Final Evaluation Report · March 2026

Anti-Racism (AR) & Mental Wellbeing (MW)
Curriculum Development Project

Post Training & Post Module Evaluation — COSTI & JIAS
Prepared by: JIAS Evaluation Team N=33 Instructors N≈67 Students Surveyed · 3 Separate Cohorts Two Modules: Anti-Racism (AR) & Mental Wellbeing (MW) Each Cohort Taught One Module Mixed-Methods Design
📊 Executive & Funder View
Executive Dashboard
Anti-Racism (AR) & Mental Wellbeing (MW) · COSTI & JIAS · March 2026
33
Instructors Trained
88%
≥10 Years Experience
+44pp
COSTI AR Understanding Gain
+40pp
JIAS MW Understanding Gain
91%
COSTI AR Students Found Lesson Useful

Key Insights

  • AR works at scale. Consistent gains in knowledge, comfort, and willingness to seek help across both organizations and learner profiles.
  • ⚠️ MW help-seeking gap requires action. Gains in understanding were strong (+40pp), but certainty about where/how to seek help did not improve — and declined on one indicator. This is the primary implementation risk before national scaling.
  • 📦 Curriculum is not yet classroom-ready for scale. Media glitches, packaging issues, and missing CLB differentiation are solvable problems. A targeted 14-week sprint would address all critical gaps.

Strategic Recommendations

  1. Fund the 14-week rapid curriculum revision sprint. A detailed phase plan is already in place (packaging, CLB tiering, representation, media fixes). Cost of inaction is loss of the evidence base.
  2. Require data collection infrastructure for COSTI Mental Well-Being delivery before the next cohort. COSTI delivered MW; student data was not collected. This gap is correctable with minimal investment.
  3. Approve national scale contingent on post-revision re-evaluation. The AR foundation is strong. MW needs one targeted fix cycle before broader deployment is defensible.

So What for Funders

  • 📐 System gap addressed. Anti-racism and mental well-being curriculum for LINC instructors has been largely absent from the IRCC-funded settlement sector. This project fills a documented need.
  • 📈 Reach estimate. 33 trained instructors × projected ~20 students/year = ~660 newcomers reached per cohort cycle from this single training investment.
  • 💡 Cross-organizational model validated. COSTI and JIAS both delivered both modules. The inter-agency approach is replicable and scalable with appropriate support.

Funding Implications at a Glance

Evidence-Based AR CurriculumFills a sector-wide gap. Pre/post data across two sites confirm meaningful, replicable gains. Ready for investment in revision and scale.
One Targeted Fix NeededMW help-seeking pathway is the only structural gap before national scaling. Addressable within existing revision sprint scope — not a new project.
Cross-Org Model WorksTwo organizations, two modules, one shared evaluation. The model reduces duplication and strengthens sector-wide evidence. Worth replicating.
📊

Impact at a Glance

N=33 instructors · N≈67 students across 3 cohorts · Anti-Racism (AR) and Mental Wellbeing (MW) modules · COSTI & JIAS
33
Instructors Trained
Post-Rollout Survey
88%
Instructors with ≥10 Years Teaching Experience
+44pp
COSTI AR: Gain in Self-Rated Understanding
+40pp
JIAS MW: Gain in Understanding
91%
COSTI AR Students Found Lesson Useful
How we get from data to impact:
📋 Pre/Post
Survey Data
🔬 Focus Group
Insights
📐 Delta
Analysis
💡 Findings &
Patterns
✅ Evidence-Based
Decisions
🏢 Sector & Staff View
🗺️

How to Use This Report

Who this is for and how to apply the findings in your context

📋 What this documents

  • Outcomes of a curriculum development and instructor training project at the intersection of language education, anti-racism, and mental well-being
  • Two modules (Anti-Racism and Mental Well-Being) delivered by two organizations (COSTI and JIAS) to 33 LINC instructors
  • Mixed-methods evaluation: pre/post student surveys across three cohorts, and post-delivery instructor focus groups

🎯 Who should read this

  • Organizations that deliver LINC or other language instruction programming
  • Settlement professionals designing instructor training in equity, anti-racism, or wellness content
  • Evaluators building frameworks for settlement program evaluation
  • Researchers interested in CLB-differentiated curriculum design

🔄 How findings transfer

  • The pre/post survey approach is replicable in any LINC program module with adaptable instruments
  • The CLB differentiation findings apply broadly: lower-CLB learners consistently need adapted materials regardless of topic
  • The help-seeking gap (MW module) is likely not unique to this curriculum — it signals a systemic challenge in integrating mental health content into language learning

⚡ Quick wins for your org

  • Use the delta table (Section 5) as a benchmark for your own pre/post outcomes
  • Adapt the Micro-Protocol script (validate → stabilize → offer choices → refer) for your own disclosure training
  • The two-book packaging model (student booklet / teacher guide) is a transferable standard for any complex LINC module
📋

Executive Summary

What this evaluation found — and what it means

🎯 What We Evaluated

  • Instructor post-training: N=33 (COSTI ≈21, JIAS ≈11) following full rollout of both Anti-Racism (AR) and Mental Wellbeing (MW) modules
  • Student pre/post surveys across 3 separate cohorts — each class taught one module: COSTI Anti-Racism (AR) ✅ · JIAS Anti-Racism (AR) ✅ · JIAS Mental Wellbeing (MW) ✅
  • COSTI Mental Wellbeing (MW) — important note: COSTI delivered the MW module but student surveys were not collected. Mental Wellbeing feedback from COSTI comes exclusively from the post-delivery teacher focus group.
  • Focus groups: Post-delivery teacher discussions at both COSTI and JIAS, covering both Anti-Racism (AR) and Mental Wellbeing (MW) delivery experience

✅ What the Evidence Shows

  • Logic Model alignment confirmed: Immediate gains in knowledge and comfort; short-term outcomes in motion
  • AR modules work at scale: Consistent gains across both COSTI and JIAS
  • MW understanding strong; help-seeking pathways remain the critical gap
  • Packaging and CLB differentiation are prerequisites for broader scaling

🔑 Central Finding

AR modules produced consistent, meaningful gains across both organizations in knowledge, comfort, and willingness to seek help — with especially large movement at COSTI (lower-CLB mix) on English discussability and self-rated understanding. MW produced strong gains in understanding and healthy talk norms but did not strengthen certainty about where to seek help nor whether to ask teachers. This signals the need to sharpen referral pathways and role clarity before broader national scaling.

👥 Project Team View
🔍

What Worked · What Didn't · What to Scale

Internal reflection for learning, iteration, and future planning

✅ What Worked

  • Pre/post survey design captured meaningful change — gains were real and visible
  • Focus groups surfaced implementation barriers not visible in quantitative data alone
  • Instructor cohort was highly experienced (88% ≥10 years) — training landed in fertile ground
  • Both organizations successfully delivered both modules — cross-organizational model validated
  • COSTI AR: dramatic gains for lower-CLB learners on English ease and understanding — strongest signal in the dataset
  • Logic model alignment confirmed at the immediate and short-term outcome level

⚠️ What Didn't

  • COSTI MW student data not collected — major gap in the evidence base; focus group cannot substitute for pre/post data
  • Assessment tasks too similar to practice tasks (isomorphic) — students could pass by repeating, not transferring
  • Media glitches and layout errors affected delivery quality and instructor experience
  • Training schedule (both AR and MW in the same week) caused cognitive overload
  • "Where to get help" and "ask teacher" outcomes did not improve in MW — the module did not close the referral loop
  • Packaging not classroom-ready — teacher notes and student pages interwoven

→ What to Scale

  • Hands-on module labs model: 2–3 hour add-on with rehearsal time before classroom delivery
  • Micro-protocol scripts: validate → stabilize → offer choices → refer (90-second framework)
  • CLB-tiered curriculum variants: Literacy/1–2 and CLB 4–5 as distinct materials
  • Two-book system: clean student booklet + separate teacher guide (print-ready and editable)
  • Non-isomorphic assessments: performance tasks and role-plays, not practice clones
  • Data collection infrastructure before each delivery — not retrofitted after

⚡ Two Tensions to Name

1. Scale vs. Specificity: The JIAS-specific content (Jewish representation, antisemitism examples) reflects genuine equity work — but it also means the current curriculum is not yet neutral enough for other-sector deployment without adaptation. Scaling requires a core + modular structure, not a single version.

2. Evidence Quality vs. Delivery Speed: COSTI MW was delivered without student surveys in place. The gap is understandable operationally, but it means the full evidence base for MW is incomplete. Before the next cycle, data collection infrastructure must be built before delivery begins — not as an afterthought.

🏢 Sector & Staff View
💡

6 Transferable Principles

What this evaluation teaches us about good practice in settlement instructor training

Instructor training must include practice, not just content

Theoretical training alone is insufficient. Instructors needed time to test videos, plan adaptations, and rehearse facilitation before entering the classroom. Apply this to any topic-specific LINC training.

CLB differentiation is non-optional in LINC contexts

Lower-CLB groups need larger fonts, slower audio, and sentence frames. Higher-CLB groups want richer tasks. One version of any curriculum creates inequitable learning conditions from the outset.

Mixed-methods triangulation surfaces what surveys miss

The focus groups revealed packaging problems, delivery barriers, and instructional design gaps that pre/post scores alone would never have captured. Both methods are required for a credible evaluation.

Help-seeking outcomes require explicit curriculum design

Attitude change does not automatically translate into help-seeking behavior. If you want students to know where to go, the curriculum must specifically map those pathways — mini-lessons, decision maps, and referral scripts.

Evaluation design must be embedded, not retrofitted

COSTI delivered the MW module without student surveys in place. The evidence gap that created is not recoverable. Evaluation instruments must be designed and tested before the first delivery, not after.

Site-specific adaptation strengthens generalizability

JIAS's request for Jewish representation in curriculum materials was not a niche ask — it was evidence of a larger principle: representation signals belonging, and its absence signals exclusion. The principle applies to every community.

👥

Participant Profiles

Who participated — instructors and students across both sites
🏫 Instructor Profile (Post-Training Rollout, N=33)
88%
≥ 10 Years Teaching
66%
COSTI Affiliated
34%
JIAS Affiliated
61%
Prior AR/EDI/MH Training
79%
Teach CLB 3–4

Student Sample Sizes by Module

Pre and post survey N for each evaluated module

CLB Level Distribution by Cohort

Post-survey sample breakdown

📈

Student Findings: Pre → Post Shifts

Measured changes across Anti-Racism (AR) and Mental Wellbeing (MW) modules · 3 separate student cohorts · Pre N=66 / Post N=67

COSTI — Anti-Racism (AR) Module

Pre (N=21) → Post (N=23) · CLB 3–5 · Separate cohort

JIAS — Anti-Racism (AR) Module

Pre (N=23) → Post (N=21) · CLB 4–5 · 90% Women · Separate cohort

JIAS — Mental Wellbeing (MW) Module

Pre (N=22) → Post (N=23) · CLB 4–5 · 74% Women · Separate cohort

Gains Comparison: Understanding

Percentage-point change (Pre → Post)

Outcome Indicator COSTI AR JIAS AR JIAS MW
Self-Rated Understanding (4–5/5)▲ +44.3▲ +19.7▲ +39.9
Comfort Discussing Topic in Class▲ +27.5▲ +9.6— n/a
"Racism Can Happen in Canada" (True)▲ +15.9▲ +32.8— n/a
"Tell Someone if Racism Occurs" (True)▲ +31.4▲ +16.0— n/a
Discussing in English (Easy/Okay)▲ +54.6▲ +21.3— n/a
MW: "Should Not Talk Feelings" (False)— n/a— n/a▲ +26.6
MW: Coping Behaviors Help (True)— n/a— n/a▲ +11.9
MW: Can Ask Teacher for Help (True) ⚠️— n/a— n/a▼ −7.1
⚠️ The MW help-seeking gap is the primary implementation priority before broader scaling. Students became more knowledgeable about mental well-being but less certain about whether they could ask their teacher for help.
💬

Qualitative Insights — Focus Groups

Themes from post-delivery teacher discussions at COSTI and JIAS · Anti-Racism (AR) and Mental Wellbeing (MW)
"Instructors consistently requested hands-on module walkthroughs — time to test videos, plan adaptations, and rehearse facilitation before entering the classroom. Theoretical training alone is insufficient."
📍 Shared — Both Sites
"CLB-tiered differentiation is not optional — it's essential. Lower CLB groups need larger fonts, slower audio, and sentence frames. Higher CLB groups want richer tasks and more grammar depth."
📍 Shared — Both Sites
"When a student disclosed emotional distress during the MW assessment, instructors needed a micro-protocol — 90 seconds: validate → stabilize → offer choices → refer. Without this, teachers freeze. Clear scripts save both teacher and student."
🔴 COSTI-Specific
"JIAS instructors strongly requested Jewish visibility in AR content — explicit antisemitism examples, synagogue and rabbi references in MW scenarios. Representation in curriculum signals belonging. Its absence signals exclusion."
🔵 JIAS-Specific
"Assessment tasks were too similar to practice tasks — students could pass by repeating what they did rather than transferring knowledge. Non-isomorphic assessments are needed to generate valid evidence of learning."
📍 Shared — Both Sites
"The interwoven teacher notes and student pages created real navigation problems in the classroom. A clean student booklet and a separate teacher guide — with print-ready and editable files — are the minimum usability requirement for scale."
🔵 JIAS-Specific

✅ What Worked

  • High instructor confidence post-training across both sites
  • Strong perceived fit of AR/MW content with LINC programming
  • Meaningful student gains in knowledge, comfort, and willingness to seek help (AR)
  • MW improved healthy norms around talking about feelings
  • COSTI AR: dramatic gains for lower-CLB learners on English ease and understanding

⚠️ Gaps Requiring Action

  • MW help-seeking clarity: "know where to find help" flat; "ask teacher" declined
  • COSTI MW student surveys not collected — evidence gap for this cohort
  • Media glitches and layout errors affecting delivery quality
  • Assessment isomorphism reducing evidence validity
  • Antisemitism and Jewish representation absent from current materials
  • Packaging not classroom-ready for seamless use at scale
🎯

Recommendations

Nine priority actions across training, curriculum, packaging, and data
Executive priority: Items 1, 2, and 9 are the highest-leverage actions before next cohort delivery. Items 3–8 form the revision sprint scope.
1
Training Delivery
Hands-On Module Labs
Include walkthroughs of AR/MW units with time to test videos, plan adaptations, and rehearse facilitation of bias incidents and disclosures.
2
Training Delivery
Micro-Protocols + Scripts
Provide ready-to-use validate→stabilize→offer choices→refer scripts, "what-not-to-say" guidance, and site-specific referral trees (names, languages, hours).
3
Training Delivery
Pacing
Avoid scheduling both AR and MW trainings in the same week. Include time for practice and reflection between sessions.
4
Curriculum
CLB-Tiered Variants
Literacy/CLB 1–2: larger fonts (18–20pt), clearer emotion images, slower audio, sentence frames. CLB 4–5: richer grammar, additional scenarios, open-ended tasks.
5
Curriculum
Assessment Redesign
Ensure non-isomorphic assessments. Add performance tasks such as help-seeking role-plays and interpersonal vs. systemic racism identification.
6
Curriculum
Representation Completeness
Explicitly include antisemitism examples in AR and Jewish community references (synagogue, rabbi) in MW alongside broader religious and cultural representations.
7
Curriculum
Reduce Repetition
Avoid reusing the same video across skill modules. Fix media glitches, layout/contrast issues, and audio speed problems across both modules.
8
Packaging
Two-Book System
Produce a clean student booklet and a separate teacher guide. Provide both print-ready and editable file formats for seamless classroom use and site-level adaptation.
9
Student Help-Seeking
Normalize Help Pathways
Add a short, leveled "Where to get help in Canada" mini-lesson and simple decision maps — when to talk to teacher vs. settlement worker vs. community provider — directly addressing the MW post-module uncertainty gap.
📊 For Funders & Decision-Makers
💰

Funding Implications

What this evaluation means for investment decisions and scale

Case for Continuation

The AR module has produced consistent, replicable gains across two organizations and learner profiles. The evidence base is sufficient to justify continuation and expansion. The question is not whether to continue but how to fix what the evaluation identified before scaling.

  • Three cohorts of student data collected and analyzed
  • Two organizations validated as delivery partners
  • Logic model alignment confirmed at the immediate and short-term outcome level

Targeted Investment Needed

The 14-week revision sprint is a defined, bounded scope — not a new project. Three phases address all critical gaps identified:

  • Weeks 1–6: Packaging, media fixes, assessment redesign
  • Weeks 2–4 (parallel): Training add-on (module labs + micro-protocols)
  • Weeks 6–14: CLB tiering + representation completeness
  • Post-revision: Re-evaluation with COSTI MW data collected

Risk of Not Investing

The curriculum exists, instructors are trained, and the evidence is in hand. Without the revision sprint, the project stalls at the proof-of-concept stage. The opportunity cost: ~660 newcomers per cohort cycle who would benefit from a scale-ready AR/MW curriculum that currently doesn't exist in the sector.

Policy Alignment

This project directly addresses IRCC priorities around settlement sector capacity, anti-racism in newcomer-serving organizations, and mental well-being supports for newcomers. The cross-organizational model (COSTI + JIAS) demonstrates collaborative sector delivery and reduces duplication. The evaluation framework itself is a transferable asset for the sector.

🗓

Proposed Next Steps & Timeline

Priority sequencing before broader national scaling
Wks
1–6
Rapid Fixes

Packaging, Media & Assessment

Fix media glitches, layout/contrast issues. Separate teacher guide from student booklet. Redesign assessments to be non-isomorphic. Provide editable and print-ready files.

Wks
2–4
Training Add-On (runs in parallel)

Module Labs + Micro-Protocols

2–3 hour add-on to existing training: module walkthroughs, disclosure scripts, site-specific referral maps (names, languages, hours) for both COSTI and JIAS.

Wks
6–14
Curriculum Sprint

CLB Tiering + Representation

Produce Literacy/1–2 and CLB 4–5 variants. Integrate antisemitism content into AR. Add Jewish community references to MW. Update all media assets accordingly.

Post
Rev.
Re-Evaluation

Student Surveys Again — Then Compare

Run MW at COSTI, and both AR and MW at COSTI and JIAS after all improvements. Compare before/after changes. Produce a one-page key results summary for stakeholders and funders.

🔬

Methodology Note

Anti-Racism (AR) and Mental Wellbeing (MW) · Design, instruments, sample sizes, and limitations

🧪 Design & Instruments

  • Mixed methods: surveys (instructor + students) + focus groups
  • Brief knowledge checks embedded in student surveys
  • Pre/post design for 3 separate student cohorts — each cohort taught one module (Anti-Racism (AR) or Mental Wellbeing (MW)); total Pre N=66 / Post N=67
  • Post-only for instructor rollout cohort (N=33)
  • Delta calculations: Post% minus Pre% for each indicator
  • Focus groups: post-delivery teacher discussions at each site, both modules
Why mixed methods? Survey data provides measurable change at scale; focus groups surface the implementation barriers, instructional design gaps, and site-specific context that numbers alone cannot capture. Both are required for a credible evaluation of a curriculum project.

⚠️ Limitations

Self-report bias in all survey instruments — gains reflect perceived change, not independently verified learning
Small class-level samples (N=21–23 per cohort) limit statistical power; findings are indicative, not inferential
COSTI Mental Wellbeing (MW) data gap: COSTI delivered the Mental Wellbeing module but student pre/post surveys were not collected — Mental Wellbeing student findings reflect JIAS only; COSTI Mental Wellbeing instructor perspective captured via focus group only
Varying CLB profiles across sites limit direct comparison of absolute scores (though delta comparisons are more robust)
No long-term follow-up data — sustained behavior change and transfer to classroom practice are not yet measured
Focus group data is qualitative and thematic — not independently coded for inter-rater reliability in this evaluation cycle