
Evaluation Data Synthesis for Program Impact Analysis
Synthesize mixed-method program data into credible, bias-checked impact findings
What You Can Do
Organize disparate data streams from surveys, interviews, admin records, and observations into a cohesive analysis framework. Cross-validate findings across quantitative and qualitative methods to identify robust patterns and surface anomalies. Generate documented, bias-checked conclusions aligned to your theory of change that stakeholders can confidently act on.
Features
systematically combine surveys, interviews, administrative data, and observation notes into unified analysis structure
compare findings across quantitative and qualitative data to identify robust patterns and strengthen evidence credibility
stress-test conclusions for common M&E interpretation errors (causality claims, selection bias, confounding variables)
organize messy real-world data into defensible outcome metrics aligned to your logic model
surface trends, anomalies, and surprising results that warrant deeper investigation
structure findings by outcome domain for board presentations and funder reports
produce audit trails showing how conclusions were reached from source data
clearly distinguish findings with high confidence from preliminary insights needing further validation
Example Output
Input: Survey responses (n=145), interview transcripts (8 participants), monthly attendance records, budget utilization data
Output:
Program Effectiveness Summary
Primary Outcome: Job Placement Rate
- Quantitative finding: 73% of program completers placed in employment within 6 months (n=92)
- Qualitative validation: Interview participants cited improved resume and interview skills; 2 reported placement barriers unrelated to skill gaps
- Robustness rating: ⭐⭐⭐⭐ High — triangulated across admin records and qualitative feedback
Secondary Outcome: Financial Stability
- Survey respondents reported average 31% income increase post-program
- Administrative data shows participant cost-per-outcome: $1,200
- Selection bias note: Early exiters (n=23) excluded from outcome data; recommend follow-up analysis
- Robustness rating: ⭐⭐⭐ Moderate — pending confirmation on comparison group
Unexpected Finding: Participants with prior work experience showed lower placement rates (62% vs. 79% for first-time workforce entrants). Qualitative probing suggests mismatch between program content and experienced worker needs.
What's Included
- SKILL.md instruction file: complete mixed-methods analysis protocol and synthesis workflow
- Data organization template: structured format for consolidating survey, interview, and administrative data sources
- Cross-validation matrix: framework for comparing findings across quantitative and qualitative methods
- Bias detection checklist: common M&E interpretation errors and validity questions to ask of each conclusion
- Impact narrative template: outcome-domain structure for donor reports and board presentations
Who It's For
- Monitoring & Evaluation Specialists managing complex multi-source datasets and synthesizing program performance
- Nonprofit Program Directors preparing evaluation reports and evidence of impact for funders
- Grant Writers building credible outcome narratives for donor proposals and progress reports
- Social Impact Researchers analyzing mixed-method program data for rigor and validity
- Evaluation Consultants conducting external program assessments and producing defensible conclusions
Best For
- Synthesizing findings from mixed-method evaluations (quantitative + qualitative + administrative data)
- Cross-validating preliminary conclusions before finalizing evaluation reports
- Extracting outcome indicators from messy real-world data with statistical integrity checks
- Stress-testing impact narratives for bias, confounding variables, and causality claims
- Building outcome-aligned findings for board presentations and funder debrief meetings







