Switch course ▾
Session 4. Data in Teacher Teams: Meetings, Targets and Hard Conversations
Objective. Lead a 30-minute data meeting in a teacher team, agree one focus and a realistic improvement target, and handle difficult conversations about data with scripts and confidentiality.
Data only improve teaching when a group of people look at them together and agree what to do. Many staff-room conversations about results end in worry, blame or silence because nobody has a shape for the talk. This session gives you a 30-minute meeting that separates looking from explaining, a way to write a target that allows for noise, and tested wording for conversations that can go wrong. It also sets the confidentiality rules that make colleagues willing to bring weak results to the table.
Key ideas
A data meeting has a shape
- In 30 minutes: LOOK (5, say only what you see) · EXPLAIN (8, give several possible reasons, 'what else could it be?') · FOCUS (5, pick one) · ACT (10, target and action) · CLOSE (2, who, what, when).
- Roles: a facilitator who keeps the steps, a timekeeper, a scribe who writes on one sheet, and a 'what else?' voice. Rotate them each meeting.
- This follows the same logic as published team-data cycles such as the Data Wise process from Harvard (see readings): prepare, inquire, act. Keep it small and regular rather than grand and rare.
From a vague wish to a target
- 'Improve results' is a wish. A target names the group, the measure, the baseline, the size of change, the date and the action.
- Set the size with noise in mind. In 87 pupils one pupil is about 1 point; a gain of 3 or 4 points could be chance, a gain of 12 pupils is not. In a class of 20, ask for something larger or measure again.
- Use two kinds: a process target (what we will do, for example a routine in every lesson) and an outcome target (what we hope to see). Process targets are in your hands.
Scripts for difficult conversations
- Five moves: say the purpose, state a fact ('I notice...'), ask for the other view ('I wonder... what do you see?'), listen and summarise, agree one next step.
- Talk about data and pupils, not about character. 'The word-problem results are lower in this class' is a fact; 'you are not teaching it well' is a judgement that closes the door.
- When a colleague is defensive, agree on what is true ('Yes, your class started with more new arrivals') and then ask for the fair check together (start bands, absences, n).
Confidentiality and fairness in a team
- Agree norms before the first meeting: class figures stay in the team; pupils by code; no photos of pages; nothing goes to a messaging group; no publication of one teacher's results.
- Data about a colleague's class is not yours to carry to the head or to the staff room. If a head asks for a ranking, offer the team's improvement plan instead and explain why raw averages mislead.
- Keep the focus on what pupils need next. A team that trusts the process shares weak results sooner, which is exactly when help is most useful.
The 30-minute data meeting
Why separating look from explain works
When a table is put on the screen, people instantly begin to explain it: 'This class is weak', 'That teacher is strict'. Explanations feel like understanding, but they usually come from habit and from who is in the room. A rule that the first five minutes may contain only observations ('I see 15 of 29 pupils solved item 4') slows the group down just enough to see what is there. Teams that follow structured cycles, such as the Data Wise process described by Oberman and Boudett, spend deliberate time on 'what do the data say?' before 'why?'.
Writing a target the noise cannot fool
Start from the baseline and its n. In a combined group of 87 pupils a gain of 3 or 4 pupils can happen by chance between two similar tests, while 12 more pupils succeeding is much harder to explain away. For a single class of 20 the same logic says: expect larger gains, repeat the measure, or pool with a parallel class. Write the target so that the measure, its marking key and the date are fixed before the work starts, and choose a parallel version of the test for the second measurement so that pupils are not simply remembering answers.
Pair every outcome target with a process target the team controls, such as a routine in two lessons a week, and a mid-point check. Process targets keep people motivated when outcomes move slowly, and a mid-point check lets you correct course in week three instead of week seven.
Examples from the classroom
Three reasons on the board
Scene. A Grade 5 team in Fergana sees that word problems with decimals are solved by 11 of 30, 15 of 29 and 10 of 28 pupils. At 'Explain' they must write three reasons before anyone defends a class: reading load, no modelling routine, a harder test item than the practice. They choose to test the second reason, because it is the only one they can change in six weeks.
Insight. Several written reasons prevent a quick blame and make the next step testable.
A target for one class of eighteen
Scene. Malika in Bukhara has only 18 pupils in Grade 7. She wants to raise the share solving a fractions item from 8 of 18 to 12 of 18. Two pupils can swing the figure by 11 points, so she asks the parallel class (20 pupils) to use the same routine, pools the 38 pupils and sets the target for the pool. She plans a second check with a parallel item in week six.
Insight. With small groups, pool classes or ask for bigger, repeated evidence.
Words for a worried colleague
Scene. In Tashkent Jasur says quietly after a meeting, 'My class is always at the bottom.' Dilnoza, the team lead, does not reassure or lecture. She says: 'I notice your class has had the lowest item results for three terms. I wonder what you see in the lessons. Would a joint lesson help?' Jasur admits that six pupils joined in September, and they agree to compare start bands and to plan a joint lesson.
Insight. Fact, curiosity, offer: the three-part script keeps the door open.
Common pitfalls
- Spending the whole meeting explaining and leaving without a written target.
- Setting a target as a single number for 100% of pupils or for individual teachers.
- Using a different test or marking key the second time and calling the difference progress.
- Carrying a colleague's class figures to the head or the staff room without agreement.
A 30-Minute Data Meeting, Then Three Hard Conversations
- ACTIVITY A (40 min). Open the Activity Kit: Table 9 (a Grade 5 team's data on a decimals unit test, three classes) and the Meeting card. All sample data. Groups of four: facilitator, timekeeper, scribe, 'what else?' voice.
- Read the data silently (5 min), then hold the meeting exactly to the clock: Look 5, Explain 8, Focus 5, Act 10, Close 2 (30 min).
- The scribe writes on the Target sheet: focus, baseline, target with size and date, process target, who does what, how we measure again. Swap sheets with another group (5 min) and give one 'noise' comment on their target.
- ACTIVITY B (30 min). Trios with the three Conversation cards in the Kit. Three rounds of 7 minutes: speaker, colleague, observer; rotate roles. Use the five moves and the script frames. The observer writes one thing that worked and one phrase to change.
- Whole group (9 min): write a short 'team data agreement' (five norms) for your own school team, using the confidentiality card. Each person signs their copy for their team.
- Alone (5 min): write your school-based task from the Kit.
Your facilitator will guide you through this activity in the session.
Questions to discuss
- At which step did your group most want to jump ahead (to explanations or blame)? What helped you hold back?
- What did the target look like before and after the 'noise' comment from the other group?
- Which phrase from the scripts will you actually use in your school this month?
Reflect
- How do your team meetings about results usually end: with a task, a complaint or silence?
- Who in your school could facilitate a data meeting without anyone feeling judged?
- What would your team need in order to measure the same thing twice in a term?
Try it tomorrow
- Run the five-step meeting once with two colleagues on a recent test and write the Target sheet.
- Agree five confidentiality norms with your team before looking at any class figures.
- Practise the 'I notice, I wonder, would it help' script aloud with one trusted colleague.
Further reading and links
Look before you explain, choose one focus, write a target with a baseline, a size and a date, and keep colleague information inside the team.