Assessment Request for
Proposal (RFP) Development
Toolkit

Customization Symbols Key

Every exemplar language block uses placeholder symbols where state-specific information must be inserted.

See the Language Legend section for the complete Customization Symbols Key, including all placeholders and example replacements.

Purpose of the Procurement

This section should outline the state's purpose of seeking proposals from qualified vendors to develop, implement, and maintain a comprehensive statewide assessment system. This system will be a critical tool for measuring student achievement, informing instructional practices, and ensuring accountability across the state's educational landscape.

Process Overview

To facilitate a fair and transparent evaluation of all submissions, the procurement process will follow a competitive sealed proposal method.

Key Dates — Example Timeline

Glossary of Terms

Rightsholders, Clients, and Partners

Rightsholders are individuals or groups directly affected by the assessments' outcomes but hold minimal to no direct decision-making power in its design or implementation. Examples include students, parents, or local community-based advocates.

Partners are individuals or groups who have a vested interest in the assessment and hold some decision-making power in the assessment development processes. Partners include legislators, the Governor's office, the state budget office, the state procurement office, the state auditor's office, and local school boards.

Rightsholder Accountability Council (RAC)

The RAC is a community engagement body comprised of individuals and groups representing intersecting marginalized populations from the local community. Input provided by RACs helps inform the intended context of the assessment, provides meaningful and actionable feedback concerning the intended outcomes, and informs the selection of the most relevant and appropriate measures to evaluate them.

Assessment System Background — Exemplar Language

Major Reference Documents

Program Management

The vendor shall establish and maintain effective program management practices throughout the contract. Key areas include:

• Key Contacts — Designated contacts on both the state and vendor sides for all communications.
• Executive Management Meetings — Regular meetings between senior leadership to review program status and resolve escalated issues.
• Documentation Repository — A centralized, state-accessible repository for all project documents, deliverables, and decision logs.
• Change Management — A formal process for managing changes to scope, schedule, and budget, with written approval from the state required before implementation.
• Annual Kickoff Meetings — Formal kickoff at the start of each contract year to align on priorities, schedule, and any program changes.
• Periodic Management Monitoring Meetings — Regular monitoring meetings (typically monthly or quarterly) that include community representation.

Project Management

The vendor shall maintain a project management approach that includes a detailed project schedule updated monthly, a project management team with named roles and responsibilities, and a plan for finalizing the Theory of Action (ToA) if applicable.

Communication Support

All communications between vendors and rightsholders/partners must always be jointly conducted with the client (state). Direct meetings between vendors and rightsholders — including presentations — must always be jointly conducted with the client to maintain transparency and equity of process.

Disaster Planning and Recovery

The vendor shall provide a documented disaster planning and recovery plan that covers system failure scenarios, data recovery timelines, and communication protocols to ensure assessment continuity.

Content Frameworks and Standards

This section requires vendors to demonstrate full alignment with state content standards, including annotations that identify the depth, breadth, and complexity the assessment must measure. Vendors must provide Content Standards and Annotations that map each standard to test blueprint specifications.

Achievement Level Descriptors (ALDs)

Vendors must develop ALDs that describe what students at each performance level know and can do. ALDs must be developed in collaboration with the Rightsholder Accountability Council at the Justice tier, and must be written in accessible language for non-specialist audiences including parents and community members.

Innovation in Assessment Design

Item Development

The foundation of a fair and accurate assessment lies in the expertise and diversity of its item writers. Item writer recruitment must emphasize qualifications that value cultural responsiveness, implement proactive recruitment strategies, and establish inclusive hiring procedures.

Principles of Culturally Responsive Assessment for Item Writing

Standard Setting

Vendors must provide a documented standard-setting plan that includes a qualified external evaluator, a detailed process for incorporating rightsholder feedback into performance level definitions, and a plan for monitoring classification accuracy and consistency.

Reporting

Reporting must consider multiple audiences — students, families, educators, and policymakers — and must be designed with input from the RAC. Reports must be written in asset-based language and must account for measurement error in all score presentations.

Item-Adaptive Testing

For assessments using adaptive testing algorithms, vendors must document the adaptive algorithm, including its psychometric basis, selection criteria, and how it handles irregularities such as connectivity failures, student disengagement, or anomalous response patterns.

Remote and Virtual Testing

Vendors must define "remote testing" and "virtual testing" as used in the contract, and must provide a comprehensive plan for students who cannot access online testing environments — including an alternative-format administration option with the same content coverage as the online assessment.

Scoring

• Non-AI-based automated scoring — Rule-based and pattern-matching systems must be documented with clear evidence of validity and reliability across student subgroups.
• Hand-scoring — Vendors must document scorer training, reliability monitoring, and the process for resolving scoring discrepancies.
• AI Scoring — If AI is used for scoring, vendors must comply with Annex A requirements, including bias auditing across all reporting subgroups, human review of a defined percentage of AI-scored responses, and community review of AI scoring rubrics.

Test Security

Vendors must provide comprehensive test security procedures and protocols covering pre-administration, administration, and post-administration periods. Security plans must address both physical and digital security, including procedures for detecting and responding to security incidents.

Online Dynamic Reporting System

The vendor must provide an online reporting system that allows educators and administrators to access disaggregated score reports, filter by reporting subgroup, and download data files in accessible formats for community partners.

Claims and Validity Evidence

Vendors must develop a validity argument that identifies the intended inferences from assessment scores (claims) and marshals evidence to support each claim. The validity argument must address five major sources of validity evidence as identified in the Standards for Educational and Psychological Testing:

• Content evidence — Alignment between assessment content and the constructs being measured.
• Response process evidence — Evidence that students engage with items in ways aligned with the intended construct, including cognitive lab evidence across targeted populations.
• Internal structure evidence — Psychometric evidence that the assessment behaves as intended, including factor analyses and fit statistics.
• Relations to other variables — Convergent and discriminant validity evidence showing the assessment relates to other measures as expected.
• Consequences evidence — Evidence regarding the impact of assessment use on students and institutions, including disparate impact analyses across all reporting subgroups.

Psychometrics

The vendor must document its psychometric model and calibration procedures, including IRT model selection, field test analyses, operational item analyses, and scaling and equating procedures. All psychometric analyses must be disaggregated by reporting subgroup to support detection of differential item functioning (DIF) and disparate impact.

Validation Argument and Peer Review

The assembled validity evidence must be synthesized into a Validation Report that is reviewed by an independent peer review panel. The peer review panel must include at least one member with documented expertise in measurement equity and one community member representing the RAC's perspective on the validity evidence.

Conflict of Interest

The vendor must disclose all actual or potential conflicts of interest at the time of proposal submission and must update this disclosure throughout the contract term. See MCR-7 for the complete list of disclosable conflicts.

Issue and Risk Management

The vendor must provide an Issue and Risk Management Plan that includes a risk register updated at each periodic management monitoring meeting, a defined escalation path for high-severity risks, and a process for involving the RAC in risk assessments that affect community engagement or data access.

Fiscal Management

The vendor must maintain documented fiscal management practices including budget tracking by deliverable, a process for managing scope change and associated budget adjustments, and annual financial reporting to the state that includes actual vs. budgeted spend on rightsholder compensation and community engagement.

Penalties

The contract must specify financial penalties for breach of MCRs and substantive deliverable failures. Penalty structures should be calibrated to the severity and impact of the breach on students and communities.

Bid Evaluation Process

The evaluation process proceeds in two stages:

• Stage 1: MCR Compliance Check — The evaluation team verifies that the vendor has accepted every MCR using the Vendor Attestation form. Proposals failing this stage are disqualified without substantive review.
• Stage 2: Substantive Scoring — Qualifying proposals are scored using the Universal Rubric (0–3 scale) applied to each section. Proposals must meet a minimum overall score threshold to qualify for contract negotiation.

The state uses a "best value" criterion rather than a "lowest cost" criterion for selecting from among qualifying proposals.

Cost Proposal

The cost proposal template must include separate line items for: core assessment development; rightsholder compensation schedule; accommodations budget (childcare, transportation, technology, translation); accessibility remediation; community engagement activities; and AI compliance activities (if applicable).

Purpose and Scope

This Annex applies if the vendor uses artificial intelligence at any stage of the assessment lifecycle. It establishes the disclosure requirements, use-case-specific guardrails, state decision rights, and verification mechanisms that govern AI use in assessment. This Annex is referenced in the AI section of MCR-5 and in Chapters IV and V.

Key Definitions

AI Use Cases

State Decision Rights

• 3.1 Pre-Deployment Approval — The state must approve any AI model before it is deployed in the assessment lifecycle.
• 3.2 Mid-Contract Model Replacement — The state retains the right to require model replacement if disparate-impact thresholds are breached.
• 3.3 Independent Algorithmic Audit — The state retains the right to commission an independent algorithmic audit at intervals defined by the state, at the state's expense.
• 3.4 Off-the-Shelf Model Veto — The state retains the right to veto the use of any off-the-shelf model that the state cannot adequately audit or that fails the disparate-impact threshold.
• 3.5 Community Sign-Off (Justice Tier) — At the Justice tier, the RAC must approve AI models used for item generation, scoring, and reporting before deployment.