Acting on item analysis in the grammar unit with three periods left
Your Class 12 Gujarati unit test has 40 candidates and 10 grammar items. The item analysis reads: sandhi item 3, 12 correct, with 22 of the 28 wrong answers landing on a single distractor; samasa item 5, 31 correct; vibhakti item 7, 9 correct, with wrong answers spread almost evenly across three options; nama/jati item 2, 34 correct. In the scripts, the 22 students who chose the single sandhi distractor had all joined the two vowels correctly but picked the wrong resultant vowel; on the vibhakti item they identified the dictionary meaning of the postposition but not the case relation it marks. You have three periods before the board examinations and a departmental instruction to 'reteach the weak chapters'. The scripts also show that students who scored above 7 on the other eight items still missed item 7.
1. Decide which items get class time out of the three available periods, and justify the allocation from the distractor pattern rather than from the percentage correct alone.
Read the worked answer
Allocation: item 7 gets two periods, item 3 gets one, items 5 and 2 get none.
Item 3 (12/40, one dominant distractor): 22 of 28 wrong answers on a single option is not noise. These students have learned the joining procedure and are failing at one decision point - which vowel results. That is a rule-level fix and one period is generous.
Item 7 (9/40, errors spread over three options): spread errors usually signal guessing or a badly built item, but here students who scored above 7 elsewhere also missed it. That rules out low ability and identifies an untaught concept: the mapping between postpositions and case relations. It needs two periods because it needs a new conceptual frame, not a correction.
Items 5 and 2: 31/40 and 34/40 mean the class can already do them; every minute spent there is a minute taken from item 7. Percentage correct alone would send you to the lowest scorer first; the distractor pattern tells you which lowest scorer is fixable in the time available.
2. Design a ten-minute diagnostic that separates two possible causes of the vibhakti errors, and state what each outcome would tell you.
Read the worked answer
Purpose: distinguish (a) not knowing what each case relation does from (b) knowing the relation but not which postposition marks it.
Part 1 - relation to postposition (4 minutes). Give five sentences with the same noun, છોકરો, each with a different postposition: ને, થી, માં, માટે, નો/ની. Students label the case relation of each underlined form (karma/sampradana, karana/apadana, adhikarana, sampradana, sambandha).
Part 2 - postposition from relation (4 minutes). Five items in the reverse direction: the case relation is given and students supply the postposition in a frame sentence.
Reading the outcome: strong on Part 1 and weak on Part 2 means the concept is intact and what is needed is a postposition-to-case table - one drill, not two periods. Weak on Part 1 with strong Part 2 means students are pattern-matching postpositions without meaning, and no amount of table practice will fix it; you must teach the relation with contrastive sentences. Even on both means the original item is faulty or vocabulary-bound.
Two minutes are left for collecting and sorting into three piles.
3. The department asks you to 'reteach the weak chapters'. Write your response, including the specific rule you will teach for the sandhi item.
Read the worked answer
Reteaching the whole sandhi chapter means re-covering material 31 of 40 students already control, and it spends the two periods that item 7 needs. The data localises the error to one decision: choosing the resultant vowel in svara sandhi. I will teach that decision alone.
Rule to teach: after the two vowels combine, the result follows a fixed hierarchy - dirgha sandhi when the same or same-quality vowel meets (વિદ્યા + આલય = વિદ્યાલય), guna sandhi when અ meets ઇ/ઈ, ઉ/ઊ or ઋ (સૂર્ય + ઉદય = સૂર્યોદય, where the result is ઓ), vriddhi sandhi when અ meets એ or ઓ. Students write the three result-sets on one card and classify ten pairs before joining any of them.
Accountability: I will report back with a five-item retest in the same format. If item-level performance does not move, the problem is my rule statement, not the students, and I will change the representation rather than add periods.