Student Questionnaire Participation

Does the questionnaire reach as many students as the achievement tests — and how do participation rates track each other across schools?

Data Extract Methodology

This page joins two separate data sources that EQAO publishes in different ZIP archives. Understanding the extraction steps is essential for interpreting the participation rates correctly.

Step 1 — Source files

Two sets of ZIP archives are downloaded from the EQAO Open Data portal:

The two ZIPs are produced independently. There is no single file that contains both achievement participation counts and questionnaire response counts.

Step 2 — Within-ZIP joining

Each questionnaire ZIP contains multiple CSVs (SQ_1.csv through SQ_7.csv). Each file covers a different block of questions but repeats the same identity columns (OrgType, BoardMident, SchoolMident, Language, Suppressed) and carries its own copy of:

prepare_questionnaire_data.py joins the 2–7 CSVs on shared identity columns using df.join(extra, on=join_keys, how="left"), keeping only columns not already present. The first occurrence of cntSQRespondents (from CSV 1) is kept.

Step 3 — Identifying the correct denominator for SQ participation rate

cntSQRespondents counts all students who completed any portion of the questionnaire. The questionnaire is administered as part of the Math assessment window — students complete it alongside the Math test.

Critical difference between G3 and G6:

In some G3 schools, the Reading and Writing cohorts (cntStudents_Read) are substantially smaller than the Math cohort (cntStudents_Math). This occurs in schools with French immersion programs or other instructional arrangements where only a subset of Grade 3 students take the English Reading/Writing assessments, while a larger population takes Math. In these schools, cntSQRespondents tracks the broader Math cohort — not the smaller Reading cohort.

Using cntStudents_Read as the denominator for G3 SQ rates produces inflated values (up to 480%) at affected schools. Using cntStudents_Math — which matches the population that actually completes the questionnaire — produces correct values ≤ 100% at nearly all schools.

This page therefore uses cntStudents_Math as the denominator for SQ participation rates, and compares against pctFullyParticipating_Math (Math achievement participation) for an apples-to-apples comparison.

Step 4 — Cross-file join

This page joins the achievement parquets (schools_g{3,6}.parquet) with the questionnaire parquets (schools_g{3,6}_sq.parquet) on (SchoolMident, school_year) using a LEFT JOIN from achievement → SQ. Schools that appear in achievement data but not SQ data (non-participants) are retained as null SQ values.

Step 5 — Province-level data

Province-level SQ counts come from province_g{3,6}_sq_all.json. Province-level registered student counts (cntStudents_Math) come from province_g{3,6}_all.json. These are joined in JavaScript on (school_year, Language) to compute province-level SQ participation rates.

Province-Level Overview

Denominator: SQ participation rate = cntSQRespondents ÷ cntStudents_Math. Math achievement participation rate = pctFullyParticipating_Math. Using the Math cohort as the common denominator is appropriate because the questionnaire is administered as part of the Math assessment window and tracks that broader population.

Year-to-Year Variation

Trend Over Time

School-Level Distribution

Each panel shows the distribution of school-level participation rates for a given year. Both the Math achievement participation rate and the SQ participation rate use the same denominator (GREATEST(cntStudents_Read, Write, Math) — the broadest registered cohort), so the two distributions are directly comparable.

School-Level Correlation

How tightly are Math achievement participation rates and SQ participation rates correlated across schools? A high Pearson r means both measures respond to the same school-level factors (e.g. student absence policies, engagement culture). A low r means the questionnaire varies independently of test participation.

Schools with the Widest Participation Gap (2024–25)

Schools where questionnaire participation falls furthest below Math achievement participation in the most recent year. A large negative gap means many students who sat the achievement test did not complete the questionnaire; a large positive gap (rare after the denominator correction) indicates a data anomaly worth investigating.