Most evaluation disappointments start in the Terms of Reference. A vague ToR produces a generic report. This pack walks you through the 9 ingredients, honest costing, and the pre-send checklist -- so the research you commission actually answers your question.
4 modules~100 minInteractiveIndia-context
Your progress
0% complete
Your Capstone
Polished ToR + Costing Sheet
By Module 4 you will have a ready-to-send Terms of Reference with all 9 ingredients, an honest budget, and a quality-control checklist. The capstone compiles your module answers into a single document.
Module 1 -- ~25 min
The 9 ingredients of a good ToR
A ToR is a contract between a commissioner and a researcher. Most bad evaluations trace to a missing or vague ingredient in the ToR -- not to bad fieldwork. If you get these nine right, even an average research firm will deliver something useful.
The 9 ingredients
Background and context -- what programme, where, how long, key facts. Not a marketing summary; an honest operational description.
Purpose and intended use -- who will use the findings, for what decision, by what date. "For learning" is not a purpose.
Research questions -- 3-5 specific, answerable questions. Each should begin with "To what extent..." or "How does..." or "What factors..."
Scope and limitations -- what is in, what is out, what you already know the study cannot answer.
Methodology expectations -- you need not prescribe the method, but name the design family (qualitative, quantitative, mixed) and any non-negotiables (e.g., "must include direct beneficiary voices").
Deliverables and timeline -- inception report, draft report, final report, presentation. Attach dates.
Team composition and qualifications -- sector expertise, language requirements, gender composition for fieldwork, prior India experience.
Budget indication or ceiling -- giving a range ("INR 12-15 lakh") attracts realistic bids. Hiding the budget attracts inflated ones.
Ethics and data management -- IRB expectations, consent protocols, data ownership, GDPR where it applies, and readiness for the DPDP Act 2023 (its core duties apply from 13 May 2027).
Worked example (fictional organisation, hypothetical figures)
Illustrative case: Vachanveer Trust, a fictional education NGO, needs a midline evaluation of its reading programme in Rajasthan. Budget: INR 18L. Timeline: report needed by March 2027 for board renewal decision.
Ingredient 2 (Purpose): "The findings will be presented to the Vachanveer Trust Board in April 2027 to inform whether to continue, modify, or discontinue the Rajasthan cluster model. The primary audience is the Board and State Director."
Ingredient 3 (Questions): (1) To what extent have foundational literacy outcomes improved in intervention villages vs. comparison villages? (2) What programme components do field staff identify as most/least effective? (3) What is the per-child cost of the current delivery model?
Your ToR Ingredients Sheet
Draft each ingredient for your real study. These save automatically and flow into the capstone.
Programme name, geography, duration, scale, key facts an outsider needs
Who uses findings, for what decision, by when?
Start each with "To what extent...", "How does...", or "What factors..."
Saved
Self-check
A ToR states: "The purpose of this evaluation is learning and accountability." What is wrong with this?
Nothing -- it covers both bases
It names no specific decision, audience, or deadline -- so the researcher cannot prioritise
It should say "impact" instead of "learning"
The word "accountability" is inappropriate in a ToR
Correct. "Learning and accountability" is a category, not a purpose. A useful purpose statement names who will read the report, what decision it informs, and by when. Without this, the researcher optimises for the wrong audience.
Module 2 -- ~25 min
Choosing the right research type
The most common ToR failure: asking for "an impact evaluation" when all you need is a process review. Or asking for "a baseline" when you actually need a situational analysis. The mismatch wastes money and produces reports nobody acts on.
Research type decision tree
If you want to know...
You need...
Illustrative cost (India, 2026)
Did the programme cause change?
Impact evaluation (RCT, quasi-exp, DiD)
INR 25-80L
What changed, but not necessarily why?
Outcome evaluation (pre-post, comparison)
INR 12-30L
How is the programme being delivered?
Process evaluation / implementation review
INR 6-15L
What do stakeholders think and experience?
Qualitative study (FGDs, interviews, case studies)
INR 5-12L
What exists before we start?
Baseline / situational analysis
INR 8-20L
Is the programme theory sound?
Evaluability assessment / ToC review
INR 4-8L
Common mismatches
"Impact evaluation" with INR 10L budget -- impossible for credible causal claims. Re-scope to outcome evaluation or process review.
"Baseline" for a programme already 2 years in -- too late. Call it a midline situational analysis.
"Mixed-methods" without specifying integration -- you get two parallel reports, not one mixed study. Specify how strands will integrate.
The evaluability question
Before commissioning a full evaluation, ask: is this programme evaluable? Does it have a clear theory of change, measurable outcomes, sufficient implementation fidelity, and enough time elapsed for effects to show? If no, commission an evaluability assessment first (illustratively INR 4-8L, 6-8 weeks). It saves you from a INR 30L study that produces "insufficient data to conclude."
Your Research Type Decision
Match your need to the right type. This feeds into your ToR methodology section.
Design family, non-negotiables, what you leave open for the researcher to propose
Theory of change clear? Outcomes measurable? Enough time elapsed? Data systems in place?
Saved
Self-check
An NGO running a 2-year nutrition programme in Jharkhand has INR 10L for evaluation. They want to claim "impact." What should you advise?
Proceed with an RCT -- it is the gold standard
Skip evaluation entirely -- INR 10L is too little
Re-scope to outcome evaluation with honest limitation statement about causal claims
Do a desk review only
Correct. INR 10L cannot support a credible impact evaluation (no RCT, no strong quasi-experimental design at that budget). An outcome evaluation with pre-post comparison is feasible and honest. The ToR should explicitly state "this study will document outcomes but cannot make causal claims."