods(4) Generation of evaluation report for management, and (5) Generation of evaluation report for partnership
• CFA used to test multi-group measurement model, construct validity & reliability, incl. groups with longer & shorter exposure Finnish-funded programme.
30
• Obtained 3 statistically significant factors: (1) evaluation capacity, (2) evaluation plans & methods, and (3) generation of evaluation report for partnership.
• No significant difference between groups with longer and shorter exposure to Finnish-funded programme.
• None of the control factors (gender, age, years of education, level of responsibility and level of administration) confounds the relationships specified in the model.
5.2 Quantitative findings (2/3)
31
5.2 The path model (3/3)
32
5.3 Reliability & construct validity
Criteria:•Construct reliability (CR) > .70•Construct validity:
Results:•24 out of 28 items satisfied construct validity and reliability criteria
33
Convergent Validity Factor loadings ≥ .50
Average variance extracted (AVE) ≥ .50Discriminant Validity Maximum shared variance (MSV) < AVE
Average shared variance (ASV) < AVE
5.1 Qualitative data analysis
34
Research question 2: How do the identified factors impact on evaluation influence at district and commune levels of Hai Lang district?•Interview data was analysed using Nvivo 8•A visual overview of qualitative data analysis:
Initially reading through text data
Dividing text into segments of information
Labelling segments with codes
Collapsing codes into categories and
themes/findings
Comparing themes/findings and categories
Pages of text
Segment of text
30-40 codes
5-7 themes/findings
Similar or different themes/findings
5.2 Qualitative findings
1. QUAL findings confirm the QUAN findings that both evaluation capacity, and evaluation plans and methods positively affect evaluation influence regardless of the 12 characteristics of interviewees.
2. Generation of evaluation report for partnership positively affects evaluation influence regardless of their college level, level of administration, sectors, groups with longer & shorter exposure to programme.
3. A group of managers perceived no effect of generation of evaluation report for partnership on evaluation influence.
35
6. Research and policy implications
36
6.1 Contributions to theory
1. The studied conceptual framework found valid in explaining evaluation influence in Hai Lang public sector.
2. The study’s results contributed construct measures capturing evaluation influence in Vietnam district public sector.
3. The study’s results provided new affirmation that evaluation capacity, evaluation plans and methods, and generation of evaluation report for partnership positively affect evaluation influence in Hai Lang.
4. Qualitative data contributed to an understanding of evaluation influence in Hai Lang.
37
6.2 Implications for future research (1/2)
1. Replicating this study in other districts with other characteristics than Hai Lang might provide a more comprehensive research setting for better insights of evaluation influence.
2. Research at interpersonal and collective levels might provide a more complete picture of evaluation influence in SEDP implementation in Hai Lang.
3. Longitudinal design might be used to capture perceived changes at different stages of evaluation processes.
38
6.2 Implications for policy (2/2)
1. Maximize significant factors by using programme staff as motivators.
2. Balance accountability and learning purposes in design and reporting of inspection and evaluative exercises.
3. Involve non-governmental actors in M&E processes of SEDP implementation.
4. Enhance evaluation capacity of Ministries of Natural Resources & Environment, and Planning & Investment climate change mitigation to improve use of M&E information.
39
Thank you for your attention!