1. General Scoring Rules
The following rules apply to every indicator across all five categories and must be applied consistently by all evaluators.
1.1 The Scoring Scale
Every indicator is scored on a 0–25 point scale. The five standardised performance bands are:
| Band |
Performance Label |
General Criteria |
| 19–25 |
Excellent performance |
Proactive compliance; exemplary, fully documented record with no significant gaps. |
| 13–18 |
Strong performance |
Consistent evidence with only minor, resolved gaps. |
| 7–12 |
Moderate performance |
Compliance mixed with neutral or weak signals; some gaps present. |
| 1–6 |
Weak performance |
Documented lapses, insufficient evidence, or notable failures on record. |
| 0 |
No evidence / serious violation |
No verifiable evidence provided, active conviction, or confirmed serious violation. |
1.2 Evidence Requirements
Eligible evidence:
Must be public, attributable, timestamped, and sourced from official or credible institutions (EACC, KRA, OAG, Hansard, official gazettes, verified media fact-checkers including Africa Check and PesaCheck, etc.).
Self-reported claims:
Score 0 unless independently verified by a recognised institution or credible third party.
No controversies / no applicable events:
Do not penalise absence. If there are no recorded events against which an indicator can be measured, award a neutral-to-high score unless evidence of avoidance is documented.
Refused or withheld data:
If a candidate or institution actively refuses to provide data that is reasonably available and request-able, treat this as Score 0. This is distinct from structural inapplicability (see N/A Rule below). Refusal must be documented in the audit trail.
Composite indices:
For indicators that aggregate multiple sub-signals, sub-signal weights are visible to evaluators in the Indicator Protocol. The final aggregated score counts as 25% of the subcategory like any other indicator.
1.3 Proxy Allowance — Private Sector Records
Where a candidate has no public office history for a given indicator, a private-sector equivalent may be used as a proxy, subject to all four of the following conditions:
| # |
Condition |
| (a) |
The proxy record has been independently audited by a recognised professional body (ICPAK, LSK, CMA, or equivalent). |
| (b) |
The proxy measures the same underlying construct as the primary indicator (e.g., a corporate audit as a proxy for an OAG audit; ICPAK disciplinary clearance as a proxy for EACC clearance). |
| (c) |
The proxy is scored on the same 0–25 scale using the same band thresholds as the primary indicator. |
| (d) |
The proxy does NOT apply to Category 1 constitutional-specific indicators where the constitutional source is an official government body only (EACC, KRA, Judiciary). These require primary evidence. |
Audit trail requirement: The evaluator must document the original indicator, the proxy document used, and the explicit basis for equivalence. Proxy use must be flagged in the score record.
1.4 N/A Rule — Structural Inapplicability
When N/A applies: An indicator is marked N/A only when it is structurally inapplicable to the candidate — i.e., the indicator presupposes a type of prior role or office the candidate has never held, and no proxy evidence exists.
Redistribution: An N/A score redistributes that indicator's 25-point weight equally across the other three indicators in the same subcategory. Each remaining indicator's maximum therefore increases by 8.33 points.
N/A vs. Refused: N/A is for structural impossibility (the candidate could not have this record). Refused/withheld is for deliberate non-disclosure of evidence that could exist. These must never be conflated.
1.5 Conditional Indicators
Five indicators are conditional — they only apply if the candidate meets the stated prior-role threshold. If the condition is not met, the indicator is automatically marked N/A and the redistribution rule applies.
| ID |
Indicator Name |
Condition for Applicability |
| 1.3.3 |
Legislative or Executive Decision Consistency Record |
Candidate must have held prior legislative or executive office. |
| 2.3.1 |
Audit Clean Opinion Certification |
Candidate must have held a budget-responsible public office. |
| 2.3.4 |
Budget Variance Management and Waste Prevention Index |
Candidate must have had direct budget oversight responsibility. |
| 5.1.1 |
Tenure Completion Rate |
Candidate must have held any prior elected or appointed public office. |
| 5.1.3 |
Crisis Management History and Success Rate |
Candidate must have held an executive, managerial, or command-level office. |
1.6 Demagogue Penalty — Category 3 Negative Trigger
Category 3 (Public Engagement) indicators are subject to a deduction if verified negative triggers are present. These are applied at subcategory level before the final category score is computed.
Hate Speech / Incitement: A verified NCIC formal finding triggers a deduction of up to 10 points from the affected subcategory score. Each finding requires a source reference in the audit trail.
Disinformation: A verified Africa Check, PesaCheck, or VeriVote platform flagged-statement finding triggers a deduction of up to 5 points per verified pattern. Each deduction requires a source reference.
Purpose: This prevents high engagement volume from masking Chapter Six violations. A candidate cannot offset an NCIC finding with high attendance at public forums.
2. The Indicator Protocol Template
Every one of the 80 indicators in the VeriVote Indicator Protocol v2.0 is specified using the following nine-field structure. Evaluators must populate each field for every score they record.
| Field |
Content and Evaluator Obligation |
| 1. Indicator Name |
The canonical name as defined in Protocol v2. Do not paraphrase or abbreviate. |
| 2. Constitutional Anchor |
The specific Article(s) of the Kenya Constitution that ground this indicator. Cited per indicator in Protocol v2. |
| 3. Definition |
The precise scope of what is being measured. Determines what counts as evidence and what does not. |
| 4. Primary Evidence Sources |
The named institutions, document types, and platforms from which evidence must be drawn. Evaluators must cite a source within this list or justify any departure. |
| 5. Measurement Method |
The formula, rate, ratio, or procedure used to convert raw evidence into a score. For composite indices, sub-signal weights are stated here. |
| 6. Scoring Guidance (bands) |
The five-band criteria (Excellent → No evidence/violation) specifying what evidence is required at each level. Evaluators must match the candidate's evidence to the correct band. |
| 7. Bias / Consistency Controls |
Rules that prevent evaluator discretion from distorting scores — e.g., minimum source requirements, what counts as corroboration, excluded evidence types. |
| 8. Evidence Type |
Classification of the evidence used, drawn from the six-type schema defined in Section 3 of this guide. |
| 9. Verification Level |
The 1–5 verification tier assigned to the evidence used, as defined in Section 3 of this guide. |
Where to find indicator definitions
The complete specification for all 80 indicators — including definition, evidence sources, measurement method, scoring bands, and bias controls — is contained in the VeriVote Indicator Protocol v2.0.
This Evaluator Guide does not replicate individual indicator definitions. Evaluators must keep a copy of Protocol v2.0 open when scoring. Any conflict between this guide and Protocol v2.0 is resolved in favour of Protocol v2.0.
3. Evidence Classification Schema
Every score submitted must include an Evidence Type and a Verification Level. These are stored in the database and displayed to voters on candidate scorecards.
3.1 Evidence Types
Assign the single most accurate Evidence Type from the six categories below:
| Evidence Type |
Description |
Kenyan Examples |
| Official Institutional Records |
Direct records from a government body with legal authority to maintain them. |
Court judgments, EACC clearance certificates, KRA compliance certificates, official gazettes, ODPP filings. |
| Independent Oversight Reports |
Reports produced by independent bodies mandated to audit or oversee public institutions. |
OAG Auditor-General reports, PPRA procurement audits, CAJ decisions, Ombudsman reports. |
| Legislative / Government Records |
Parliamentary or executive records documenting decisions, debates, and proceedings. |
Hansard transcripts, committee minutes, executive orders, presidential/governor's assent records. |
| Verified Public Communications |
Official statements, manifestos, or disclosures published through verifiable institutional channels. |
Candidate's official website (archived), IEBC-registered manifesto, official press releases, government portal announcements. |
| Independent Data Sources |
Evidence produced by credible independent organisations with no stake in the candidate's outcome. |
Africa Check fact-check reports, PesaCheck analyses, KNBS statistical publications, academic peer-reviewed studies. |
| Structured Surveys |
Quantitative data from formally designed surveys with documented methodology. Use sparingly; only where no primary source exists. |
Afrobarometer, IPSOS Kenya polling with published methodology, CAJ perception surveys. |
3.2 Verification Levels
Assign the single Verification Level that best reflects the independence and authority of the primary evidence used for a given score:
| Level |
Name |
Standard and Examples |
| Level 1 |
Official Verified Evidence |
Highest tier. Direct records from a government body or court with legal authority. Examples: court judgments, EACC certificates, KRA compliance records, gazette notices. |
| Level 2 |
Independent Institutional Evidence |
Third-party institutional reports from bodies mandated to audit or oversee. Examples: OAG audit reports, EACC clearance reports, PPRA findings, CAJ decisions. |
| Level 3 |
Documented Public Evidence |
Officially published records in the public domain. Examples: Hansard transcripts, gazetted notices, registered manifesto documents, official press releases on institutional portals. |
| Level 4 |
Analytical or Survey Evidence |
Aggregated or modelled evidence. Used sparingly and only where Level 1–3 evidence is unavailable. Examples: fact-check aggregates, perception survey results with documented methodology. |
| Level 5 |
Indirect / Proxy Evidence |
Lowest tier. Inferred or equivalent evidence where no primary source exists. Must document the basis for equivalence. Examples: private-sector audit standing in for OAG audit; corporate disciplinary clearance as EACC proxy. |
3.3 UI Translation for Voter Scorecards
The evidence classifications above are translated into plain-language confidence labels for voters on the VeriVote platform:
| Confidence Label |
Verification Level(s) |
Meaning for Voter |
| Evidence Confidence: High |
Levels 1–2 |
Score is grounded in official government or independent oversight records. Highest reliability. |
| Evidence Confidence: Moderate |
Level 3 |
Score is grounded in documented public records. Reliable but not from a primary government source. |
| Evidence Confidence: Limited |
Levels 4–5 |
Score relies on analytical, survey, or proxy evidence. Confidence is Limited — treat with appropriate caution and review the audit trail. |
3.4 Database Schema — Required Columns per Score Record
Every score record in the VeriVote database must include the following fields in addition to the raw score:
| Column |
Data Type |
Description |
| evidence_type |
ENUM / string |
One of the six Evidence Types defined in Section 3.1. |
| verification_level |
INT (1–5) |
Verification tier assigned per Section 3.2. |
| source_institution |
String |
Named institution that produced the primary evidence (e.g., "OAG", "EACC", "Africa Check"). |
| source_document_ref |
String / URL |
Document title, file reference, or stable URL linking to the evidence. Screenshots must be archived. |
| score_date |
DATE |
Date the score was assigned or last verified. |
| evaluator_notes |
Text |
Brief justification for the band assigned, including any proxy use, N/A declaration, or refusal flag. |
| safeguard_triggered |
BOOLEAN |
Set to TRUE if the Integrity Veto (Step 5) or Category Floor Rule (Step 6) was applied to produce the candidate's final score. The UI reads this flag to determine whether to display a safeguard banner adjacent to the score. |
4. Audit Trail Requirements
Every score must be defensible. The audit trail is the mechanism that makes a score disputable, correctable, and transparent to voters. These requirements are mandatory, not advisory.
Source links or screenshots:
Every score must include either a stable URL to the primary evidence document, a document reference number, or an archived screenshot. Ephemeral links (e.g., social media posts without archiving) are not acceptable as standalone sources.
Brief justification:
The evaluator must record, in plain language, why the evidence warrants the band selected. One to three sentences is sufficient. The justification must address the specific band threshold, not just describe the evidence.
Composite index documentation:
For indicators that use sub-signals (Calculated Fields), the evaluator must record the score awarded to each sub-signal, not just the final aggregated score. This enables independent verification of the composite.
Proxy documentation:
If a private-sector proxy was used (Section 1.3), the evaluator must state: the original indicator, the proxy document, the institution that audited it, and the explicit basis for equivalence.
N/A declaration:
If an indicator is marked N/A, the evaluator must state why the indicator is structurally inapplicable and confirm that no proxy evidence was available or eligible.
Refusal flag:
If a candidate or institution refused to provide evidence, the evaluator must document the date and method of the request, the party from whom evidence was sought, and the nature of the refusal or non-response.
Audit Trail Retention
All audit trail records must be retained for a minimum of five years from the date of scoring, or for the duration of the elected term plus two years, whichever is longer.
Records must be stored in a format accessible to the VeriVote compliance team for independent review. Evaluators cannot unilaterally delete or amend a submitted audit trail entry. Corrections must be logged as a new entry with reference to the original.
5. Newcomer Scoring Rules
Candidates without prior public office history are not penalised by design. The following rules govern how evaluators handle indicators where no public record can exist.
5.1 Category 2 — Institutional Accountability (Pillar-Level Threshold)
Category 2 relies heavily on public office records. For newcomers, a four-stage process applies:
| Stage |
Name |
Evaluator Action |
| Stage 1 |
Proxy Mapping |
Search for private-sector equivalents for each null indicator (e.g., ICPAK-audited corporate accounts as proxy for OAG audit). Score eligible proxies using the same band thresholds. |
| Stage 2 |
Indicator Nullification |
Mark remaining indicators NULL if no proxy exists after Stage 1. |
| Stage 3 |
Threshold Evaluation |
Count NULL indicators in Category 2. If NULL count ≤ 75%: score the pillar normally; remaining indicators fill the category proportionally. |
| Stage 4 |
Structural Redistribution |
If NULL count exceeds 75%: zero Category 2 and redistribute its 25% weight to the remaining four categories using the standard matrix (see How We Calculate Scores v2, Section 4). |
5.2 Category 3 — Public Engagement (Newcomer Proxy)
Candidates without a constituency office are scored using leadership proxies on the same 0–25 scale:
- Community service records (verified by a registered NGO or community organisation)
- Pro-bono work (verified by a professional body such as LSK)
- NGO board engagement (verified by NGO Coordination Board registration records)
- Professional body participation (verified by the relevant professional body)
Evaluators must document which proxy type was used and confirm independent verification of the evidence.
5.3 Category 4 — Policy Development (Newcomer Proxy)
Candidates without a parliamentary record may be scored using the following policy proxy types, each on the same 0–25 scale:
- White Papers / Publications: Peer-reviewed articles, policy reports, or academic publications authored by the candidate.
- Organisational Strategy Records: Strategic plans, restructuring frameworks, or growth strategies the candidate authored or led in a corporate, NGO, or institutional role.
- Advocacy Frameworks: Formal legislative petitions, policy memos, or submissions to government bodies authored by the candidate.
- Professional Thought Leadership: Expert panel participation or published governance commentary, verified by the convening body or publication.
All proxies must be independently verifiable. Evaluators must identify the original indicator, the proxy document, and the basis for equivalence in the audit trail.
6. Worked Examples
The following five examples illustrate correct application of the evaluation template. One example is drawn from each category. Each shows the full nine-field structure as an evaluator would complete it. For the indicator's canonical definition and measurement rule, refer to the corresponding entry in Protocol v2.0.
| Indicator ID |
1.4.4 |
| Constitutional Anchor |
Art. 73(1)(a), 10(2)(a) |
| Definition |
Evidence of publicly recognising and formally rectifying errors, misstatements, or ethical lapses. A Documented Retraction is a formal, publicly traceable correction of a specific previously identified false or erroneous statement. Informal social media posts without an official correction notice do not qualify. |
| Primary Evidence Sources |
Africa Check / PesaCheck retraction logs; VeriVote platform flagged-statement log; official press releases; candidate's institutional website (archived); EACC resolutions. |
| Measurement Method |
Rate = Documented Retractions / Identified Major Errors or Controversies. Normalise by public exposure level (high-profile candidates face more scrutiny; adjust threshold accordingly). |
| Scoring Guidance |
Excellent (19–25): Timely formal retraction for every identified major error; proactive correction with documented follow-up action. Strong (13–18): Formal retraction for most errors; minor delays or one unaddressed minor error. Moderate (7–12): Partial or delayed retractions; some errors publicly identified but not formally addressed. Weak (1–6): Few or no formal retractions despite multiple identified errors. No evidence / violation (0): No corrections despite documented controversies; or active misrepresentation. |
| Bias / Consistency Controls |
Require ≥1 verifiable public record per correction claimed. Neutral score if no controversies are on record — do not penalise absence of errors. Exclude informal social media acknowledgements that lack an official correction notice. |
| Evidence Type |
Independent Data Sources (Africa Check / PesaCheck) — or Official Institutional Records if sourced from EACC resolutions. |
| Verification Level |
Level 2 (Independent Institutional Evidence) if from Africa Check / PesaCheck. Level 1 if from EACC resolutions. |
| Indicator ID |
2.4.3 |
| Constitutional Anchor |
Art. 232(1)(e), 10(2)(c) |
| Definition |
Effectiveness of institutional systems for complaints, oversight, and enforcement. Scored as a Calculated Field using four equally weighted sub-signals. |
| Primary Evidence Sources |
OAG audit reports; oversight body recommendations and implementation status records; public complaint logs; CAJ decisions; disciplinary records. |
| Measurement Method |
Weighted average of four sub-signals: (30%) complaint tracking system existence and resolution rate; (30%) oversight body recommendation follow-up compliance rate; (20%) public reporting transparency score; (20%) documented enforcement evidence. Final score = weighted average normalised to 0–25. |
| Scoring Guidance |
Excellent (19–25): All four sub-signals fully operational with documented evidence; enforcement actions on record. Strong (13–18): Three sub-signals strong; one partial. Moderate (7–12): At least two sub-signals present; partial compliance. Weak (1–6): Minimal or non-functional systems. No evidence / violation (0): No accountability mechanisms in place or documented obstruction of oversight. |
| Bias / Consistency Controls |
Use only official reports and verified logs; exclude perception surveys. Document each sub-signal score separately. If this indicator is N/A (candidate has no public institutional history), apply proxy allowance rules from Section 1.3. |
| Evidence Type |
Independent Oversight Reports (OAG, CAJ) for primary; Official Institutional Records for complaint logs. |
| Verification Level |
Level 2 for OAG / CAJ reports. Level 1 if sourced from official institutional complaint registries. |
| Indicator ID |
3.1.1 |
| Constitutional Anchor |
Art. 10(2)(a), 38(3) |
| Definition |
Frequency and geographic reach of documented two-way public engagement sessions. Office Open Days are the primary session type. Campaign Rallies are explicitly excluded (one-way communication, not two-way dialogue). |
| Primary Evidence Sources |
Official constituency office logs; georeferenced event records with attendance registers; meeting minutes; verified media coverage with timestamps; ward/sub-county administration records. |
| Measurement Method |
Score = (Frequency Score × 0.5) + (Geographic Reach Score × 0.5). Frequency benchmarks are role-specific: MCAs = monthly; Senators, Women Reps, Governors = quarterly; President = national coverage. Geographic Reach Rule: >75% of applicable sub-units covered = 25 pts maximum; <30% covered = 12 pts maximum. |
| Scoring Guidance |
Excellent (19–25): Meets or exceeds frequency benchmark; Office Open Days documented in >75% of applicable sub-units. Strong (13–18): Meets benchmark in most periods; coverage 50–74% of sub-units. Moderate (7–12): Below benchmark frequency; coverage 30–49% of sub-units. Weak (1–6): Rare Office Open Days; coverage <30% of sub-units. No evidence / violation (0): No documented Office Open Days. |
| Bias / Consistency Controls |
Require attendance register or official minutes per session. Campaign rally records must be excluded. Newcomer proxy: community leadership forums, civic society facilitation, or professional body engagement — scored on same scale with geographic reach adjusted to the candidate's relevant professional domain. |
| Evidence Type |
Official Institutional Records (ward/sub-county administration logs); Verified Public Communications (official office schedules). |
| Verification Level |
Level 1 for ward/sub-county administration records. Level 3 for verified media coverage with timestamps. |
| Indicator ID |
4.4.2 |
| Constitutional Anchor |
Art. 232(1)(d), 35(3) |
| Definition |
The extent to which a candidate's policy proposals or manifesto cite verifiable evidence sources for major claims. Applies the Feasibility Multiplier: Vague (0.0) / Specific (0.8) / Costed (1.0). |
| Primary Evidence Sources |
Published manifesto (IEBC-registered copy); bills or legislative proposals (Hansard); executive orders; official policy documents with date and authorship verified. |
| Measurement Method |
Citation Rate = (Major claims with cited, verifiable sources) / (Total major claims). Apply Feasibility Multiplier per claim: 0.0 = vague/no source; 0.8 = specific mechanism named; 1.0 = full cost-benefit breakdown with verified revenue source. Final score normalised to 0–25. |
| Scoring Guidance |
Excellent (19–25): >80% of major claims cite verifiable sources; majority are Costed (1.0). Strong (13–18): 60–80% of claims cited; mix of Specific and Costed. Moderate (7–12): 40–59% cited; mostly Specific (0.8). Weak (1–6): <40% cited; mostly Vague (0.0). No evidence / violation (0): No citations; or document cannot be authenticated. |
| Bias / Consistency Controls |
Document-based only; evaluate each major claim independently. Binary per claim (cited or not). Newcomer proxy: white papers, published policy commentary, or advocacy framework submissions — scored on same scale. |
| Evidence Type |
Verified Public Communications (IEBC-registered manifesto); Legislative / Government Records (Hansard for bills). |
| Verification Level |
Level 3 for IEBC-registered manifesto. Level 1 for Hansard-recorded legislative proposals. |
| Indicator ID |
5.1.4 |
| Constitutional Anchor |
Art. 73(1), 232(1)(h) |
| Definition |
Degree of confidence demonstrated by credible stakeholders in the candidate's professional leadership. Measured as a Calculated Field using four equally weighted signals. |
| Primary Evidence Sources |
Signal 1: Senior staff retention rate (HR records / official org charts). Signal 2: Documented organisational endorsements (formal letters, joint statements from registered bodies). Signal 3: Verified staff turnover records (official HR or payroll records). Signal 4: Independent leadership evaluations (published 360-degree reviews, board evaluations, or professional body assessments). |
| Measurement Method |
Final Index = (S1 + S2 + S3 + S4) ÷ 4. Each signal scored 0–25 independently before averaging. Exclude partisan endorsements and anonymous testimonials. Only attributable, public sources count. |
| Scoring Guidance |
Excellent (19–25): Strong, documented signals across all four; no credible condemnations on record. Strong (13–18): Three signals strong; one moderate. Moderate (7–12): Mixed signals; some credible endorsements, some neutral. Weak (1–6): Few endorsements; documented credible criticism. No evidence / violation (0): Widespread credible condemnation or no verifiable signals. |
| Bias / Consistency Controls |
Exclude political party endorsements and endorsements from direct family members. Require each endorsement to be attributable to a named, registered organisation. Document each of the four signal scores separately in the audit trail. |
| Evidence Type |
Official Institutional Records (HR / org chart records for S1 and S3); Verified Public Communications (endorsement letters for S2); Independent Data Sources (published leadership evaluations for S4). |
| Verification Level |
Level 1 for HR records. Level 2 for published independent evaluations. Level 3 for formal endorsement letters from registered organisations. |
7. VeriVote Document Hierarchy
This guide operates within a suite of VeriVote documents. Evaluators should understand how they relate:
| Document |
Version |
Evaluator Relevance |
| VeriVote Indicator Protocol v2.0 |
v2.0 (current) |
PRIMARY REFERENCE. Contains all 80 indicator definitions, measurement rules, and scoring bands. Keep open when scoring. |
| How We Calculate Scores |
v2.0 (current) |
Explains how individual indicator scores combine into the final Suitability Index. Includes Trust Coefficients, redistribution logic, and score range interpretation. |
| Master Technical Dictionary & Governance Core |
v2.2 (current) |
Defines all technical terms used across the VeriVote framework. First point of reference for any term or concept requiring clarification. |
| VeriVote Indicator Protocol — Evaluator Guide |
v2.0 (current) |
This document. Operational instructions for evaluators. Does not restate indicator definitions. |
Conflict Resolution
In the event of any conflict between this Evaluator Guide and the VeriVote Indicator Protocol v2.0, the Protocol takes precedence on all matters of indicator definition, measurement method, and scoring band criteria.
This guide takes precedence on matters of evaluator procedure: how to document evidence, how to declare N/A, how to apply the proxy allowance, and how to record the audit trail.