Skip to content
Avanti Docs

Motion in Two Dimension: questions and student responses

Avanti Fellows, built 2 November 2026. Shared for research only. Please do not redistribute.

This folder holds every question in Avanti’s bank for one Physics chapter, Motion in Two Dimension (Class 11, JEE and NEET), that students at our Centres of Excellence answered at least 100 times this year. For each question there is the full text, the options, the answer key, the worked solution and its figures. Every individual answer sits beside it, so the difficulty of each question, and what students got wrong, can be measured directly.

If you are using Claude or another AI assistant, give it this whole folder and start it on this README. Everything it needs to read the files is written here.

What is in it

Section titled “What is in it”
Questions 63: 49 single-answer MCQs, 4 multiple-answer MCQs, 10 numerical-answer
Responses 34,234 question-student-test rows, of which 19,066 were answered and the rest skipped
Students 1,914, in Grade 11 (a handful in Grade 12), at 3 programmes: JNV CoE, EMRS CoE, Punjab CoE
Tests 128 test sittings between 29 July and 2 November 2026: chapter tests (59% of rows) and full-length major tests (41%)
Figures 35 questions have a figure in the question, an option or the solution

Why only CoE students. Avanti’s Centres of Excellence are residential, and their tests are sat on site under supervision. Students in our online and school programmes also answer these questions, but their tests are not proctored, so their answers are left out.

Why at least 100 answers. That is enough to judge a question’s difficulty with some confidence. 34 questions have at least 200. The counts are in items.csv, so you can set a stricter cut.

Files

Section titled “Files”
File One row per What it is
questions.md question Every question in readable form: text, options with letters, key, the answer split, the solution and figures. Start here.
questions.jsonl question The same content as JSON lines: item_id, question_type, text, options (list), key, solution, images (file names)
items.csv question Statistics per question (below)
responses.csv student × test × question Every answer, under a study ID (below)
images/ figure PNG and JPG files named q<item_id>_<n>, referenced from questions.md and questions.jsonl

Text is plain markdown with mathematics in LaTeX between \( and \). A figure appears where it sits in the question as ![figure](images/q2148_1.jpg). Some options are figures themselves, for example the four graphs in item 2148.

items.csv

Section titled “items.csv”
Column Meaning
item_id The question’s ID in Avanti’s question bank. Joins to responses.csv and questions.jsonl
question_type single-choice, multi-choice, numerical-integer or numerical-float
key The answer students were marked against: a letter, letters such as B,C, or a number
cms_answer The answer in the question bank today. For numericals it may be an accepted range, e.g. 0.45,0.55
key_matches_cms_today Whether key agrees with cms_answer (for a range, whether key falls inside it). True for all 63
n_shown Rows for this question: every time a student in a sitting reached it
n_answered How many of those answered rather than skipped
skip_rate 1 − n_answered / n_shown
p_correct Share of answers that were right, among those answered. Skips are not counted as wrong
discrimination_r Correlation between getting this question right and the student’s percentage on the whole test (point-biserial, answered rows only). Above about 0.3 the question separates stronger from weaker students well; near zero or negative, it does not
median_time_s Median seconds spent on the question, among those who answered
n_students, n_tests Distinct students and test sittings
choice_counts How the answers split, most common first: option letters for MCQs, values for numericals
has_figure 1 if the question, an option or the solution has a figure
cms_last_edited When the question was last edited in the question bank

responses.csv

Section titled “responses.csv”
Column Meaning
study_id A random ID for the student, made for this dataset (S and six digits). It carries no information, and the same student keeps the same ID across tests. Avanti holds the only key back to the student
test_id, test_name, test_date The test sitting. One test can carry several questions from this set
test_type, test_format chapter_test covers this chapter or a few; major_test is a full-length mock of JEE or NEET
test_grade, student_grade The grade the test was set for, and the student’s grade
programme JNV CoE, EMRS CoE or Punjab CoE
item_id, question_type, position_in_test The question and where it appeared in the paper
response What the student answered: a letter (or letters) for MCQs, the number entered for numericals. Blank if skipped
key The answer they were marked against
answered 1 answered, 0 skipped
correct 1 right, 0 wrong, blank if skipped
marks, marks_if_right, marks_if_wrong Marks earned, and the scheme: usually +4 for right and −1 for wrong, as in JEE Main
time_spent_s Seconds the student spent on the question, from the test engine
student_test_percentage The student’s score on the whole test, as a percentage. A measure of ability on that day

Things to know before analysing

Section titled “Things to know before analysing”
  • Skipping is a choice. With −1 for a wrong answer, skipping a question you are unsure of is often the right move. So p_correct is accuracy among those confident enough to answer, and skip_rate is a difficulty signal in its own right. For an item response model, treat a skip as missing, not as wrong.
  • Options never shuffle. Option A is the same option for every student who saw the question, so the answer split can be read as the share of students holding each wrong idea. Question order within a test may be shuffled.
  • The key comes from the test as it was given. We checked it against the question bank today. Numerical keys stored as 5 in one sitting and 5.0 in another have been made the same.
  • Multiple-answer questions are marked right only when the student picks exactly the right set. Partial credit, where the scheme gives it, shows in marks.
  • The same student can meet a question twice, in a chapter test and later in a major test. Use test_date to separate a first sight from a repeat.
  • student_test_percentage includes the question itself. For a strict discrimination index, remove the question’s own marks from the total first.
  • Tests differ in difficulty and length, so a percentage is comparable within a test, not across tests.
  • Small groups. EMRS CoE and Punjab CoE are a few hundred students each. Results by programme are fine. Please do not try to narrow down to a centre or a person.

Starting points

Section titled “Starting points”

Things this data can answer:

  1. How hard is each question? Rank them by p_correct and skip_rate together, or fit an item response model from responses.csv.
  2. Which wrong answers are popular, and why? In items.csv, a wrong option chosen by more students than the key usually points to a shared misconception. Read the question in questions.md and name the wrong idea. Item 2148 (slope of a trajectory against time) and item 3680 are good first examples.
  3. Which questions measure the chapter well? Those with discrimination_r above 0.3. Look again at those near zero, such as 3697 and 3699.
  4. Can a model predict difficulty from the question alone? Predict p_correct from the text, options and figure, and compare with what students actually did.

A first prompt for Claude:

Read README.md, then questions.md and items.csv. For the ten questions where the most popular wrong option has the largest share, state the likely misconception behind that option in one sentence each, and say how confident you are.

Privacy and use

Section titled “Privacy and use”
  • No personal data. No names, contact details, dates of birth, schools or centres, and no Avanti student IDs. The study IDs are random.
  • Do not try to re-identify students, and do not combine this data with other data to do so.
  • Shared for this research collaboration only. Please do not publish the student-level rows. Aggregates and question-level findings are fine to publish, with a mention of Avanti Fellows.
  • Questions: contact Avanti Fellows (Akshay Saxena).

How it was built

Section titled “How it was built”

build.py beside this README, in Avanti’s data-assistant repository at analysis/ai-coach/datasets/motion_2d/.

  • Responses: Avanti’s warehouse table fact_student_test_results_question_level, academic year 2026-27, question bank nex-gen-cms, chapter “Motion in Two Dimension”, programmes whose name contains “CoE”.
  • Test percentages: fact_student_test_results_overall, section overall, for the same sittings.
  • Question text, options, solutions and figures: Avanti’s question bank, English versions, read on 2 November 2026. Figures are stored there as embedded images and were written out as files.
  • Study IDs: random. The key from study ID to student is kept privately at Avanti and is not part of this folder.