Solvi Türkiye Learning Report 2026: Where do students get stuck at home?
We studied the moments when students ask for help with homework and exam preparation. Using Solvi data from April to November 2026, we looked at which topics they get stuck on, what time of day they study and what they do when they get something wrong, and compared the findings with ÖSYM, Ministry of Education and OECD data.
Half of all help requests are maths
Within maths, geometry comes first, and almost every geometry question includes a figure.
Homework happens in the evening
During school weeks, nearly half of all questions arrive between 3 pm and 11 pm. Sunday is the busiest day of the week.
Most mistakes are never revisited
Only 28% of practice questions answered incorrectly are attempted again. Most students who do retry get it right on the second attempt.
The load is in maths, and within maths in geometry
More than half of the questions students photograph and ask for help with are maths. Among students preparing for the LGS high school entrance exam, the share rises to two thirds.
Help requests by subject
Random sample of Turkish-language questions
- Mathematics
- 54.6%
- Turkish
- 14.1%
- Physics
- 9.9%
- Chemistry
- 5.9%
- Biology
- 5.5%
- History
- 2.7%
- Geography
- 2.5%
- Other
- 4.0%
Maths topics with the most help requests
Share of maths questions
- Geometry
- 28.6%
- Equations, inequalities
- 15.6%
- Divisibility, GCD and LCM
- 12.2%
- Rational numbers, fractions
- 9.9%
- Exponents and roots
- 9.9%
- Data and probability
- 5.9%
- Word problems
- 5.9%
- Logarithms
- 4.2%
| Topic | LGS | YKS |
|---|---|---|
| Geometry | 27.6% | 30.4% |
| Exponents and roots | 18.6% | few |
| Equations and inequalities | 13.1% | 17.3% |
| Divisibility, GCD and LCM | 13.8% | 12.5% |
| Rational numbers | 13.1% | 6.0% |
| Logarithms | none | 7.1% |
"Few" means fewer than 30 examples for that exam. Questions tagged as hard make up 14.9% of YKS (university entrance exam) questions and 7.6% of LGS questions.
Official exam data points to the same place
Average LGS 2026 item difficulty (closer to 1 means easier)
- Mathematics
- 0.28
- Foreign language
- 0.44
- Science
- 0.45
- Turkish
- 0.47
- History of the Republic
- 0.53
- Religious culture
- 0.63
Source: Ministry of National Education (MEB ÖDSGM), 2026 Central Exam Report. In TYT Basic Mathematics, the average net score is 15–17% of the number of questions (ÖSYM 2024–2026). In PISA 2025, mathematics is Türkiye's only domain below the OECD average.
In Turkish, the problem is reading passages, not grammar
A third of the Turkish questions students ask for help with are reading-passage questions. They are short, have no figures and almost none are tagged as hard. Students are not struggling with calculation but with understanding what they read.
Turkish topics with help requests
Share of Turkish questions. In the YKS, reading passages account for 48%.
- Reading passages
- 37.3%
- Meaning in words and sentences
- 16.9%
- Sentence structure
- 8.4%
- Phonology
- 7.8%
- Spelling rules
- 7.2%
- Verbals
- 6.0%
Very few prefer reading
Answers to "How do you learn best?"
- Watching videos
- 55.3%
- Solving questions
- 39.3%
- Reading
- 5.4%
In-app onboarding survey. Percentages are of those who answered.
Physics
- Motion, projectiles
- 25.2%
- Torque and equilibrium
- 11.6%
- Pressure
- 11.6%
- Electrostatics
- 11.6%
97% of motion questions include a graph or figure.
Chemistry
- Atoms, periodic table
- 17.8%
- Chemical bonds
- 15.3%
- Mole calculations
- 13.6%
- Redox
- 12.7%
Concepts (atoms, bonds) and calculation (moles) carry equal weight.
Biology
- DNA, genetic information
- 20.8%
- Cells, membrane transport
- 16.0%
- Inheritance
- 13.6%
- Energy metabolism
- 13.6%
DNA stands out in May, cells in September–October.
Students want reasons, not shortcuts
After seeing a step-by-step solution, students ask a follow-up question on roughly one in six solutions. Most follow-ups are not about the answer itself but about where a step comes from.
Follow-up themes, most to least common
- 1"Why this step?"Where a relationship, formula or value comes from.
- 2Missing prior knowledge and misconceptionsA gap in a concept that should have been learned in earlier grades.
- 3"I didn't get it at all"Asking for the explanation again, more slowly or in a different way.
- 4Remembering and exam strategyHow do I remember this in the exam, and how do I get faster?
Qualitative review of conversations. Requests for "a shorter way" or "a similar question" are almost absent.
School questions prompt more follow-ups than exam questions
Share of solutions with a follow-up question, by question type
- School and homework
- 20.8%
- LGS
- 18.2%
- YKS
- 14.0%
The exam type is estimated automatically from the question photo. Follow-ups are not concentrated on hard questions: the rate for questions tagged as hard (12%) is below that of easy and medium questions (16–19%).
Homework in the evening, studying on Sunday
During school weeks, students start studying after school. Activity peaks between 7 pm and 9 pm, and about 8% of questions arrive between 11 pm and 3 am.
Hour of day and day of week
Two school weeks from September and October 2026, Türkiye time. Darker means busier; hours with too little data are shown faded.
Distribution by day of week
Autumn term, 8 September – 1 November 2026
- Monday
- 13.0%
- Tuesday
- 13.5%
- Wednesday
- 13.9%
- Thursday
- 14.0%
- Friday
- 10.9%
- Saturday
- 16.8%
- Sunday
- 17.9%
The weekend share is 34.7%; an even distribution would give 28.6%.
Studying stops as soon as exams end
June 2026, daily question volume (1–12 June average = 100)
In July and August, the share of app visitors who ask a question also falls from 57–59% to 38–42%, recovering in September.
Those who go back fix it, but most never go back
The app generates practice questions from each student's weak topics. The results show how students deal with their mistakes.
Practice questions answered correctly on the first attempt
Incorrectly answered questions that are attempted again
Retried questions corrected on the second attempt
Retried questions eventually answered correctly
"What do you struggle with most?"
Onboarding survey respondents
- Solving questions
- 63.4%
- Understanding the topic
- 27.8%
- Time
- 5.6%
- Motivation
- 3.1%
Big goals, weak habits
47% of students initially aim to solve more than 30 questions a day. Practice data shows that volume crowds out going back to mistakes. On a four-option question, the chance of guessing right on the second attempt is about 33%. The observed 62% shows that students who go back largely do fix their mistakes.
Rates in this section are calculated per question. Practice and survey data cannot be split by country; the content is mostly YKS and LGS.
Six recommendations for educators
Practise reading figures separately
Geometry is 29% of maths help requests, and 92% of those questions include a figure. Exercises on extracting data from figures may be more targeted than reviewing formulas.
Treat reading passages as a reading skill
37% of Turkish questions are reading passages, and only 5% of students prefer learning by reading. Short, regular reading tasks address both findings at once.
Answer "why this step?" up front
The reasoning behind steps tops the list of follow-up questions. Saying why each step is taken builds the connection students cannot make on their own.
Make revisiting mistakes part of homework
72% of mistakes are never retried, yet most are fixed when they are. Repeating fewer questions may be more productive than solving many new ones.
Plan for unit starts and evenings
Students look for help on the unit taught that week, between 7 and 9 pm, and on Sundays. Extra resources at the start of a unit and a short reminder before Sunday reach them when help is actually needed.
Build a bridge for after exams
Studying drops quickly after exam week. A light but regular review plan for the summer can make the return in September easier.
What these figures measure, and what they don't
Data
- The source is Solvi, an app where students photograph a question and get a step-by-step solution. Period: 3 April – 2 November 2026.
- Subject and topic distributions were calculated from a sample of more than 2,000 Turkish-language questions, weighted towards May and September–October 2026.
- Hour and day patterns were derived from all question submissions from Türkiye. Times are in Türkiye time (UTC+3).
- Practice and survey findings cannot be split by country. Because almost all content is YKS and LGS, they mainly reflect Türkiye.
Reading notes
- The report shows the topics students ask for help with most, not the topics they get wrong most.
- Subject, topic, difficulty and exam tags are estimated by AI from the question photo. They are not ÖSYM or Ministry of Education classifications.
- The data represents students who use Solvi, not all students in Türkiye. Seasonal patterns are based on a single year.
- All data is aggregated and anonymous. No rate based on fewer than 30 examples is published. Student messages are not quoted; only themes are given.