When a new tutoring centre opened near a Texas high school, students' state test scores stayed flat. What moved were their SAT scores, the colleges they chose and, years later, their paycheques (Chen, 2026).
This digest covers five pieces of research published between October 12 and October 26, 2026. Two are Canadian: a Statistics Canada release on who finishes high school on time, and a Statistics Canada study of how the school-age population and the classroom are changing. Three come from the United States: a working paper on what private tutoring centres change for students, a national study of state science of reading laws, and a Hechinger Report look at an audit of research on AI in education. Read together, each one reports an average, and each one finds that the average hides a split. Who a student is, and what the help is aimed at, decides most of what happens.
Do Tutoring Centres Improve Grades? A Study of Five Million Texas Students
Marty Haoyuan Chen, an economist at the University of Maryland, linked business records for every private tutoring establishment in Texas from 1997 to 2025 with state records that follow more than five million public school students from high school into college and work. His paper is dated October 12, 2026 and was released as Annenberg Institute EdWorkingPaper 26-1619. Over that period the number of tutoring centres in Texas grew from 234 to more than 900, and the share of public high schools with at least one centre within two miles rose from 19 to 37 per cent. The study compares students at the same high school in different years, as centres opened and closed nearby (Chen, 2026).
More centres nearby had no measurable effect on state standardized test scores or on high school graduation. It did raise college admission exam scores. Scaled to students who actually attended a centre for a year, the paper estimates a gain of roughly 8 to 13 SAT points, a shift toward more selective and out-of-state colleges, and earnings in the late twenties that were $1,130 to $1,940 a year higher, or 2 to 4 per cent. Those gains appeared only among students who were not economically disadvantaged. For students from lower-income families, the effects were close to zero on nearly every outcome. A survey of 1,150 Texas parents in July 2026 helps explain why: 25 per cent of children were receiving paid private tutoring, and 13 per cent of children from better-off families attended a tutoring centre, compared with 3 per cent of children from lower-income families. Among parents of tutored high school students, the top reasons were test preparation (58 per cent), competitive college admission (54 per cent) and a child struggling academically (47 per cent) (Chen, 2026).
This is a single-state US working paper that has not yet been peer reviewed, and it measures how many centres were nearby, not what any one program taught. The author notes that the gains on admission exams are smaller than the effects of 0.3 to 0.4 standard deviations reported for publicly funded tutoring programs that serve struggling students and are measured on state tests. The lesson that travels to Canada is simple: tutoring changes what it is aimed at. Exam prep moved exam scores. If a child's real gap is fractions, reading fluency or essay structure, the help has to be aimed squarely at that gap. Our guide on how to tell if your child needs a tutor walks through the signs worth acting on.
Who Finishes High School in Canada? The Gender Gap Is Narrowing
On October 21, 2026, Statistics Canada reported that 84 per cent of students in 2023/2024 finished high school on time, meaning within three years of starting Grade 10. Within five years, the extended-time rate reached 91 per cent, up from 87 per cent in 2016/2017. The gap between young women and young men has narrowed sharply. In 2016/2017, 84 per cent of young women graduated on time compared with 76 per cent of young men, a gap of 8 percentage points. By 2023/2024 the rates were 86 and 82 per cent, a gap of 4 points (Statistics Canada, 2026).
Provinces differ. Young women had higher on-time rates in every province, by 3 percentage points or less, except Quebec (9 points) and Ontario (5 points). With extra time, the gaps shrink further: Quebec's went from 12 points in 2021/2022 to 7 points in 2023/2024, and Ontario's from 5 to 3. Prince Edward Island, Nova Scotia and New Brunswick show no gender gap once the extra two years are counted. Taking those extra years is common: 9 per cent of students in Quebec, 8 per cent in Ontario and Saskatchewan and 7 per cent in Manitoba graduated in extended time, compared with 2 to 5 per cent in the Atlantic provinces (Statistics Canada, 2026).
On-time high school graduation in Canada, by gender
On time means within three years of starting Grade 10. Source: Statistics Canada, The Daily, October 21, 2026
For a parent of a teenager, two things stand out. Boys are closing the distance, and an extra year is a common route to a diploma, not a verdict on a student. The time to act is earlier, when a course starts to slip in Grade 10 or 11 and there is still room to recover the credit. A quick conversation with your teen's teacher about which assignments matter most is often the most useful first step.
How Is Canada's Classroom Changing? A New Statistics Canada Study
The same day, Statistics Canada published Demographic Echoes, a study of the school-age population by Graham Ziegler. Canada had about 5.8 million children aged 5 to 17 in 2025. Under the agency's medium-growth projection, that falls to about 5.5 million by 2035, a drop of 4.5 per cent, even as the total population grows by 5.9 per cent. The picture varies by province: the school-age population is projected to fall 11.7 per cent in Quebec, 7.4 per cent in British Columbia and 2.9 per cent in Ontario, and to grow 3.3 per cent in Alberta and 4.6 per cent in New Brunswick (Statistics Canada, 2026).
Fewer children does not mean simpler classrooms. Drawing on the Canadian Health Survey on Children and Youth, the study reports that between 2019 and 2024 diagnosed attention deficit disorders rose from 6.7 to 9.6 per cent of children, autism spectrum disorder from 2.0 to 3.8 per cent, and anxiety disorders from 5.1 to 5.5 per cent. Diagnosed learning disabilities fell from 8.4 to 6.6 per cent, about 1 in 15 school-age children. The share of children with excellent or very good mental health dropped from 83.2 to 78.1 per cent, and the share of parents reporting high stress rose from 26.2 to 31.9 per cent (Statistics Canada, 2026).
Canadian children, 2019 and 2024
| Measure | 2019 | 2024 |
|---|---|---|
| Attention deficit disorders (diagnosed) | 6.7% | 9.6% |
| Autism spectrum disorder (diagnosed) | 2.0% | 3.8% |
| Anxiety disorders (diagnosed) | 5.1% | 5.5% |
| Learning disabilities (diagnosed) | 8.4% | 6.6% |
| Excellent or very good mental health | 83.2% | 78.1% |
| Parents reporting high stress | 26.2% | 31.9% |
Share of children and youth, Canadian Health Survey on Children and Youth, as reported in Statistics Canada, Demographic Echoes, October 21, 2026
The study also counts more adults in schools. The number of educators in public schools rose 20 per cent, from 350,619 in 2002/2003 to 422,151 in 2023/2024, while teacher job vacancies more than doubled, from 2,200 in 2018 to 5,050 in 2024. Statistics Canada notes that national data on class sizes are not available. The figures describe diagnoses, so they cannot say how much of the rise reflects more screening and awareness and how much reflects more need. For families of a child with ADHD or autism, the practical takeaway is to ask how the school's support plan, or IEP, handles the subjects your child finds hardest, and what you can reinforce at home.
Do Science of Reading Laws Raise Reading Scores?
By 2025, 41 US states and the District of Columbia had passed laws aligned with the science of reading: requirements for explicit, systematic teaching of phonics and other foundational skills, early screening, teacher training and approved materials. Florence Bason and Christopher Redding of the University of Florida tested whether those laws changed Grade 4 reading results on the National Assessment of Educational Progress (NAEP), using state data from 2003 to 2024. Their paper was released in October 2026 as EdWorkingPaper 26-1616 (Bason and Redding, 2026).
The gains did not arrive right away. Four years after a state adopted its law, Grade 4 reading scores were about 2.8 NAEP points higher than expected, roughly 0.43 of the state-level standard deviation, and the estimate held at five years. The gains were not shared evenly. For students not eligible for free or reduced-price lunch, the estimated effect was 2.37 points and statistically significant. For students who were eligible, it was 1.00 point and not significant. The authors also found that a coordinated package of policies appeared to matter more than any single requirement on its own (Bason and Redding, 2026).
Canada is on a similar road. The Ontario Human Rights Commission's Right to Read inquiry report, released February 28, 2022, called for direct and systematic instruction in foundational reading skills and for screening every student at least twice a year from Kindergarten to Grade 2. The US findings suggest two things for parents here. Change at the system level takes years to show up in scores, and children who start furthest behind may need more direct, individual practice than a new curriculum delivers on its own. Our explainer on what the Right to Read report means for your child covers the screening steps in Ontario.
Can You Trust the Research on AI in Classrooms?
On October 26, 2026, The Hechinger Report's Jill Barshay reported on an audit by Patrick O'Neill, a community college professor at Ivy Tech in Indiana. O'Neill examined 14 of the 39 published meta-analyses on AI in education, every one of which had reported positive findings. He found serious conceptual and statistical flaws, including statistics that could not be correct, references that did not exist and conclusions that went further than the data. The papers have drawn more than 3,000 citations and gain about 200 more each month. One has since been retracted (The Hechinger Report, 2026).
O'Neill's preliminary reanalysis of 44 well-designed experiments found results that ranged from very helpful to very harmful and averaged out to roughly zero. He was clear that this does not mean AI is harmless or useless: the effect depends on how a tool is used. Susanna Loeb of Stanford University told Barshay her team reviewed almost 900 studies and found only 20 strong enough for its evidence summary, and very few high-quality studies have been done in US K-12 schools. The audit is a preprint and is still awaiting peer review (The Hechinger Report, 2026).
For families, this is a reason for healthy skepticism rather than alarm. When an app promises that it raises learning, ask what exactly it does: does it explain a step, check work, or simply hand over answers? Last week's research roundup covered a university trial where an AI tutor lowered grades, and our post on how students themselves view AI shows many teens already notice the difference.
What This Means for Families and Teachers
Every study in this digest has a headline number, and every one has a split underneath it. Tutoring centres moved SAT scores, but only for students from better-off families who used them for test prep. Reading laws lifted scores, but mostly for students who were already less likely to struggle. Graduation rates are high, but boys and girls still travel at different speeds. AI studies looked positive on average because the weak ones were counted alongside the strong.
So the most useful question for a parent is not whether something works in general, but whether it fits your child. What is the specific gap? What is the goal this term? What would show that it is closing? A short conversation with the classroom teacher answers the first question more often than a test does. If your child needs one-on-one help aimed at that gap, Connect Education can match you with a tutor for the subject and grade. Every tutor on Connect Education is a licensed teacher in good standing. You can book a free 15-minute assessment or browse certified tutors to see who is available.
Certified teachers
Online, home or library