GCSE Computer Science Tutor 2026: One-to-One Online Sessions

Live coding in a shared editor, aimed at the question types the AQA and OCR examiner reports say lost candidates marks.

Book a Free Consultation

For a student entered for AQA 8525 or OCR J277, one-to-one online sessions with a Leading Tuition tutor can run in a shared code editor, in the language the school entered them for. 91% of our students achieve their desired grades. Below are the question types that cost marks in the AQA and OCR examiner reports for June 2025, and one national figure we derived from JCQ's June 2026 England results, with the working shown.

Preparing for AQA 8525 or OCR J277 Computer Science?

Our expert GCSE Computer Science tutors help students across the UK with both papers, writing and checking code with students in the language their school entered them for.

Book a Free Consultation Message us on WhatsApp

Where do candidates lose marks on coding questions? Eight patterns

AQA's report on 8525/1 for June 2025 says the coding questions were answered well on the whole in all three languages. The marks it records as lost on those questions are set out below, question by question. OCR's J277/02 report for Summer 2025 describes a noticeable increase in ability on programming questions, and it lists imprecision in high-level programs, including unchecked boundary cases, among the habits of candidates who did less well. The two reports overlap more than the boards' branding suggests. Each of the eight points below is attributed to the report it comes from.

1. Loops that run one time too few

On 8525/1 Question 5, the AQA report says the range function proved problematic, with many students writing range(4) instead of range(5), and other answers using while counter <= 5 where while counter < 5 was needed. In Question 11.5 it records students writing for i in range (7), which would run one time too few. A tutor and student can fix this by running the loop and counting the passes, rather than by rereading it.

2. Conditions checked in the wrong order

AQA 8525/1 Question 3 was a second straightforward programming question. Over 60% of students gained 7 of the available 8 marks and only 30% gained full marks. The report says errors in logic caused most of the trouble for the final mark: a lot of students checked whether the number of students was greater than 15 before checking 25. (In our reading, a check for more than 15 made first would catch every larger number, so the test for 25 would never be reached.) The report adds that students should check their solutions carefully to catch these errors.

3. Boundaries that throw out valid input

In OCR J277/02 Question 1(b)(i), the report records candidates writing num > 1 AND num < 100, which would exclude the valid inputs 1 and 100. It also records if num >= 1 AND <= 100 as incorrect, because a condition joined by AND must reference the variable twice, once in each condition. Question 6(e) of the same paper is named in the report as a place where candidates did not check the boundary case of the exact number of tickets available.

4. Output is not return

The AQA 8525/1 report says students should understand the difference between outputting a value and returning a value, and that where a question states one of these, marks cannot be awarded if they are confused. It also records a common error of writing the word output instead of a correct print or writeline statement, in all three languages.

5. Symbols typed as they are in maths

The same report says students continue to use mathematical symbols incorrectly, such as the single character for greater than or equal to, which would not work in a programming language. It also records the order of symbols being wrong, with => written instead of >=, and some students confusing greater than with less than, which cost marks.

6. The wrong language, or pseudocode when a programming language was required

AQA 8525/1 for June 2025 records an increase in students answering in a different language from the one they were entered for, or using pseudocode when a high-level language was required, and says not all of the marks for that question are then available. OCR J277/02 works differently: its Section B accepts OCR Exam Reference Language or a high-level language the candidate is familiar with, and its report says the majority of high scorers used a high-level language consistently throughout the paper.

7. A minimum that starts at zero

In AQA 8525/1 Question 10.3, the report says only a small number of students initialised the shortest value to one of the array elements or to a large number. The others, in the report's words, instead chose to initialise it as zero, which logically would not work. It also records loops taking the list or array out of bounds in the same question.

8. A variable used outside the function that made it

In AQA 8525/1 Question 11.5, students were asked to extend a program using a given subroutine. The report says a significant number thought the local variable newhidden would be accessible outside the findletter function, instead of storing the function's return value in the variable hidden. Over 40% of students gained over half marks on that question, and 17% did not attempt it.

AQA's final note on 8525/1 says design marks are available for the coding questions, so some marks can be gained even where the code is syntactically incorrect, and that teachers should encourage students to attempt every question. For a student who leaves a coding question blank, that is the most direct mark to recover.

Coding marks slipping on boundaries and loops?

Our expert GCSE Computer Science tutors help students with the loop, condition and boundary errors the AQA and OCR reports describe, in one-to-one online sessions.

Book a Free Consultation Message us on WhatsApp

A selection of AQA questions where half or fewer of candidates scored in June 2025

The table is a selection of eleven questions where the AQA reports give a share at or below about half; it is not a complete list, and other questions in the reports have shares in the same range. We reproduce the share as AQA words it ("only", "just over", "approximately"), so the final column states exactly what each share counts. The shares are not comparable across rows, because some count full marks, some count both marks and some count a single mark.

PaperQuestionTaskShareWhat the share counts
8525/1Q4.1Choose the correct option on the purpose of validation40%chose the correct option
8525/1Q5Describe what a data type is20%could correctly describe it
8525/1Q6Complete a merge sortunder 40%gained full marks
8525/1Q7Give two characteristics of the structured approach to programmingunder 10%gave 2 valid characteristics
8525/1Q10.2Show how records are formed17%gained marks
8525/2Q9.3Calculate the minimum file size of a sound filejust under 50%gained all three marks
8525/2Q11.2State differences between RAM and secondary storagejust over 40%gained both marks
8525/2Q15Identify penetration testingabout 30%gained both marks
8525/2Q16.1Explain the file transfer protocolabout 25%gained the mark
8525/2Q19.2State an advantage of fibre optic cableabout 20%gained both marks
8525/2Q20.2Define a foreign keyabout 15%gained both marks

Source: the AQA GCSE Computer Science 8525 Report on the examination for Paper 1 (1A/B/C) and Paper 2, June 2025, listed on AQA's assessment resources page. Most of the rows are not code. They ask for a definition, a named characteristic or a difference, and AQA's comments on them are about wording, which the next section takes up.

What does define mean in a Computer Science exam?

The AQA 8525/2 report for June 2025 is direct about this. Some students recalled definitions and used appropriate terminology; others appeared not to have learned the definitions from the specification. The report says that when a question uses the command word define, the student must give the actual definition, not a vague or even good description of it.

Words that change the meaning

The report names slips such as using "amount" when "number" is meant, and "information" when "data" is meant. In Question 9.1 on sample rate, it says too many students wrote about the amount of sound or the number of samples with no reference to a time period. In Question 20.1 on relational databases it says too many referred to multiple databases instead of multiple tables, and in Question 20.2 it says some described a foreign key as a physical key.

One difference and its opposite count as one

In Question 11.2 the report says just over 40% gained both marks, and that students need to give two differences, not one difference with the inverse as the second. Writing that RAM is volatile and then that secondary storage is non-volatile earns one point, not two. The same pattern appears in Question 6.2 on assembly language and machine code. In Question 19.2 the report says the cable is not faster; it is the transmission of data that is faster, and answers such as "reliable" needed to refer to reliability over a long distance.

Being precise about a data type

On 8525/1 Question 5, only 20% could correctly describe what a data type was, and the report says over 65% could identify the data type of a value with a decimal point but too many answered Real, which it says is not a data type in any of the three languages in the specification. Question 7 shows the same slip from the other side: the report says the majority gave the benefits of the structured approach when the question asked for characteristics.

Level-of-response questions

For the Question 14 level-of-response discussion in 8525/2, the report says a lot of answers were limited to level 1 (1 to 3 marks) because students wrote a list of statements or some brief descriptions, and other answers were limited to level 2 (4-6 marks) because they covered only legal and/or ethical issues or only cybersecurity. OCR's J277/02 report lists being precise with definitions and instructions among the habits of the candidates who did well.

What Parents Say

DP

Deborah Prior

★★★★★
Confidence and progress

My son was low in confidence after COVID and the gaps in learning were building up. Using this service helped plug those gaps, rebuild confidence, and improve his ability to learn properly again.

NI

Nicola

★★★★★
Amazing tutors!

We arranged maths and English tuition ahead of GCSEs and both tutors were brilliant. My daughter's grades improved, and by results day she achieved two grades higher than in her mock exams.

S

Saifan Khan

★★★★★
Great Teaching and Focused Learning

I attended an English bootcamp with Leading Tuition and the experience was excellent. The teacher gave focused attention to the small group while keeping the lessons interactive and engaging.

Leading Tuition is rated 4.8 out of 5 on Trustpilot from 59 reviews, read on 3 October 2026, and 91% of our students achieve their desired grades.

Want help with definitions and database questions?

Our expert GCSE Computer Science tutors help students give the exact definitions and SQL answers that the reports show marks depending on, in one-to-one online sessions.

Book a Free Consultation Message us on WhatsApp

SQL and records in the June 2025 examiner reports

OCR's J277/02 report for Summer 2025 says that creating and using functions, and SQL syntax, remain areas for improvement. In Question 6(b)(i) it says most candidates identified the incorrect table name and that IF should have been WHERE, while far fewer replaced the AND between field names with a comma or noticed the missing string delimiters around Saturday. The full-mark answer in the report is:

SELECT ActId, Stage
FROM TblMusic
WHERE Day = "Saturday"

The report adds that field and table names do not contain spaces, so Tbl Music is wrong. On the next part, Question 6(b)(ii), which asked for an example of a record as an entire row of the table, OCR writes of it: "This was the lowest scoring question on the entire examination paper." The report says examiners were told to be generous and award the mark wherever four field values were identified.

AQA reports the same topic from its own papers. In 8525/2 it says the database section remained a difficult section for some students. On Question 20.4, which carries up to 6 marks, approximately 25% gained 0 or did not attempt it, while some students reached 5 or 6, and common mistakes included == instead of a single =, misused brackets and problems in the WHERE clause. Most students gained marks by choosing the correct fields in the SELECT clause and the correct tables in the FROM clause. On 8525/1 Question 10.2, only 17% gained marks on records, and the report notes that this type of question posed issues in previous years too.

For how AQA and OCR set up their papers, the GCSE Computer Science exam board comparison sets the boards side by side; the point here is narrower. Both reports comment on these topics, so a student on either board can practise writing queries and record structures before the written answer is attempted.

What can a shared code editor do for these marks?

An online session can use one editor that tutor and student both type in. That matters for a subject where the mark often turns on one boundary, and the sections above give a practical use for each finding.

The editor does not replace the written paper. For subject-level detail beyond this page, the computer science tutoring overview sets out how our tutors work across GCSE and A level.

Preparing for both Computer Science papers?

Our expert GCSE Computer Science tutors help students practise the coding, definition and SQL questions in the sections above, one-to-one and online.

Book a Free Consultation Message us on WhatsApp

Girls are 23.0% of number sat (entries) and 27.6% of grade 7 and above awards: how do the shares compare?

This is a figure we calculated ourselves, from one official table, so that every step can be checked. Source: JCQ GCSE results, June 2026, Table 2, "GCSE results by main grades, subjects, age and sex, England, 2026" (published 20 August 2026). Scope: England, all awarding bodies, JCQ subject "Computing", all ages. JCQ heads its count column "Number sat" and its notes refer to entries, so we write "number sat (entries)". JCQ's notes say: "Figures reported in this publication are provisional as they are not informed by late entries and review of marking."

The inputs, as JCQ publishes them

Number sat (entries)Grade 9 %Grade 7 and above %
Girls18,91912.0%36.0%
Boys63,5098.4%28.1%
Male and Female82,4289.3%29.9%

The arithmetic

  1. Girls' share of entries. 18,919 / 82,428 = 23.0%.
  2. Grade 7 and above awards. Girls: 18,919 x 36.0% = about 6,815. Boys: 63,509 x 28.1% = about 17,840. Together about 24,655. Counts are rounded to 5.
  3. Girls' share of those awards. 6,815 / 24,655 = 27.6%.
  4. Check against JCQ. 24,655 / 82,428 = 29.9%, which is the Male and Female figure JCQ itself publishes at grade 7 and above, so the two sexes' rows reproduce the published total.
  5. Grade 9 awards. Girls: about 2,275. Boys: about 5,355. Together about 7,630. Girls' share: 2,275 / 7,630 = 29.8%.
  6. The gap at grade 7 and above. 36.0% minus 28.1% = 7.9 percentage points.

In other words, girls were 23.0% of the number sat (entries) and received 27.6% of the grade 7 and above awards and 29.8% of the grade 9 awards. The calculation uses unrounded percentages from the table and rates are shown to one decimal place. The figure describes the national distribution of results in one subject in one year. It does not explain why the shares differ and it says nothing about any individual student, so the useful reading for a family is the plain one: in this subject, the top grades are reached by pupils of both sexes.

What do families ask about GCSE Computer Science tutoring?

Which programming language should my child answer in?

The one the school entered them for. AQA's 8525/1 comes as Paper 1A in C#, 1B in Python and 1C in VB.NET, and its June 2025 report records an increase in students answering in a different language from the one they were entered for, or using pseudocode where a high-level language was required, so not all the marks for that question were available. OCR's J277/02 Section B accepts OCR Exam Reference Language or a high-level language the candidate knows, and its report says most high scorers used a high-level language consistently.

Why do candidates lose marks on code that runs?

The AQA June 2025 report for 8525/1 says the errors in logic and boundaries seen that year are seen each year across all three languages. Examples it gives: range(4) instead of range(5), checking for more than 15 before more than 25 in Question 3, and starting a shortest-value search at zero in Question 10.3. In Question 3, over 60% of candidates reached 7 of the 8 marks but only 30% reached all 8, because the last mark depended on getting the logic right.

What do the two boards say about SQL and records?

OCR's J277/02 report for Summer 2025 names SQL syntax, and creating and using functions, as areas for improvement, and calls its Question 6(b)(ii), which asks for an example of a record, the lowest scoring question on the entire paper. AQA's 8525/2 report says the database section remained a difficult section for some students, with about 25% of candidates gaining 0 on Question 20.4 or not attempting it, and its 8525/1 report records only 17% gaining marks on the records question, Q10.2.

How do girls' and boys' top grades compare in GCSE Computer Science?

In JCQ's June 2026 England results for the subject it labels Computing, 36.0% of girls and 28.1% of boys reached grade 7 or above, and 12.0% of girls and 8.4% of boys reached grade 9. Girls were 23.0% of the 82,428 number sat (entries). These are national percentages across all awarding bodies, and JCQ describes them as provisional. They describe results; they do not say why the two groups differ, or what any one child will achieve.

How should a student answer a define question?

AQA's 8525/2 report for June 2025 says that when a question uses the command word define, the student must give the actual definition, not a vague or even good description of it. The same report found only about 15% of candidates gained both marks for defining a foreign key in Question 20.2, with about a further 30% gaining one mark for a partial definition, and some used database in place of table.

Where can I read the examiner reports for my child's board?

AQA lists its 8525 reports on its GCSE Computer Science assessment resources page, by paper and series, including Paper 2 for June 2025 and Paper 1A (C#) for June 2025. OCR lists its examiners' reports for J277/01 (Computer systems) and J277/02 (Computational thinking, algorithms and programming) on its J277 assessment page. Both reports comment on individual questions, so a parent can set a mock paper's weak spots beside the same question types.

Related pages

Book GCSE Computer Science tuition

Our expert GCSE Computer Science tutors work one-to-one online with students entered for AQA 8525 or OCR J277. Tell us the board and the paper your child finds hardest.

Book a Free Consultation
Message us on WhatsApp