Item analysis after a sandhi and samasa unit test
A 20-item grammar unit test is taken by 38 second-year students. Item analysis flags three items below 40% correct. For సూర్యోదయం, 24 students wrote 'సవర్ణదీర్ఘ సంధి'; the key says గుణ సంధి. For మాయిల్లు, 19 wrote 'సంధి లేదు' and 12 wrote 'సవర్ణదీర్ఘ సంధి'. For పద్మాక్షి, 22 wrote 'కర్మధారయ'; the key says బహువ్రీహి, and two students who wrote బహువ్రీహి explained it as 'పద్మములు కలవి'. Separately, three students split మహర్షి as 'మహ + ఋషి' and named it సవర్ణదీర్ఘ; the key accepts 'మహా + ఋషి, గుణ సంధి'. The teacher has one 40-minute period before the revision exam, and the students have already been shown the answer key and have had two days to argue with it.
1. Diagnose the single misconception that produces both the 24 wrong answers on సూర్యోదయం and the 12 wrong answers on మాయిల్లు, and state the check that would have prevented each.
Read the worked answer
The misconception: 'a vowel junction that changes the spelling is సవర్ణదీర్ఘ.' Students have learned to look at whether two vowels met, not at which two.
సూర్యోదయం = సూర్య + ఉదయం. అ + ఉ yields ఓ. ఓ is neither vowel lengthened - it is a third vowel. సవర్ణదీర్ఘ requires same-family vowels (అ/ఆ, ఇ/ఈ, ఉ/ఊ, ఋ/ౠ) producing the long form of that same vowel. అ + ఉ = ఓ is గుణ సంధి. Check to teach: write the two vowels, write the result, then ask 'is the result one of my two vowels, lengthened?' Yes → సవర్ణదీర్ఘ. No, a new vowel → గుణ.
మాయిల్లు = మా + ఇల్లు. Nothing lengthens and no vowel is lost, so 19 students concluded 'no sandhi'. But a య్ has been inserted - యడాగమ సంధి. The 12 who wrote సవర్ణదీర్ఘ were running the same over-generalisation as the first group. Check to teach: ask 'was a consonant born between the vowels?' If yes, it is an āgama sandhi, never a vowel sandhi.
Why these are one error, not two: in both cases the students classify by effect (spelling changed / changed a lot) instead of by operation (which vowels, and what happened to them).
2. The two students who wrote బహువ్రీహి for పద్మాక్షి explained it as 'పద్మములు కలవి', while the key's model explanation is 'పద్మముల వంటి కన్నులు కలది'. Full marks, part marks, or none - and why?
Read the worked answer
Full marks for the identification; no deduction. Put the canonical vigraha on the board, but do not take the mark back.
Reasoning. The discriminating feature of బహువ్రీహి is not the comparison - it is reference: the compound points to a third entity (the person who possesses the eyes), not to the lotuses and not to the eyes. The students' explanation 'పద్మములు కలవి' preserves exactly that: something/someone possesses lotus-eyes. They have demonstrated the one thing that separates బహువ్రీహి from కర్మధారయ, which is what the item was built to test.
What the 22 కర్మధారయ answers missed: they treated పద్మాక్షి as an attribution inside the compound ('lotus-eye') and never asked who the word finally refers to. That is the misconception to remediate.
Board line: 'Ask what the whole word names. If it still names the thing in the first member, it is కర్మధారయ; if it names someone who owns it, బహువ్రీహి.'
Marking principle: in a 1-mark identification item, credit the reasoning that discriminates, not the wording of the key. A key that only accepts one vigraha will generate avoidable appeals - which is precisely what the two-day argument has been about.
3. Build the 40-minute remedial period: what you teach, in what order, with timings, and what you deliberately leave out.
Read the worked answer
Order: operation first, labels last.
- 0-8 min - vowel chart drill. On the board: అ+ఉ=ఓ, అ+ఇ=ఏ, ఆ+ఋ=అర్, అ+ఆ=ఆ. Students say the two vowels and the result aloud for each. The chart, not the rule, is what they lack.
- 8-18 min - six live items, choral naming. రామాలయం, దేవేంద్రుడు, మహర్షి, మాయిల్లు, సూర్యోదయం, and one new one (e.g. మహేశుడు). For each, two questions only: which vowels met; did a consonant appear?
- 18-28 min - samasa reference test. Five compounds (పద్మాక్షి, నీలమేఘం, తల్లిదండ్రులు, పంచపాండవులు, రాజుగారి). For each, one question: what does the whole word name?
- 28-35 min - re-test of the three flagged items plus two new ones, closed book, collected.
- 35-40 min - answer-key reconciliation. Show the key, take only evidence-based challenges ('which vowels met?' matters; 'our teacher said so' does not).
Deliberately left out: a re-teach of all 20 items (18 already above 80% - re-teaching them trains students to ignore revision); the full theory of sandhi classification; and any new content. The period is for correcting one operation, not for covering the syllabus again.