Decide what belongs in an appendix
Sort supporting material with a worked disposition ledger, precise appendix references and a check that your report's main argument still stands alone.
About 11 min read · Allow extra time for the practical checks
An appendix gives a reader somewhere to check relevant supporting detail without interrupting the report’s main argument. It should not become the place where the argument finally appears.
The difficult decision is often not how to label Appendix A. It is whether a particular item belongs there at all. A long table might provide useful verification; a short limitation might be essential to the interpretation and need to stay in the body. Length alone does not decide.
This guide uses an original fictional report pack to show four dispositions: keep in the body, summarise in the body and append permitted detail, omit, or hold until a question is resolved. A material ledger and a reverse dependency check make those decisions traceable. These are practical editorial recommendations, not a universal assessment format.
Establish what your actual submission allows
Read the brief, handbook or journal instructions before moving material. Record whether appendices are permitted or required, what counts towards the word limit, which file formats are accepted, and any limits on length or supplementary files. Do not assume an appendix is excluded from the word count or that an assessor will read it in full.
If the instructions are unclear, ask a specific question: “May the question wording appear in an appendix, and does that text count towards the limit?” Until answered, keep the uncertainty visible in your plan. Do not redesign the submission around an imagined exemption.
Formatting conventions also depend on the required style. Purdue OWL’s appendix guidance for APA papers explains APA labels, titles and references to supplementary material. Use its detailed formatting rules when APA governs your work, not as rules for every report.
First settle which report structure the brief requires. An appendix is a supporting destination within that structure, not a substitute for deciding what the main sections must accomplish.
Separate essential content from useful verification
Ask what the reader would lose if the item were unavailable.
If they would lose the answer to the question, the evidence needed to understand that answer, a necessary reasoning step or a consequential limitation, retain that content in the body. You may still put additional detail elsewhere, but the body must carry its meaning.
If the reader could understand and assess the main argument but might reasonably want to inspect exact wording, extended workings or a detailed record, an appendix may help. MIT’s Mayfield Handbook frames appendices as useful reference material rather than essential main-text content. That distinction is more useful than a rule such as “anything longer than a page goes at the end.”
For example, a methods section may need to identify what a questionnaire measured, how it was used and its relevant limits. The exact permitted questionnaire wording can then support that explanation in an appendix. Moving the entire explanation out of the methods section would leave the reader unable to understand the study without detouring.
An appendix is also not automatically a safe home for confidential or third-party content. Relevance, disclosure permission and assessment requirements are separate decisions. A requirement to demonstrate a process does not by itself establish permission to publish identifiable records.
Build a material-disposition ledger
For each candidate item, record:
- its exact identity and working-file location;
- the body claim or section it supports;
- the reader’s reason for consulting it;
- the proposed disposition and a short reason;
- the intended appendix label and specific locator, if applicable;
- any unresolved requirement, permission or confidentiality question.
The ledger is a working aid. It need not become another submitted appendix unless the task calls for it. Its purpose is to stop “might be useful” from turning into an indiscriminate collection of files.
Use four decisions consistently. Body means the content is needed to understand the report. Body summary plus appendix means the argument stays visible while permitted detail supports checking. Omit means the item does not earn a place in this submission. Hold means a relevant uncertainty prevents a final inclusion decision; it does not mean “include anyway and add a disclaimer.”
Work through a fictional report pack
Wholly fictional teaching case: Rowanbridge Study Centre trialled a staffed equipment-return desk for four weeks. A learner is writing a report on whether the available records justify continuing the trial. The centre, records and figures below are invented; they are not evidence from a real organisation.
The fictional brief permits relevant appendices and requires the main report to explain its evidence and limitations. Its word-count treatment is unspecified, so that question remains open. Nothing in the brief authorises disclosure of personal records.
The pack contains a checked weekly summary of 12, 15, 11 and 18 completed returns; a 56-row transaction record with names and email addresses; a short question sheet created by the fictional report author; twelve response sheets, nine of which favour continuing; an extended staffing-cost worksheet with a checked total of $320; an unrelated events brochure; and a note that no pre-trial comparison is available. Completed returns count transactions, not distinct people. The twelve responses are not known to represent all users, and the cost worksheet does not establish full operating cost or affordability.
The following ledger separates the usefulness of a detail from whether it can be shared. The appendix labels are local teaching choices, not a prescribed numbering system.
On a narrow screen, scroll the table sideways to read every column.
| Item | Body location and reader need | Decision and reason | Locator or unresolved gate |
|---|---|---|---|
| M1 · Checked weekly return summary | Findings §3.1: understand activity and its counting unit. | Body. The four weekly counts and transaction definition are part of the evidence being assessed. | Table 1 in the body; no appendix copy needed. |
| M2 · Named transaction record | Methods §2.1: document where activity counts came from. | Hold the raw record. Identifying details are not justified by a desire to show all the evidence. | No public appendix assigned. Confirm authorised disclosure and verification arrangements with the responsible person. |
| M3 · Author-created question sheet | Methods §2.2: let readers inspect what respondents were asked. | Describe the response task in the body; append the blank wording because it helps readers inspect its scope without reproducing responses. | Appendix A, “Trial feedback questions,” questions 1–2. No third-party or personal content in this fictional blank sheet. |
| M4 · Twelve response sheets | Findings §3.2: assess the narrow feedback claim. | Keep the aggregate result and representativeness limit in the body. Hold the sheets pending disclosure review; do not assume removing names is sufficient. | No appendix assigned. Whether any excerpts may be shared remains unresolved. |
| M5 · Staffing-cost worksheet | Findings §3.3: inspect the limited cost total and its composition. | Body summary plus appendix. Retain $320 and the excluded-cost limitation in the body; place permitted extended workings in supporting detail. | Appendix B, “Recorded staffing-cost workings,” section B.1, total row. In the fictional specification, this worksheet has no personal or restricted content. |
| M6 · Events brochure | No claim or analytical step depends on it. | Omit. It adds background about the centre but does not help answer the report question. | None. A file's availability is not a reason to submit it. |
| M7 · Missing-comparator note | Discussion §4.1: assess what the evidence can establish. | Body. Absence of a pre-trial comparison limits the interpretation; it must not be hidden in an appendix. | Explicit limitation in §4.1 and carried into the conclusion. |
M2 and M4 demonstrate why “append the raw data” is not an automatic answer. The supplied checked summaries can be discussed with their limits, but this exercise does not invent permission to circulate the underlying records. In a real project, restricted access may also limit what an assessor or reviewer can verify. Resolve that through the appropriate institutional or publisher process; do not fabricate a shareable replacement or claim that withholding has already been approved.
M6 shows a different decision. It is omitted because it has no useful role, not because it is sensitive. Keeping those reasons distinct makes revision easier if the report’s question changes.
Write the body first, then point to supporting detail
Compare these two fictional passages:
Weak: The trial was successful. All results and limitations are in the appendices.
Revised: The four-week record contains 56 completed returns, distributed as 12, 15, 11 and 18 by week (Table 1). These are transactions rather than distinct users. Nine of twelve respondents favoured continuing the trial; the response task asked about continuation, not learning outcomes (Appendix A, questions 1–2). Recorded staffing expenditure was $320; the extended workings appear in Appendix B, section B.1. This is not a full operating-cost estimate. Without a pre-trial comparison and evidence that respondents represent users more broadly, these records do not establish an improvement caused by the trial.
The revised passage contains the findings and the limits needed to interpret them. Its appendix references offer inspection rather than postponing the argument. It does not establish a final continuation decision: the report still needs a reasoned assessment against the brief’s criteria and any evidence gaps.
The arithmetic is deliberately simple: the four weekly values sum to 56, and nine favourable responses belong to twelve respondents, not to all 56 transactions. An appendix label cannot repair a mismatch in those units.
The proposed Appendix A would preserve the fictional author’s exact question wording: “1. Do you favour continuing the staffed return-desk trial? 2. What is your main reason?” It would not contain participant answers. Appendix B would carry the checked worksheet, including its $320 total and stated cost exclusions. This guide does not invent the individual worksheet entries or present these short descriptions as a complete submitted report.
If Appendix B’s total later changes, update the body and interpretation too. A correct supporting file and an obsolete main-text number still produce an inconsistent submission.
Make every appendix reference earn its place
Use an informative title and a stable identifier. “See supporting information” asks the reader to search; “Appendix B, section B.1, recorded staffing-cost total” names a destination and tells them why it matters.
TU Delft’s appendix guidance connects extended workings or code with the main report through relevant text references and organised topics. Apply that principle without treating its course-specific formatting as your own submission instructions.
Check the relationship in both directions:
On a narrow screen, scroll the table sideways to read every column.
| From the body | Destination to verify | Return check |
|---|---|---|
| Methods §2.2 and Findings §3.2: what the feedback task asked | Appendix A, Trial feedback questions, questions 1–2. | The blank wording supports the continuation-response interpretation, not a claim about learning outcomes. |
| Findings §3.3: $320 recorded staffing expenditure | Appendix B, Recorded staffing-cost workings, §B.1, total row. | The total, units and exclusions agree with §3.3; neither location calls this full cost or affordability. |
This is a planned audit for the fictional pack, not certification that a real attachment was checked. Perform it on your actual final files. After a reordering or deletion, check every affected label rather than relying on “above,” “below” or an old page number.
Tables and figures inside appendices still need intelligible titles, notes and source credit under the required style. The caption-decoding guide helps check that separate task. Supporting material is not exempt from citation, accurate labelling or accessible presentation.
Run the reverse dependency check
Hide the appendices temporarily and read the body in submission order. Can you still identify the question, approach, principal evidence, reasoning, answer and consequential limitations? Highlight every sentence that requires an unseen attachment merely to make sense.
Then reopen the appendices. For every item, name the precise body claim or method it supports and the extra inspection it enables. If there is no useful return path, reconsider inclusion. If its only job is to supply the missing argument, move the essential content back into the body.
In the fictional pack, hiding A and B removes exact question wording and extended cost workings, but not the response task, result, cost total or interpretation limits. The body remains understandable. The reader can return to A or B for more detail; confidential records are still held, not silently counted as disclosed evidence.
Before submission, check:
- The actual brief permits the planned appendices and any word-count uncertainty is resolved.
- The body carries essential evidence, interpretation and limitations.
- Every appendix has a relevant job and an informative title.
- Every body callout reaches the intended final item, section, question or table.
- Supporting totals, labels and versions agree with the body.
- Confidentiality, attribution and reuse requirements have been addressed through the appropriate process.
- No unresolved “hold” item slipped into the submission.
- The exported files preserve readable content, navigation and any required text alternatives.
If moving material was mainly an attempt to escape a word limit, return to structural word-count revision instead. Removing duplication and narrowing scope are different from hiding essential reasoning.
Where your course or publisher permits external support, academic editing can help locate broken references, duplicated detail and missing explanations in your own draft. The author remains responsible for the evidence, permitted disclosure and final decisions about the submission. An editor should query an unclear evidential relationship, not invent one.
