This Socratic seminar rubric uses three criteria with observable descriptions. Practice applying a descriptor to a short exchange before using it to review a full seminar. The example includes the complete source text, discussion prompt, and original dialogue.
Give the discussion a shared text
The original fictional source reads:
S1. The gallery extends Friday opening to 20:00 for a four-week trial.
S2. Weekday afternoon hours stay unchanged.
S3. Visitor counts and staff comments will be reviewed after the trial; results are not yet available.
The seminar prompt is:
What would you need before recommending a permanent Friday extension?
A teacher-authored seminar lesson hosted by UC Berkeley includes expectations for supporting claims with evidence, developing other speakers' ideas, and considering differing views and source issues. The three-level descriptions and gallery exchange below are original examples for this task.
Use descriptions students can apply
These are local discussion criteria, with Not observed available separately for each one:
| Criterion | Level 1 | Level 2 | Level 3 |
|---|---|---|---|
| Textual reasoning | States a position without relevant text support, or misrepresents the text. | Identifies relevant text but leaves the link to the claim partly unexplained. | Explains how specific text supports the claim and recognizes a relevant limit. |
| Developing a contribution | Repeats a position without engaging with the prior idea. | Refers to a prior contribution and adds a relevant point. | Identifies the prior idea and develops, qualifies, or challenges it with a reason. |
| Productive inquiry | Asks a question that assumes an unsupported answer or closes off examination. | Asks a relevant question about the text or argument. | Asks for evidence or a distinction that could advance the interpretation. |
“Not observed” means the available observation does not support a judgment for that criterion. It is not the lowest performance level. An excerpt may show textual reasoning while giving you little evidence about how that speaker develops others' ideas.
Share the descriptors before the discussion. Students should know that the rubric attends to what a contribution does, rather than simply how many times someone speaks.
Calibrate the descriptors against an exchange
The fictional dialogue is:
A: “It should stay because everyone likes evenings.”
B: “S1 says four-week trial, and S3 says results come afterward; I would wait for those records.”
C: “Building on B, which count comparison would help distinguish new visits from visits moved from afternoons?”
Select Textual reasoning and compare A with B. A's universal claim about everyone's preference has no support in the three sentences. It fits the Level 1 description for this excerpt.
B identifies the trial and the pending results, then connects them with a reason to wait for records. The contribution explains a source-based limit on recommending a permanent change, matching the Level 3 textual-reasoning description.
Next, select Developing a contribution for C. C explicitly picks up B's idea about reviewing records and adds a distinction between new visits and visits shifted from afternoons. That develops the earlier point with a reasoned question, matching Level 3 for this criterion.
C's question also illustrates Level 3 productive inquiry: it identifies the kind of comparison needed to examine a proposed benefit. The passage supplies no count results, so the rubric explanation should not answer the question for the group.
Keep the evidence beside the descriptor
For an actual observation, record the criterion, the relevant words, the descriptor they support, and a next step. The example's blank review requires an evidence note before assembling that local review text.
| Criterion reviewed | Evidence from this excerpt | Useful next practice |
|---|---|---|
| A's textual reasoning | “Everyone likes evenings” supplies no passage support. | Replace the universal claim with a claim the text supports, or frame the preference as a question to investigate. |
| C's development of B | “Building on B” introduces a question about what the counts would establish. | Use the same approach to explain what a proposed evidence source could and could not show. |
These judgments concern supplied excerpts. A whole-seminar review needs the observation range and the evidence your class's assessment process requires. Keep other criteria Not observed when the excerpt does not address them.
Make a rubric for your own text
Create one complete, self-contained HTML Socratic seminar rubric.
Learner group and purpose: [FILL IN]
Original or permitted source text with references: [PASTE]
Seminar question: [PASTE]
Criteria and observable level descriptions: [PASTE]
Fictional practice exchange, if wanted: [PASTE]
Checked descriptor-to-excerpt explanations: [PASTE]
Actual review evidence, if available: [PASTE OR LEAVE BLANK]
Show the source, question, criteria, and complete level descriptions.
Keep Not observed separate from the lowest level. Explain the quality
of the reasoning or contribution without converting speaking frequency,
eye contact, or confidence into an evidence score.
Let a reader select a criterion and supplied excerpt to inspect which
observable words support a descriptor. Label practice dialogue fictional.
Keep an excerpt-specific judgment separate from a whole-seminar grade.
Do not automatically evaluate arbitrary typed prose or invent evidence.
Include a blank criterion/evidence review with a proposed next practice.
Require an evidence note before assembling its review copy. Preserve
notes when switching criteria, mark old review output stale after edits,
and ask before resetting typed work. Keep input local with no claimed
saved student record, submitted grade, or assessment certification.
Use text of at least 16px, labeled keyboard-accessible controls, and
a layout readable at 360px without horizontal scrolling. Put CSS and
JavaScript inline with no external dependencies. Return complete HTML.
Make the expectation clear before the seminar
Ask students to compare A and B, then explain what changed in C's contribution. Their discussion can reveal a vague descriptor before it becomes a grading disagreement.
If the rubric is an HTML page in NOS, revise the instruction above the criterion selector while reviewing the worked exchange. Ask students to identify the earlier idea and explain whether the next speaker develops, qualifies, or challenges it with a reason. Give students that shared instruction before the next discussion.