Customize: Analysis Plan Quality

Decide what to keep, what to fork and what to add, then take away an AI review prompt. Change nothing and you get the published prompt exactly as it stands.

Describe your purpose, work stage, intended users and available evidence in ordinary language. Leave blank to use the populated starting rubric. Missing records remain evidence requests.

Several lets you run one check at a time, and keeps each one short enough to open in an assistant.

Your Prompt

This is the published prompt, unchanged.

Analysis Plan Quality

7,546 characters
Establish the document purpose, design, stage, applicable requirements, and available source evidence before rating. Distinguish N/A with a reason, not evidenced, unresolved conflicting evidence, and observed inadequate performance. Missing records do not prove an activity did not occur. Missing records alone justify evidence requests and a deferred readiness decision, not an inferred failure or mandatory repeat activity. Do not invent documentary, signature or approval requirements. Cite document locations for every finding. Keep supported facts separate from proposals, assumptions, and illustrative examples. Never invent completed fieldwork, validation, consent, approval, values, dates, quotations, or assigned ownership. Label unconfirmed owners and proposed revisions for confirmation. Report blocking ethics, safety, confidentiality, measurement-validity, or inference failures first regardless of scores. Give a total and band only when all five dimensions are applicable and sufficiently evidenced with no unresolved evidence conflict; otherwise show the dimension profile, coverage, and evidence requests without a total, band, or rescaled equivalent. A documentation score never authorizes use, fielding, compliance, or approval.

You are an expert M&E analyst with experience across both quantitative and qualitative analysis. Score the analysis plan section of the document I will provide using the rubric below. The document may be a MEL plan, inception report, evaluation report, contribution analysis, needs assessment, or any document with an analysis plan section.

SCORING RUBRIC - Analysis Plan Quality
Rate each applicable, evidenced dimension 1-5. Choose the highest anchor fully supported by the evidence; a material defect takes precedence over a favorable element count. An absent required feature in a complete submitted artifact is an observed documentation defect; an unavailable source or record is not evidenced, not proof the underlying activity failed.

DIMENSION 1: Analysis Plan-to-Question Mapping
- Score 5: All 4 elements: Each research or evaluation question has a defined analytical approach (no orphan questions). The approach is appropriate to the question type (descriptive, inferential, exploratory, explanatory, normative). The order of analytical steps is logical and the sequence is clear. A mapping table or equivalent makes the question-method link visible to a reader.
- Score 4: Three of four elements adequately specified; the remaining element incomplete: Question-method mapping covers most questions; ordering or visible mapping table partial.
- Score 3: Mapping is implicit. Some questions have approaches; others left undefined. Order of analysis loosely described.
- Score 2: Methods listed without clear mapping to questions. Order absent.
- Score 1: No mapping between questions and analysis. Questions and methods sit in unrelated sections.

DIMENSION 2: Analytical Methods Specification
- Score 5: All 4 elements: Specific methods are named (e.g., "logistic regression with the four covariates listed and clustered standard errors at facility level," not "regression analysis"). For qualitative work, the specific approach is named (thematic, framework, grounded theory, content analysis) with a coding strategy described. Software or tools are named where relevant (Stata, R, NVivo, Dedoose, Excel). Method choice is justified (why this method, not another).
- Score 4: Three of four elements adequately specified; the remaining element incomplete: Methods specific and tools named; justification or coding strategy partial.
- Score 3: Methods named at category level only. Tools mentioned. Coding strategy and justification thin.
- Score 2: Generic method labels ("we will analyze qualitatively"). No tools. No justification.
- Score 1: No methods named or methods incompatible with the data described.

DIMENSION 3: Data Preparation and Integration Plan
- Score 5: All 4 elements: Handling of missing data is specified (rules, thresholds, imputation approach if any, or documented exclusion logic). Outlier or anomaly handling is defined with rules for inclusion or exclusion. Data integration logic is clear when combining sources (how survey and admin data merge, how qualitative and quantitative findings connect). Data transformations are documented (recoding, scaling, derived variables).
- Score 4: Three of four elements adequately specified; the remaining element incomplete: Missing data and outliers handled; integration or transformations partial.
- Score 3: Generic statement of "we will handle missingness." Outliers mentioned. Integration vague. Transformations undocumented.
- Score 2: No missingness rules. No outlier rules. Integration assumed.
- Score 1: No data preparation plan.

DIMENSION 4: Triangulation and Synthesis Logic
- Score 5: The evidence strategy fits the questions and justified method or methods. Necessary checks and handling of contradictions are explicit. Integration or weighting is specified where relevant. A single-method analysis can meet this anchor.
- Score 4: The strategy is appropriate and operational; one nonmaterial check or synthesis detail is incomplete.
- Score 3: The strategy is partly described, but important contradictions or the evidence-to-claim link need clarification.
- Score 2: Necessary checks or synthesis are absent, or integration is asserted without an operational method.
- Score 1: The strategy cannot support the claims or falsely presents unsupported checks as validation.

DIMENSION 5: Quality Checks and Validation
- Score 5: Quality checks fit the analytic tradition and risks, with responsibilities and timing. Sensitivity or alternative-interpretation checks are specified where meaningful or nonapplicability explained. Decisions have an audit trail, and limitations state what cannot be answered. Reflexive qualitative work need not use inter-rater reliability.
- Score 4: Appropriate checks, audit trail, and limitations are substantially specified; one minor operational detail remains.
- Score 3: Some appropriate checks are specified but responsibilities, audit trail, or material limitations are incomplete.
- Score 2: Quality assurance is generic or method-inappropriate; material analytical risks lack a response.
- Score 1: No appropriate checking is specified, or validation is claimed beyond what the procedures can establish.

OUTPUT FORMAT:
Return your assessment as a table followed by a summary:

| Dimension | Score (1-5), N/A, or Not Evidenced | Evidence from Document | Priority Revision |
|-----------|-------------|------------------------|-------------------|
| Analysis Plan-to-Question Mapping | | | |
| Analytical Methods Specification | | | |
| Data Preparation and Integration Plan | | | |
| Triangulation and Synthesis Logic | | | |
| Quality Checks and Validation | | | |

First report blocking issues, unresolved conflicts, applicability and evidence coverage. Only if all five dimensions are applicable and sufficiently evidenced with no unresolved source conflict, report **Total: X/25** and **Band:** Strong (22-25) / Adequate (17-21) / Needs Revision (11-16) / Substantial Revision (5-10), as a secondary documentation summary. Otherwise omit both total and band; report the dimension profile and evidence requests without rescaling.
**Single Most Important Revision:** [One specific sentence]

For any dimension scored 1 or 2, add a brief explanation and a concrete revision example.

DOCUMENT TO SCORE:
[Paste your analysis plan section or full document here]
Open it, with the prompt already in:ChatGPTClaude
Or paste it into:GeminiCopilot

Gemini and Copilot open empty. Copy the prompt, then paste it in.