This module develops scientific writing, criterion-based manuscript review, and responsible use of AI in writing and review. Students learn to write for a reader’s expectations rather than their own memory of the work, to revise an argument rather than polish a surface, to read and produce a real peer review, and to verify every claim a tool hands them before it enters a manuscript.
Teaching pattern: Each 80-minute meeting has a bounded source or anchor task, a short application or decision task, a concrete output, closure time, and an explicit contingency reserve. Published papers, review files, and author responses are used to inspect how writing and review actually work—not to infer which specific sentence caused an acceptance.
| # | Lecture | Key Topics and Output |
|---|---|---|
| 1 | Reading, Reporting, and the Reader’s Work | Reader-expectation prose; CARS moves; reporting-guideline orientation; portfolio topic, audience, and scope |
| 2 | Writing Principles and Their Boundaries | Whitesides and lab guidance; testing a rule against a real corpus; Draft 1 |
| 3 | Revision as Argument Restructuring | A public published abstract and a clearly constructed comparison; number auditing; Draft 2 |
| 4 | Peer Review Systems, Venue Requirements, and What They Publish | Real referee reports; two eLife regimes; assessment-vocabulary calibration; venue-property and version check; independent review |
| 5 | Responding to Review | A published point-by-point rebuttal; evidence-based disagreement; response letter |
| 6 | AI in Scientific Writing: Verification, Disclosure & Submission | Public-input AI audit with direct-use and no-direct-use paths; verification note; final revision sprint and submission readiness |
The module deliverable is a team Writing Portfolio: a research abstract that is drafted, revised, formally reviewed, answered, and submitted with a documented revision history, plus a public-source AI-feedback audit whose verification lessons are applied to the final submission without uploading the team draft.
| Lecture | Milestone | Due |
|---|---|---|
| 1 | Topic, target audience/venue, and scope-and-permission statement | In class |
| 2 | Draft 1: conclusion-first abstract with a named central claim | Oct 19 |
| 3 | Draft 2: revised for argument structure, specificity, and number discipline | Oct 21 |
| 4 | Venue-property and version check for the target venue, recorded with sources | In class |
| 5 | Peer review of another team’s Draft 2, plus a point-by-point response to the review your team received | Oct 28 |
| 6 | AI feedback audit using a public source, with a complete verification record; final revision sprint, one-page revision history, and submission-readiness sign-off | Nov 4 |
Scope and permission. Teams may write about a member’s own research only if that member is willing to share it with classmates and has any co-author or advisor permission required. A public, published, or instructor-created alternative topic is always available. Nobody is required to disclose confidential, unpublished, proprietary, or embargoed work in order to complete this module. Permission to share within the course does not authorize uploading the work to an external AI service.
Contribution record. Because the portfolio is a team product about work that usually belongs to one member, each submission carries a short contribution note identifying who drafted, who revised, who verified sources, and who signed off. This uses the procedure the team already wrote in Module 1 Charter Appendix C. Authorship eligibility and order are Module 4 topics; this note only records what each person did.
Goal: Establish that clarity is a property of the reader’s experience rather than the writer’s intention, then choose a portfolio topic, audience, and scope that the team can actually write about.
Interpretive limit: CARS is a description of moves that Swales observed in published introductions, not a rule a journal enforces and not a guarantee that a paper containing all three moves is good. Gopen and Swan’s claims are about reader expectation in English scientific prose; they are argued from worked examples, not from a controlled experiment, and they do not transfer automatically to every language, genre, or section. Use both as diagnostic lenses students can apply and check, not as scoring rubrics.
| Time | Activity |
|---|---|
| 6 min | Frame the distinction: what the writer knows, what the text states, and what a reader can reconstruct from the text alone |
| 17 min | Tiered reading: after common instructions, teams receive the same paper at staggered start times so all stop together after an assigned 2, 5, 10, or 15 minutes; nobody answers during the reading window |
| 10 min | In the same post-reading window, all teams answer the same five questions without reopening the text and record which questions their assigned depth could and could not answer; the class tabulates the result |
| 14 min | CARS move audit: mark Moves 1–3 in the assigned introduction, then mark one sentence whose topic position or stress position works against the reader |
| 11 min | Select the portfolio topic, name the target audience and a plausible venue, and write the one-sentence claim the abstract must support |
| 5 min | Reporting-guideline orientation: search the EQUATOR Network library for a guideline matching the study type, or record that no fitting guideline was found; the applicable reporting items shape what the abstract must eventually support |
| 8 min | Scope-and-permission check: confirm sharing permission, or switch to the public/instructor-created alternative; record the contribution plan |
| 5 min | Exit ticket: one reading depth that was sufficient for a stated purpose, and one that was not |
| 4 min | Transition and contingency reserve, distributed across the activities above |
| 80 min | Total |
The tiered reading task produces a small classroom observation, not a finding about reading in general. The class records which of the five questions each depth answered correctly, with the number of teams at each depth as the denominator. That table is the evidence. It supports statements such as “in this class, on this paper, the abstract-only group could not answer the question about controls.” It does not support claims about optimal reading strategy, about other papers, or about other readers.
Goal: Apply a conclusion-first, outline-driven drafting workflow, and learn to distinguish a well-argued practice from a field-wide rule by testing a specific stylistic claim against real published papers.
Attribution discipline: Advice circulates faster than its evidence. Before repeating a rule in your own writing group, name who states it, in what source, and for what kind of paper. Guidance associated with a specific research group describes that group’s practice; it does not establish a norm across fields, and a rule that serves a physical-chemistry letter may not serve a methods paper or a review.
The instructor supplies three short, preselected excerpts and an operational definition of one observable feature—for example, parentheses per sentence, passive constructions under a stated coding rule, or whether a discussion opens with the principal finding. Each team records:
The output is a one-sentence rule with a stated boundary, not a verdict that the rule is right or wrong. Three excerpts cannot establish a field-wide convention. Process claims such as whether an abstract was drafted first or last require version history or author evidence and cannot be inferred from a finished paper.
| Time | Activity |
|---|---|
| 9 min | Principles deck: the outline as instrument, conclusion-first drafting, and figures as the argument’s spine |
| 14 min | Rule test against three preselected excerpts, using one operationally defined observable feature |
| 14 min | Write the conclusion first: one or two sentences naming the central claim and what supports it |
| 14 min | Rebuild a supplied weak paragraph using topic and stress positions from Lecture 1 |
| 11 min | Peer audit: does the abstract state a claim, or only a topic? Is every number attached to a comparison? |
| 8 min | Assemble Draft 1 and record the contribution note |
| 5 min | Exit ticket: one rule with its scope condition, and one thing the team’s draft still cannot support |
| 5 min | Transition and contingency reserve, distributed across the activities above |
| 80 min | Total |
Output: Writing Portfolio Draft 1, due October 19 on Blackboard. Draft 1 must contain an explicit central claim, at least one figure concept described in words, and no quantitative claim the team cannot source.
Goal: Map the argument structure of a public published abstract, distinguish structural revision from surface editing through a clearly constructed comparison, audit every number for its referent, and apply the same lenses to the team’s own draft.
The source text is the final published abstract of:
Hyunwoo Yuk, Teng Zhang, Shaoting Lin, German Alberto Parada, and Xuanhe Zhao, “Tough bonding of hydrogels to diverse non-porous surfaces,” Nature Materials 15, no. 2 (2016): 190–196, doi:10.1038/nmat4463. Published online November 9, 2015. Accepted manuscript available via PubMed Central, PMC4762474.
The course comparison text is constructed for teaching from facts in the public accepted manuscript. It deliberately presents the same public claims in a weaker order so students can diagnose structure without treating an unpublished draft as evidence. It is not an earlier manuscript version and must not be described or cited as one.
| Move in the published abstract | Work it does for the reader |
|---|---|
| Biological benchmark and unmet synthetic challenge | Establishes the comparison and the problem |
| General bonding strategy | States the approach before procedural detail |
| Interfacial mechanism, including silanation | Explains why the approach can work |
| Greater-than-1,000 J m⁻² result | Reports the achieved performance against the benchmark |
| Materials and applications | States the demonstrated scope without turning it into a universal claim |
Evidence boundary: The public abstract establishes what the published paper says. The constructed comparison supports a classroom revision exercise only. Neither source reveals the paper’s actual drafting history, the order in which its authors wrote sentences, or which wording influenced editorial decisions. Lecture 4 uses a different paper whose referee record is public when the course needs evidence about review.
The published abstract contains two toughness values that are easy to confuse, and students must separate them:
A number in an abstract must carry its referent, its comparison, and its direction. Reporting “800 J m⁻²” as the paper’s achievement inverts the argument: the benchmark becomes the finding, and the finding disappears. Each team audits every number in its own Draft 1 the same way.
| Time | Activity |
|---|---|
| 8 min | Provenance briefing: distinguish the public abstract, the constructed comparison, and claims neither can support |
| 12 min | Map the published abstract’s problem → strategy → mechanism → result → scope sequence |
| 12 min | Compare the constructed weak version with the published sequence; classify differences as specificity, structure, scope, or surface |
| 10 min | Number audit on the published abstract: for each value, name the referent, the comparison, and who achieved it |
| 17 min | Apply both lenses to the team’s Draft 1: restructure the argument, then audit every number |
| 8 min | Cross-team check: can another team state your central claim and name your comparison after one read? |
| 5 min | Exit ticket: one structural change made and the reason it is not merely cosmetic |
| 8 min | Transition and contingency reserve, distributed across the activities above |
| 80 min | Total |
Output: Writing Portfolio Draft 2, due October 21 on Blackboard, with a two-sentence note naming the one structural change the team considers most consequential.
Goal: Read an actual set of referee reports, compare what two different journal models make public, check the target venue’s actual stated requirements, and write an independent criterion-based review of another team’s draft.
All three records below are public and can be read in full before class.
1. A bioinspired paper with a detailed published referee file. Danli Luo, Aditi Maheshwari, Andreea Danielescu, Jiaji Li, Yue Yang, Ye Tao, Lingyun Sun, Dinesh K. Patel, Guanyun Wang, Shu Yang, Teng Zhang, and Lining Yao, “Autonomous self-burying seed carriers for aerial seeding,” Nature 614 (February 16, 2023): 463–470, doi:10.1038/s41586-022-05656-3. Springer Nature publishes the Peer Review File (PDF, 96 pages): reports from three referees across three rounds and the authors’ point-by-point responses. The paper’s acknowledgement thanks Barbara Mazzolai, Naomi Nakayama, and another anonymous reviewer; the public file does not map names to report numbers. The instructor is a co-author and can distinguish personal recollection from what the published record establishes. See the case brief and source-linked extracts for the assigned passages.
2. eLife under its earlier model. Victor Kang, Robin T. White, Simon Chen, and Walter Federle, “Extreme suction attachment performance from specialised insects living in mountain streams (Diptera: Blephariceridae),” eLife 10 (2021): e63250, doi:10.7554/eLife.63250. Publishes a consolidated decision letter with essential revisions and the author response. Reviewers anonymous; reviewing and senior editors named.
3. eLife under its current model. Natalia Jagielska, Thomas G. Kaye, Michael B. Habib, Tatsuya Hirasawa, and Michael Pittman, “New soft tissue data of pterosaur tail vane…” (Reviewed Preprint v2, November 29, 2024; doi:10.7554/eLife.100673.2). Publishes an eLife assessment using controlled vocabulary for significance and for strength of evidence, alongside public reviews. Its assessment calls the contribution “valuable” and the methods “compelling“—two terms with published definitions rather than free prose. A later Version of Record was published December 18, 2024; the version number is pinned here because the assessment exercise uses v2.
eLife publishes separate term sets and definitions for significance and strength of evidence. For the calibration exercise, the instructor supplies the terms in randomized order and withholds the definitions and intended order until every student has recorded an individual ordering. This section and the linked deck are revealed only after that first response.
What an isolated-phrase experiment found. Thomas E. Hardwicke, Sarah R. Schiavone, Beth Clarke, and Simine Vazire, “An empirical appraisal of eLife’s assessment vocabulary,” PLOS Biology 22, no. 8 (2024): e3002645, asked 301 individual participants holding doctoral or graduate degrees to rate short hypothetical statements on 0–100 scales. The implied ranking matched eLife’s intended order for 20% [15–24%] on significance and 15% [11–20%] on strength of evidence. A more explicitly ordinal alternative reached 62% [57–68%] and 67% [62–72%] (odds ratios 8.17 and 11.4, both p < .001).
This tests how individuals interpreted isolated phrases under the study’s conditions. It does not test team deliberation, actual editorial assessments embedded in explanatory prose, reviewer quality, or the overall value of eLife’s publishing model. The intended ranking was the most common, and typical deviations were small. The assessment is also not a rubric: eLife instructs editors and reviewers to include terms “when appropriate,” and a single assessment may mix levels across a paper’s parts.
Students first rank the bare terms individually without seeing eLife’s definitions, then compare their ranking with the intended order and the study. Applying the terms to an abstract is a separate classroom task and is not compared numerically with Hardwicke’s individual phrase-ranking result.
| Nature peer review file | eLife, earlier model | eLife, current model | |
|---|---|---|---|
| Referee reports published | This article’s file includes reports and rebuttals across three rounds | Consolidated into a decision letter | Yes, as public reviews |
| Reviewer identities | The article acknowledges two by name and another as anonymous; report numbers are not mapped to names | Anonymous; editors named | Anonymous unless the reviewer signs |
| Author response published | Yes, point by point | Yes | Yes, when provided |
| Accept/reject decision | Yes, and it precedes publication | Yes | Editorial invitation gate, but no post-review accept/reject decision; assessment plus versions |
| Controlled assessment vocabulary | No | No | Yes, for significance and strength of evidence |
Read the source, not the summary. Journal models change. eLife’s model changed substantially after 2022, and Springer Nature has extended transparent peer review across its titles. Confirm each journal’s current practice on its own policy page before teaching or citing this table.
Publishing models are not judged only on their merits; they interact with the infrastructure that counts them. eLife’s decision to stop making accept/reject decisions had a measurable institutional consequence:
This is a concrete case for the class: a journal changed its post-review publication decision, and an indexing policy that required a validation decision no longer treated every Version of Record as eligible. Students should be able to distinguish eLife’s design choice from Clarivate’s eligibility rule and ask what this interaction implies for reforming assessment and publishing at the same time.
Selectivity, review model, indexing, and access model are distinct properties that may interact. Before this class, each team checks its own target venue (selected in Lecture 1) against current primary pages:
What the public record lets a reader check: the version, date, DOI, venue, author list, and stated article history; the venue’s published criteria; the methods, data-availability statement, disclosures, and corrections that are actually published; and the venue’s current statements about review, access, scope, and fees. What it does not establish: that every contribution appears in the byline, that reviewers agreed or every concern was resolved, what an editor privately weighted, that a result is true or will replicate, or that reputation accurately predicts rigor or indexing for this article.
| Time | Activity |
|---|---|
| 6 min | Briefing: what each review model chooses to publish and what no model guarantees |
| 12 min | Read an assigned extract of the Nature referee reports; classify each comment as evidence requested, analysis questioned, presentation, or scope |
| 6 min | Locate the referee comment that proposed the paper’s final title; discuss what it means that framing can change through review |
| 8 min | Before seeing definitions, rank the bare eLife terms individually; then reveal the intended order and read the bounded Hardwicke result |
| 9 min | Apply the defined terms to a supplied abstract and defend each choice from its text; state why this is not the Hardwicke task |
| 10 min | Module 3 Microcase 1: “The Review That Was Never Closed”—8-minute case plus 2-minute debrief |
| 12 min | Write an independent review of another team’s Draft 2: one strength, one material weakness, one specific request, each tied to text |
| 7 min | Venue-property check: from the team’s pre-class notes, confirm review/access/scope/fees from venue pages, indexing from the named index, selectivity only with a reported denominator, and the exact version cited; record every source or mark not publicly reported |
| 5 min | Exit ticket: one venue requirement your team had assumed rather than checked |
| 5 min | Transition and contingency reserve, distributed across the activities above |
| 80 min | Total |
Reviews address the draft, not the author. Do not speculate about the team’s ability, effort, or intent.
Goal: Analyze a published point-by-point rebuttal, then write a response that agrees, partially agrees, or disagrees with evidence—without conceding a point the record does not require conceding.
The Lecture 4 Nature Peer Review File contains the authors’ published responses across three rounds. Its structure is a usable template, and students can verify every element:
Worked examples visible in that file include a request for replicate counts and mean ± standard deviation reporting; a challenge to whether simulations were dynamical and how they should be parameterized; a request that an empirical correction factor be labeled as empirical; and a request to clarify how model parameters were calibrated. The hardest comment in the file is a necessity challenge—one referee argued the work showed no significant advance and asked why natural seed awns would not serve instead—and the case brief walks through how the authors answered it by conceding what was true, supplying quantities, running the experiment the challenge implied, and adopting the referee’s own counter-example as design rationale.
What a published rebuttal shows and does not show: It documents what the authors wrote and what the referees said next. It does not reveal what the editor privately weighted, and a response that preceded acceptance is not proof that it was the response that produced acceptance.
| Category | When to use it | What it must contain |
|---|---|---|
| Agree and change | The request is right, or the cost of complying is lower than the cost of arguing | The change, its location, and the new text |
| Agree in part | The concern is valid; the proposed remedy is not the only one | What you changed, what you did not, and why the alternative addresses the concern |
| Disagree with evidence | The comment rests on a misreading, or the requested work would not test what it is meant to test | The evidence, the reasoning, and a clarification in the manuscript so the next reader does not misread it the same way |
| Cannot do within scope | The request is legitimate but out of reach | An explicit statement of the limitation in the manuscript, not silence |
Disagreement is a normal part of the genre. A response that concedes every point produces a worse paper and does not, on this record, produce a faster acceptance.
| Time | Activity |
|---|---|
| 8 min | Retrieve Lecture 4 reviews; each team receives the review written about its Draft 2 |
| 18 min | Rebuttal anatomy: read assigned extracts from the published response; classify each as agree / agree in part / disagree with evidence / out of scope, and identify what makes each verifiable |
| 14 min | Draft a point-by-point response to the review the team received, using the quote–response–modification structure |
| 10 min | Reviewer–author exchange: reviewers state whether each response resolves the comment, and name what would |
| 10 min | Tone and evidence test: read the round-three exchange in which a referee challenged a single changed adjective as a good-faith problem and the authors answered with a plain correction; then rewrite one response that concedes a point the evidence does not require conceding, and one that disagrees without giving a reason |
| 7 min | Assemble the response letter and record the contribution note |
| 5 min | Exit ticket: one point on which the team disagreed, and the evidence it cited |
| 8 min | Transition and contingency reserve |
| 80 min | Total |
Output: Peer review written for another team and point-by-point response to the review received, both due October 28 on Blackboard. Both documents are assessed.
Goal: Evaluate AI-mediated feedback on a public scientific text, verify every factual claim in the output, produce the same evidence record whether a student uses a tool directly or audits an instructor-provided output, and complete the portfolio with a final revision sprint, a one-page revision history, and a submission-readiness sign-off.
Journal and committee policies on AI are version-sensitive and differ by role and venue. Check these maintained primary pages directly rather than relying on remembered rules or secondary summaries:
These policies are not interchangeable. Students build a role × input × tool matrix before acting:
| Question | What must be checked |
|---|---|
| Role | Author, reviewer, editor, applicant, or student? A person can move between roles, and the applicable duties change. |
| Input | Public, synthetic, authorized but unpublished, confidential, privileged, personally identifiable, or review-protected? Disclosure after use cannot cure an unauthorized upload. |
| Tool | Has the relevant publisher, sponsor, institution, data owner, and course authorized this tool and its data handling for this input? |
Authors remain responsible for accuracy, attribution, originality, and any required disclosure; that is not a universal permission to use every tool on every manuscript. Reviewers and editors handle confidential or privileged material and may use only systems expressly allowed by the controlling policy. The exact current policy—not a role stereotype—controls.
Module 2 covered proposal-specific permission, confidentiality, and role-specific rules for applicants and reviewers. This lecture covers writing and review: what AI feedback is worth on a manuscript, and what must be verified before any AI-mediated claim enters a document. Module 4 takes up integrity, accountability, authorship, and governance. The course AI Use Policy governs throughout, and sponsor, publisher, institutional, and tool rules outrank course permission.
One common public input. Nothing in this lecture authorizes entering another team’s draft, a team’s own unpublished draft, or any confidential, proprietary, personally identifiable, privileged, or review-protected material into a public AI tool. Every team evaluates the same instructor-supplied publicly available abstract and verifies it against its DOI record and public article. The instructor records the selected source and confirms that course, institutional, and tool rules permit the planned use.
Both paths begin with an independent human review of the same text and end with the same verification record and grading criteria. The second path is not a lesser or substitute assignment.
The attributional question exists because structural and factual competence are different capabilities, and a tool can be useful at one while failing at the other. Where a tool supplies an author list, an institution, a venue, or a year, students check it against the DOI record and count the errors. Where a tool is asked to correct itself, students record whether the correction was right, since a confident correction can be wrong in a new way.
What one class run can establish: that a named tool, at a named version, on a named date, given a specific prompt, produced a specific correct or incorrect output. It cannot establish a general capability ranking between products, a stable error rate, or that a competitor would do better. Report what happened in your run and stop there. Model versions change without notice, and a result from a prior semester is a historical observation, not a current benchmark.
| Time | Activity |
|---|---|
| 6 min | Build the role × input × tool matrix and confirm the public-input boundary |
| 6 min | Write the independent human baseline on the common text |
| 9 min | Direct-use path: produce and save an output; no-direct-use path: inspect the first provenance-complete artifact |
| 11 min | Verify every factual claim against the DOI record and public article; tabulate correct, incorrect, and unverifiable |
| 9 min | Module 3 Microcase 2: “The Reference That Checked Out” and debrief |
| 8 min | Compare the output with the human baseline: supported, unsupported, missed, and newly useful feedback |
| 6 min | Repeat the fixed prompt or compare the second saved run; record variation without generalizing |
| 5 min | Complete the AI-use and verification note |
| 11 min | Final assembly: short revision sprint incorporating the Lecture 5 responses; assemble the one-page revision history (what changed at each draft, who made the change, what prompted it); run the submission-readiness check against the Lecture 4 venue-property record and sign off the contribution note |
| 4 min | Exit ticket: one useful observation from the AI audit and one claim that failed or resisted verification |
| 5 min | Transition and contingency reserve |
| 80 min | Total |
Record the task; whether the student used the tool directly or audited a provided artifact; the tool/model label available; date; exact prompt; public input and DOI; full output identifier; every claim checked and the source against which it was checked; comparison with the human baseline; any material change adopted; human verifier; and unresolved limitation. This is a Module 3 assignment record, not a universal journal, publisher, or institutional disclosure form. Students must separately check the rules of any venue to which they actually submit.
The final assembly uses materials already produced in the module: the response letter from Lecture 5, the venue-property record from Lecture 4, and the verification record from this lecture. Nothing new is drafted from scratch in the last 11 minutes:
Output: Final abstract, one-page revision history, AI verification record, and submission-readiness check, due November 4 on Blackboard.
This module treated authorship only as a record of who did what. Module 4 takes up the questions that record raises: who qualifies as an author, how order and corresponding responsibility are decided, what CRediT roles capture, what accountability means when a paper is wrong, and how AI assistance interacts with all of it. “Contributor,” “author,” “inventor,” and “rightsholder” remain distinct categories.
| Material | Status and Required Boundary |
|---|---|
| Lecture 1 Gopen & Swan and Swales packet | Ready from cited primary sources; teach CARS as description, not requirement |
| Lecture 2 Whitesides article and principles deck | Ready; each principle must name its source and its scope condition |
| Lecture 3 abstract restructuring case | Ready from the public accepted manuscript and a clearly labeled constructed comparison. The comparison is a teaching artifact, not an earlier draft or evidence of the authors’ process |
| Lecture 4 Nature Peer Review File | Ready; public 96-page PDF with a case brief of assigned extracts. Assign extracts rather than the whole file, and check the link before class |
| Lecture 4 eLife comparison | Ready from the pinned November 29, 2024 reviewed preprint (v2) and December 18, 2024 Version of Record (v3); recheck both exact versions before teaching |
| Lecture 4 venue and version check | Ready as a pre-class task against current venue pages; author instructions change frequently and must be rechecked each offering; prepare the Think. Check. Submit. and DOAJ criteria in advance |
| Lecture 5 rebuttal extracts | Ready from the same public file |
| Lecture 6 AI protocol | Ready as a two-path protocol using one public input. Prepare two provenance-complete saved outputs for the no-direct-use path and recheck all governing policies |
| Lecture 6 final assembly | Ready; reuses the Lecture 5 response letter, the Lecture 4 venue-property record, and the Lecture 6 verification record |
| Module 3 microcases | Constructed teaching composites; see the scenario pack |
Before the relevant class, the instructor should:
By the end of this module, students should be able to:
Detailed submission portals and any instructor-supplied packets are available on Blackboard. Public instructions on this page define the course expectations and evidence boundaries.