Eight groups, one product: why the group work collapsed
You teach Natural Science to Grade 6 in a school in Shymkent. The class has 32 students, and the desks are bolted to the floor in five rows facing the board. For a lesson on photosynthesis you planned eight groups of four, each with a handout, a role card and 15 minutes to build a concept map on A2 paper. What happened: Group 1 finished in four minutes and then talked about football; Group 3 sat silent for 12 minutes and produced nothing; in Group 5 one student wrote everything while three watched; three students whose reading is documented at 1.5 years below grade level were never given a turn; you ran 12 minutes over and never collected the maps. The exit slips asked students to name the inputs and outputs of photosynthesis: 19 of 32 could not name more than one. Your deputy head has asked for a written reflection and a revised plan for the same objective next week.
1. Diagnose the failure. Which design decisions caused it, and which principle of learning theory does each violate?
Read the worked answer
The common conclusion - 'this class cannot do group work' - is not supported. The design, not the children, produced the outcome.
- No individual accountability. One A2 sheet per group means one student's work can carry four marks. This predicts both the domination in Group 5 and the loafing elsewhere; cooperative learning research is consistent that group products without individual accountability produce exactly this pattern.
- No positive interdependence. The handout could be completed alone, so there was no structural reason for members to need each other.
- Low ceiling and no checkpoints. Group 1 finishing in four minutes is evidence about the task, not the students: an open 15-minute task with no intermediate milestones lets fast groups disengage and slow groups stall.
- ZPD mismatch without mediation. A text 1.5 years above three students' reading level, with no glossary, audio, diagram version or pre-teaching, places those students outside the zone where assistance can work. Vygotsky's point is that the gap must be bridged by a more knowledgeable other or a tool; you supplied neither.
- Decorative roles. Role cards were not tied to the product, so they imposed no obligations.
- Physical arrangement. Bolted rows force groups of four to turn and shout, which converts noise into off-task behaviour.
- No collected product. Students correctly inferred the map did not count.
So 19 of 32 naming fewer than two inputs is not evidence against group work; it is evidence that no individual ever processed the content.
2. Redesign the same objective for the same 45 minutes and the same bolted-down room. Be concrete about structure, roles, timing and the three low readers.
Read the worked answer
Structure: jigsaw, four roles, individual output.
- Pre-teach six key terms (carbon dioxide, water, light, oxygen, glucose, chlorophyll) with pictures in the first 6 minutes, whole class. The three low readers get the diagram-only version of the text and hear the audio once during this phase.
- Four expert texts at three reading levels: 'what goes in', 'what comes out', 'what light does', 'where it happens'. Pairs are formed by turning two desks to face each other - groups of four become two facing pairs, which fits the room.
- Roles tied to the product: reader (reads aloud to the pair), recorder (writes), checker (tests the other pair with one question), reporter (gives the 20-second summary). Roles rotate at each checkpoint.
Timing with visible checkpoints. 0-6 vocabulary; 6-14 expert pairs read and complete a 4-line frame; 14-22 home groups teach each other, one minute per expert; 22-32 each student writes their own three-sentence summary on a strip - this is the individual product, collected; 32-40 pairs check against the textbook diagram and correct in a different colour; 40-45 exit slip.
Scaffolds. Sentence frames on the board ('Plants take in ___ and ___'), a labelled diagram as an alternative to prose, a timer projected so the class, not you, polices the phases, and a rule that the reporter is chosen at random by you.
Why this fixes the diagnosis. The individual strip supplies accountability, the jigsaw supplies interdependence, the checkpoints cap the ceiling, and the diagram version keeps the low readers inside the zone where support works.
3. What evidence will you collect to show the redesign worked, and what would make you conclude it did not?
Read the worked answer
Baseline and target. Baseline is 19 of 32 unable to name more than one input or output on the same exit slip. A defensible target is at least 24 of 32 naming both inputs and at least one output; below 20 is no improvement.
Three evidence sources, deliberately different.
- Individual exit slip, identical in format to the baseline, so comparison is possible. Score it against fixed criteria: three items correct, two, one, none.
- On-task sampling by a colleague or on video: count students visibly engaged at minutes 10, 20 and 30. Record the percentage, not an impression.
- Student talk: rough count of who speaks in the home groups, to check the three low readers had turns. A simple tally sheet is enough.
What would count as failure. If exit slips improve but on-task sampling stays low, the task is still misdesigned. If scores improve only for the strongest students, the scaffolding is not reaching the students it was built for. If everything improves once, do not conclude success: novelty is a plausible rival explanation, so repeat the same structure in a different topic within two weeks and require the gain to hold.
Reflection for the deputy. Write it as diagnosis, intervention, evidence, next step - not as an apology. The conclusion to defend is that the first lesson failed on design, not on the class.