Prepare for a collaborative discussion, follow its rules, ask questions that elicit elaboration, and modify a view when the evidence warrants it

Learning goal: Engage effectively in a range of collaborative discussions, one-on-one, in groups and teacher-led, with diverse partners on a grade 7 text or issue. Come to the discussion prepared, having read or researched the material under study, and draw explicitly on that preparation by referring to evidence. Follow rules for collegial discussion, track progress toward specific goals and deadlines, and define individual roles as needed. Pose questions that elicit elaboration, respond to others with relevant observations that bring the discussion back on topic, acknowledge new information expressed by others, and modify your own view when it is warranted. Build on others' ideas and express your own clearly.

Original article — A Rooftop Garden at Birch Lane?

Birch Lane Middle School is invented, and every number in this article is hypothetical. The figures exist so that a class can practice discussing evidence; nobody should memorize them as facts about a real school. Paragraphs are numbered so that a discussion can cite them.

(1) In January the Birch Lane student council proposed turning the flat roof of the cafeteria wing into a vegetable garden. The idea came from a science unit on plant growth. In a survey of 310 students, 187 supported a rooftop garden, 74 opposed it and 49 were undecided.

(2) Supporters point to learning. The seventh-grade science teacher wrote to the council that a garden could host four lessons a year for each grade, on soil, water, pollination and food webs, and that students who plant something tend to remember the unit. She added that she had not measured this and was reporting her impression from eleven years of teaching.

(3) Three concerns have been raised. First, cost. A landscaping company estimated 42,000 dollars for raised beds, soil, a watering system and a safety railing. Second, structure. The district engineer said the roof must pass a load check before anything is placed on it; the check costs 3,500 dollars and takes about six weeks. Third, access. The head custodian pointed out that the only way onto the roof is a maintenance ladder, so a proper staircase would be needed before any student could go up, and the landscaping estimate does not include one.

(4) The custodian also mentioned an alternative. An unused courtyard beside the library gets sun for most of the day and could hold six raised beds at ground level for roughly 9,000 dollars, with no load check and no staircase. It is about a third of the roof's area.

(5) The council must recommend one of three options to the principal by March 15: the rooftop garden, the courtyard garden or no garden this year. The recommendation must name the evidence it relies on and list what the council still does not know.

What a collaborative discussion is. A collaborative discussion is a conversation in which people reason together about a topic, text or issue so that the group ends with a better-supported position than any member began with. It is not a debate that someone wins and it is not a circle of opinions in which nothing changes. The purpose explains every rule that follows. The range of settings matters too: in a one-on-one exchange you carry half of the thinking; in a group of four you must track three other voices; in a teacher-led discussion the teacher manages turns and you must connect your comment to whoever spoke before you. Diverse partners, people whose experience and reading of the article differ from yours, are the point, not an obstacle, because the group only gets smarter if someone brings something you did not.

Come prepared, and use the preparation (SL.7.1a). Preparation means you have read or researched the material under study before you sit down, and you arrive with notes that point to evidence by location. For the article, a prepared student has marked at least three items: the survey split of 187, 74 and 49 (paragraph 1); the teacher's admission that her learning claim is an impression, not a measurement (paragraph 2); and the missing staircase in the 42,000-dollar estimate (paragraph 3). Preparation only counts if you draw on it explicitly. "I like the roof idea" gives the group nothing to test. "Paragraph 3 says the estimate leaves out a staircase, so 42,000 is a floor, not a total; before we compare it with 9,000 we need the staircase cost" refers to evidence, probes it and invites the group to reflect on what the figure can and cannot show.

Follow the rules, track the goal, define roles as needed (SL.7.1b). Collegial rules let people disagree about ideas without attacking each other: one speaker at a time, paraphrase a claim before you challenge it, criticize the claim and not the person, and let a speaker finish. Tracking progress means someone keeps the specific goal and the deadline in view. The council's deadline is March 15; a class group modeling it should set its own, for example "recommend one option and list two unknowns in twelve minutes," and check at the halfway point which options have not been discussed. Roles are defined as needed, not by habit. A pair rarely needs any. A group of five with three options and a clock benefits from a facilitator who invites quiet members in, a recorder who writes decisions and unknowns, and an evidence checker who returns to the article whenever someone quotes it from memory.

Pose questions that elicit elaboration; respond and bring the talk back (SL.7.1c). A question that elicits elaboration asks the speaker to explain, give evidence or an example, or clarify the thing that matters. Its grammar is not the test: "Can you explain why six beds are enough?" could be answered yes, yet it plainly asks for an explanation, while an open-sounding "What do you think?" can invite a shrug. "Do you like the courtyard?" asks for a verdict and nothing more. "You said the courtyard is a third of the roof's area; how many of the science teacher's four lessons would still work in six beds?" asks for elaboration tied to the text. Responding means answering the question or comment that was actually asked, with a relevant observation, not with a restatement of your own view. Discussions drift; when someone begins describing their grandmother's tomatoes, the collegial move is a relevant observation that returns to the topic: "That reminds me that watering is in the roof estimate, but is anything said about watering the courtyard?"

Acknowledge new information and modify your view when warranted (SL.7.1d). When another speaker adds information you did not have, say what it is and what it does to your position. Sometimes it changes your view: a student who favored the roof learns that the load check alone takes six weeks and the estimate omits the staircase, and moves to the courtyard. Sometimes it does not, and then you say why: "The staircase is a real cost, but the survey shows 187 of 310 students wanted the roof specifically, so I still think we should price the staircase before giving it up." Both are honest. What is not honest is repeating your conclusion as though the new information had not been spoken.

Make comments that further the discussion: elaborate on and draw conclusions from the remarks of others. A comment furthers a discussion when the group knows more, or can decide more, after it than before it. Two moves do this reliably, and both start from what someone else said rather than from your own notes. To elaborate on a remark is to take another speaker's point and extend it: add a detail from the text that supports it, spell out a consequence the speaker did not state, or show where else it applies. To draw a conclusion from the remarks of others is to put two or more comments together and say what follows from them that none of the speakers said alone. Both moves name the speaker whose remark you are building on, so the group can follow the thread: "Building on what Rosa said..." or "If Ben is right about the survey and Malik is right about the timing, then..." A comment that only agrees ("Yeah, what she said") or restarts with an unrelated point of your own does not further the discussion, even when it is true. The model below uses both moves. Ben elaborates on Amara's survey figure by pointing out what the survey question did and did not ask, and the recorder's final entry is a conclusion drawn from several speakers' remarks, not from any one of them.

The misconception. Many students believe a good discussion is one in which they say a lot. It is not. A member who speaks four times, each time citing a paragraph, asking one question that elicits elaboration and changing one position for a stated reason, contributed more than a member who spoke twelve times without evidence. The measure is whether the group's final position is better supported than its starting positions.

A worked model in prose. Imagine four students with the goal "recommend one option and two unknowns in twelve minutes." Amara opens with the survey: 187 of 310 favor the roof. Ben responds with a relevant observation, not a rival opinion: the survey asked about a roof garden, not about a courtyard garden, so the 187 may or may not prefer the roof over the courtyard. Chloe asks a question that elicits elaboration: "Ben, what would the survey have needed to ask for us to use it?" Ben answers: a question offering both gardens. Amara acknowledges the new information and modifies her view: the survey shows support for a rooftop garden, but not a preference for the roof over the courtyard, which nobody was asked about. Dev, tracking the clock, notes six minutes remain and that cost has not been discussed. The recorder writes: recommendation, courtyard; unknowns, the staircase cost and whether six beds fit four lessons. The group's position is better supported than any member's opening line, which is the whole point.

Worked example 1 — an exchange with the moves labeled. Rosa: "Paragraph 3 says the load check costs 3,500 dollars and takes six weeks. If the council recommends the roof on March 15, nothing can be placed on it until at least late April." (Draws on preparation with a location and probes a consequence.) Malik: "So does the timing rule out planting this spring, or just delay it?" (Poses a question that elicits elaboration.) Rosa: "Delays it, but the estimate also leaves out a staircase, so I do not know the total cost." (Responds to the question asked; states an unknown.) Malik: "I came in favoring the roof. Given the check and the staircase, I now favor the courtyard for this year and the roof as a later goal." (Acknowledges new information and modifies his view, with the reason stated.)

Worked example 2 — a role and goal plan for a group of five. Goal: recommend one option and list two unknowns. Deadline: twelve minutes, with a check at six. Facilitator: invites each member to give one piece of evidence with its paragraph before anyone gives an opinion. Recorder: writes each decision and each unknown in two columns. Evidence checker: when Priya says "the teacher proved gardens improve memory," reads paragraph 2 aloud, which says the teacher called it an impression and had not measured it. Timekeeper: at six minutes announces that cost has been covered but access has not. Fifth member: floats between roles. The plan works because every role is tied to the goal; a role nobody needs is dropped.

Worked example 3 — bringing a discussion back on topic. The group is comparing the 42,000-dollar and 9,000-dollar estimates when Theo begins describing a garden at his cousin's school that had raccoons. A weak response ignores him or says "that is off topic." A collegial response makes a relevant observation that connects his story to the issue: "Raccoons raise maintenance, for the courtyard and perhaps the roof too; the article says nothing about maintenance for either option, so that is a third unknown." The group is back on the text, Theo's contribution has been used rather than dismissed, and the recorder has a new entry.

Worked example 4 — elaborating on a remark and drawing a conclusion from two. Priya: "The courtyard is about a third of the roof's area." Jonah: "And paragraph 2 says the science teacher wants four lessons a year for each grade." Rosa elaborates: "Building on Priya's point, paragraph 4 says a third of the area still holds six raised beds, so the question is not whether the courtyard is smaller but whether six beds are enough." (She extends Priya's remark with a detail from the text and states what it implies.) Malik draws a conclusion: "Putting Priya's and Jonah's points together: if the courtyard is a third of the roof and the teacher wants four lessons for each grade, then before we recommend it we need to ask her whether six beds can serve all four lessons, and none of us can answer that from the article." (He states something that follows from two speakers' remarks and that neither said alone.) The recorder adds the question to the unknowns. Neither comment repeats a speaker's point; each moves the group closer to its goal.

Discussion preparation sheet: Before a group discussion of the article, complete a sheet with three columns: evidence I bring (each with its paragraph number), two questions that elicit elaboration and name a speaker or a claim, and my current position with the condition that would change it. During the discussion add a fourth column: new information I heard, and whether I modified my view or justified keeping it. A strong sheet has at least three evidence entries, two questions that ask a named speaker to explain, give evidence or clarify, and one position with a stated condition.

Independent performance: Take part in three settings on the same issue: a one-on-one exchange, a group of four or five with a goal, a deadline and roles, and a teacher-led whole-class discussion. An observer with a checklist records whether you referred to evidence by location at least twice, followed the agreed rules, tracked progress toward the goal or held a role, posed at least one question that elicited elaboration, responded to a question with a relevant observation rather than a restatement, brought the discussion back on topic at least once, and acknowledged new information by modifying or justifying your view with a reason. Success means every item is observed at least once across the three settings and you can write afterwards, in three sentences, one place where another speaker's evidence changed what you said. The written questions below check understanding; they do not replace the observed discussion.

Building-on item, with a model answer: Theo says, "My cousin's school garden had raccoons." Amara says, "The article gives no maintenance plan for either option." Write one comment that elaborates on one of these remarks and one that draws a conclusion from both. Model answer: Elaborating on Amara: "Building on Amara's point, paragraph 3 lists raised beds, soil, a watering system and a safety railing in the roof estimate, and paragraph 4 gives the courtyard's cost, sun and size; neither says who would weed, water during the summer or deal with pests." Drawing a conclusion from both: "If Theo's raccoons are a real risk and, as Amara says, neither option has a maintenance plan, then the council cannot yet compare the full cost of the two options, so a maintenance plan belongs on the list of unknowns for March 15." The first comment adds text evidence to one speaker's remark; the second combines two remarks into a conclusion neither speaker stated.