How should you read a chapter-wise mock test result?

A chapter test ends with a score, and it is the first number anyone looks at. It is also the one that tells you least about what to do next. Two chapters can score almost the same for opposite reasons, and need opposite work. The counts printed under the score tell them apart.

Why is the score the least useful number?

Because it adds together two things you fix in different ways. A low score can mean you attempted few questions and got most of them right, or that you attempted almost all of them and got many wrong. The first needs more practice on the chapter. The second needs more care: slower reading, checking, or fewer guesses. The score alone cannot say which.

Which two numbers should you work out?

From the three counts, two rates. Call the number right \( r \), wrong \( w \), skipped \( s \), and the total \( n = r + w + s \).

$$ \text{attempt rate} = \frac{r + w}{n} \qquad \text{accuracy} = \frac{r}{r + w} $$

In words: the attempt rate is the share of questions you answered at all; accuracy is the share of those answers that were right. Skipped questions count against the first and not at all in the second.

Under JEE Main's marking, +4 for a right answer and −1 for a wrong one, the marks from a test are

$$ \text{marks} = 4r - w $$

In words: four for every right answer, minus one for every wrong one. A skipped question adds nothing and takes nothing.

What are the four kinds of chapter?

Put each chapter on a chart, attempt rate across and accuracy up, and draw a line through the middle of each. The lines here are at 50%; where to draw them is a choice, not a rule.

• Strong: you attempt most questions and get most right. Keep it up with mixed tests.
• Rushed: you attempt most questions but get many wrong. The marks go to wrong answers: slow down, and check before you mark.
• Avoided: you get right most of what you attempt, but attempt little. The marks go to skipped questions: you know more of the chapter than you use.
• Weak: you attempt little and get much of that wrong. Go back to the chapter itself before testing it again.

A worked example

Here are five 25-question JEE Physics chapter tests. The numbers are made up to show the method; they are not anyone's results.

Take the two that look alike. Rotational motion scores 26 out of 100, Thermodynamics 31: close enough to file together as “weak”. They are not alike at all. In Rotational motion, 19 of the 25 questions were attempted and 10 of those were wrong: ten marks went to wrong answers. In Thermodynamics, only 9 were attempted, 8 of them right, and 16 were skipped. One chapter needs care, the other needs practice.

A note on guessing: under +4 and −1, an answer is worth giving if it has better than a one-in-five chance of being right, so an accuracy below 20% loses marks on average. The arithmetic is worked through in Should you guess when there is negative marking?

What should you review first?

The wrong answers, with their explanations, and the ones you were sure of before the rest. That order comes from research on learning from mistakes. Janet Metcalfe's review of that research (Annual Review of Psychology, 2017) reports that making errors and then receiving corrective feedback helps learning, and that the benefit is greatest for errors made with high confidence. That last finding, first reported by Butterfield and Metcalfe in 2001, is called the hypercorrection effect: a confident mistake, once corrected, is more likely to stay corrected.

A skipped question gives you nothing to correct; a wrong one shows you what you believed. So read the explanation of every wrong answer before moving on, and only then turn to the questions you skipped, as practice.

What this does not tell you

One test is a small sample: 25 questions can swing on a few. Look for the same kind of result across two or three tests of a chapter before deciding what it is. And this is a way to read a result, not a promise of marks; no study is cited here that measured its effect on scores.