Most Indian commerce internals are still last year’s question paper with the firm name changed. Students know this. Coaching centres know this. First-generation students who work evenings know this. They therefore allocate effort to memorising formats, not to deciding which format applies. The mid-semester test then certifies recall speed and handwriting stamina. It does not certify whether a student can choose an accounting treatment, reject a bad ratio, or attach a condition to a credit decision. This briefing is a working method for writing a 60- or 90-minute internal that measures judgment in B.Com., BBA and first-year MBA sections without requiring a Western testing office or a teaching-assistant army.
The method assumes the constraints that actually exist: 80 to 120 scripts, a university pattern that still wants “short notes,” mixed English fluency, a department that fears student unrest if the paper looks unfamiliar, and a teacher who already carries four courses. It does not assume item-response software. It assumes a blueprint, three item families, a public rubric, and a return hour that treats the script as a case, not as a verdict.
1. Why the Indian mid-sem still rewards the photocopy
Three institutional habits keep the internal stuck. First, pattern lock. Boards and affiliating universities publish a unit-wise mark split and a “short note / long problem” skeleton. Departments copy that skeleton into internals so that students will not complain that the mid-sem “was not like the university paper.” The result is two identical assessments instead of one diagnostic and one summative. Second, item recycling. A problem on FIFO versus weighted average with the same three purchases appears every autumn because it is easy to mark. Students buy the previous five papers and practise the arithmetic. Third, marking as addition. Scripts are scored by counting completed steps. A student who applies the wrong method but finishes the table can outscore a student who names the right method and stops when the data run out.
First-generation students are hit twice. They often have weaker English for the “discuss” items, so they dump definitions. They also have less time to sit five previous papers. A memory test therefore looks fair in the staff room and unfair in the hostel. The teacher who wants judgment then faces a political problem: a paper that looks new will be called “out of syllabus” even when every concept was taught.
Do not wait for the Board of Studies to rewrite the pattern. Use the existing mark split. Change what sits inside each bucket.
2. Write a one-page blueprint before you write a single item
Start with a table, not with a favourite problem. The blueprint is one page. Share it with the co-teacher if the course is shared. Do not share the items. The blueprint is what you will defend if a student says the paper was unfair.
| Block | Marks | Claim being tested | What a memory-only student cannot do |
|---|---|---|---|
| A. Choice under constraint | 8–10 | Select a method or treatment and name the constraint that forces the choice | Cannot dump both methods and hope one matches the key |
| B. Computation with a trap | 12–16 | Compute after discarding one irrelevant figure | Cannot copy the full table from last year’s solution booklet |
| C. Judgment paragraph | 8–10 | Reject one alternative and attach one condition | Cannot score with a memorised “advantages and disadvantages” list |
| D. Short definition only if mandated | 4–6 | One term used in Blocks A–C, not a separate unit dump | — |
Map every item to a lecture date and a worked example. If you cannot point to a Tuesday when the class practised the choice, the item is a surprise, not a test of judgment. First-generation students will treat surprise as cruelty. Judgment items need rehearsal of the type of decision, not rehearsal of the exact numbers.
2.1 Tell the class the blueprint in week three
Publish the four blocks and the claim column. Do not publish the trap. Students then know that “write a short note on inventory methods” will not appear as a free-standing 8-mark gift. Coaching centres lose some of their edge because the photocopy no longer matches the claim column.
3. Build three item families that cannot be answered from last year’s paper
You do not need twenty original items. You need three families and the discipline to write one fresh member of each family every term.
3.1 Family 1 — Forced choice with a named constraint
Give two acceptable methods and one constraint that kills one of them. Example in cost accounting: a job-order firm that has just accepted a rush export order with a contractual delivery date. Ask which costing treatment the firm should use this week, and which figure in the data set is the constraint, not the cost. The student who writes both treatments in parallel and waits for the key has failed the item even if both workings are arithmetically pretty.
In financial reporting, give a lease fact pattern with one clause that determines classification. Ask for the classification and the clause. Do not ask for a 400-word comparison of Ind AS 116 and the old standard. Comparison essays reward memory of slide headings.
3.2 Family 2 — Computation that contains one unused number
Every recycled problem uses every number on the page. Students learn that unused figures do not exist. Put one plausible but irrelevant figure in the stem: a prepaid expense that does not belong in the working-capital cycle you asked for, or a market price that is not the measurement basis required. Award marks for striking that figure and naming why it is idle. A student who folds it into the ratio has failed a judgment test even if the calculator is correct.
3.3 Family 3 — Reject-and-condition paragraph
Cap this at 120 to 150 words. The stem must force a verb: reject the extra shift, accept the vendor’s 60-day terms only with a deposit, or refuse to capitalise the advertising campaign. The last sentence must attach a condition or a figure. Ban the words “in conclusion,” “holistic,” and “in today’s competitive scenario.” Those phrases are how students hide the absence of a decision.
A good mid-sem item is one you could hand to a junior accountant on Monday morning. If the item only makes sense inside a university unit list, it is still a memory item wearing a case costume.
4. Publish a marking scheme that separates arithmetic from judgment
If the key only lists the final number, markers will award the number and ignore the choice. Split every long item into three visible bands and print those bands on the question paper itself. Students then know where judgment lives.
| Band | Share of the item | What earns the band | What cannot earn it |
|---|---|---|---|
| Choice / treatment | 30% | Names the method or classification and the constraint | Listing both methods “for completeness” |
| Working | 50% | Correct units, one discarded figure labelled, arithmetic that follows the chosen method | A correct table under the wrong method |
| Condition / limit | 20% | One sentence on what would reverse the choice | Generic risk language with no figure |
Train co-markers for twenty minutes with three photocopied scripts: one memory dump, one clean arithmetic under the wrong method, and one short but correctly chosen working. Agree the band scores before the pile is split. Inconsistency across two teachers is how students conclude that judgment marking is mood.
For first-generation writers, mark English only when it changes the claim. A missing article is not a missing ratio. Circle the unit error; do not rewrite the paragraph into staff-room English and then deduct for style. Fluency can be a course outcome in the assignment file. It should not silently consume the judgment band in the internal.
5. Run the hall so the design survives the invigilator
A judgment paper dies in administration as often as it dies in item writing. Three practical controls matter in an Indian hall.
5.1 Stem on one side, workings book on the other
Print the data set so that neighbouring seats do not share the same column order if you run two versions. Version B can swap the unused figure and change the firm name. The blueprint stays identical. Students who came to copy a neighbour’s table discover that the idle number has moved.
5.2 Allowed artefact, named in advance
If the course claims to teach “professional practice,” permit one page of handwritten formulae with no worked example. Collect it with the script. A clean formula sheet rewards preparation of structure. A banned-calculator panic rewards the student who memorised last year’s layout. State the artefact rule in week one, not at the door.
5.3 Time that matches the judgment load
A 20-mark memory paper can be finished in forty minutes by a drilled student. A 20-mark judgment paper cannot. If the slot is sixty minutes, cap reading load at 350 words of stem plus one small table. Overlong cases in a mid-sem are how faculty accidentally reintroduce speed-as-merit. First-generation readers pay that tax first.
6. Return the script as a teaching hour, not a score dump
The internal is wasted if the only public artefact is a marks list on the notice board. Use one lecture after marking to treat three anonymised answers as a case.
- Project the stem and the printed band scheme.
- Show Answer A: complete table, wrong method. Ask the room which band dies first.
- Show Answer B: short working, correct choice, missing condition. Ask what sentence would have bought the last band.
- Show the unused figure and why it was idle. That five-minute reveal teaches more than a new chapter.
Return scripts with three marks only: a tick on the choice band, a circle on the unused figure, and one imperative in the margin—“Name the clause that kills operating-lease treatment.” That is the same written-feedback discipline this series has argued for first-generation commerce students: a next action, not a sermon. Students who want a longer comment can come to a twenty-minute window with the script open at the failed band. Do not re-teach the whole unit in office hours because a student is bargaining for two marks.
If the department requires a “model answer,” publish a model of the choice and the discarded figure, not a twelve-page solution that recreates the photocopy market for next year. A complete key is a gift to the coaching file. A partial key is a teaching tool.
7. Protect teaching load without collapsing back to MCQs
Judgment papers fail in Indian departments when the teacher who designs them also marks 240 scripts in four days. Design for marking time as a constraint, not as an afterthought.
Spend time here
Blueprint page; two versions of the unused figure; twenty-minute marker calibration; one return lecture.Refuse to spend time here
Rewriting every paragraph into elegant English; unique comments on 120 scripts; a new case narrative every unit.Cap the judgment paragraph at a third of a page. Number the lines. Markers then scan for the verb and the condition. Pair with a computation that has a binary discard decision. Those two structures mark faster than “discuss working capital management in India” and they discriminate better.
If you teach two sections, mark Section A’s Block B first, then Section B’s Block B. Fatigue is item-specific. Switching sections but staying inside one family keeps the standard still. Rotating randomly through whole scripts is how the last thirty papers become generous.
For NAAC files, keep the blueprint, the two versions, the calibration trio, and the return-hour slides. That packet is better evidence of assessment reform than a paragraph in the SSR that says “continuous internal evaluation is student-centric.”
8. Field checklist and closing brief
Run this list the week you set the paper, the morning of the test, and the afternoon you return scripts.
- Blueprint page exists, with four blocks and a claim column mapped to lecture dates.
- Class saw the claim column in week three; they did not see the unused figure.
- Every long item has a printed three-band scheme on the question paper.
- At least one figure in a computation stem is idle and must be labelled as idle.
- At least one item forces a reject-and-condition sentence under 150 words.
- Two versions exist if neighbouring seats can see tables.
- Allowed artefact was announced in week one and is collected with the script.
- Three calibration scripts were marked before the pile was split.
- Return hour used anonymised answers; the published key does not reprint a full working.
- Margin comments are imperatives tied to a band, not essays.
The mid-semester test is the one assessment most Indian commerce teachers still fully control. University finals will remain pattern-locked for years. Internals need not imitate that lock. A paper that forces a choice, discards a decoy number, and asks for a condition does more for employability language in the IQAC report than another guest lecture on “industry readiness.”
Start with one course this term. Keep last year’s computation if you must, but add the idle figure and the printed choice band. Students will complain that the paper was “twisted.” That complaint is evidence that the photocopy stopped working. Teach into that complaint in the return hour. By the second internal, the class will prepare differently, and the coaching file will be one season out of date.
Further reading on DrDPKlass: How to Give Written Feedback First-Generation Commerce Students Will Actually Use · How to Run the Case Method in an 80-Seat Indian Commerce Classroom Without Losing the Back Row.
Teaching Series · 24 September 2026