A preparation checklist can look orderly across two columns and read like two conversations spliced together.
The first heading belongs to document preparation. The next sentence belongs to scheduling. A condition arrives after the action it controls.
Review the sequence that carries the instruction. Identify which words belong together before deciding where the columns should sit.
The method below uses an invented documentation-review page. Its checklist, defects, decisions, and worksheet are synthetic. They describe no real audit or measured outcome.
Identify the Order That Carries Meaning
WCAG 2.2 SC 1.3.2, Meaningful Sequence, is Level A. When presentation order affects meaning, a correct reading sequence must be programmatically determinable. Normative SC 1.3.2.
W3C's informative Understanding guidance allows more than one correct order. Independent sections can appear in different orders if their content remains coherent. Interleaving their sentences can change that result. Understanding Meaningful Sequence.
Start with a content dependency. Does the reader need a condition before an action? Does a label identify the paragraph that follows? Does a numbered step depend on completing an earlier step?
Those questions define what the review must preserve. “The layout changes” is an observation about presentation. “The approval condition follows the submission instruction” names a possible loss of meaning.
For the fictional page, the buyer must obtain approval before sending samples. The page also describes how to choose a review window. Keep each instruction attached to its own condition.
Write the intended meaning in plain text. Ask the content owner to accept it before asking the implementation owner to change the layout.
Separate Reading Order From Focus Travel and Headings
Keep three review records. One describes content sequence. One describes movement between interactive controls. One describes section structure.
For this worksheet, reading order means the sequence available when someone reads the content. Focus travel means the sequence encountered while moving between controls. The two records can involve the same page without answering the same question.
A keyboard check that reaches every link leaves the ordinary preparation paragraphs unexamined. A paragraph-order check leaves the menu's keyboard behavior unexamined.
Use the keyboard-navigation review for the interactive task. Record its outcome separately. Avoid copying a successful keyboard result into a reading-sequence field.
Use the accessible-headings worksheet to inspect section names and ranks. A useful outline can identify the preparation section while its paragraphs still appear in a confusing order.
Enlargement creates another review condition. Use the zoom and reflow guide to record availability and layout at the specified dimensions.
Keep this review focused on the instruction itself. The content owner should be able to read the intended sequence without needing a screenshot or an explanation of the stylesheet.
Inventory the Fictional Two-Column Checklist
The invented page presents two cards. The left card contains preparation instructions. The right card contains scheduling instructions. Each card has a heading, a condition, and an action.
The preparation condition says to obtain approval for sample documents. Its action says to send only the approved set. The scheduling condition says to confirm reviewer availability. Its action says to choose a window after confirmation.
These are fictional service terms. A real reviewer must obtain the actual conditions from the responsible content owner.
The intended sequence is:
| Position | Content unit | Intended text |
|---|---|---|
| One | P1, preparation heading | Prepare the sample documents |
| Two | P2, preparation condition | Obtain approval for the sample set before sending it. |
| Three | P3, preparation action | Send only the approved set. |
| Four | S1, scheduling heading | Choose a review window |
| Five | S2, scheduling condition | Confirm reviewer availability before selecting a window. |
| Six | S3, scheduling action | Choose a window after confirmation. |
The letters identify content units in the worksheet. They are not labels the fictional buyer needs to decode.
Keep a complete copy of each unit. Counting six boxes does not establish that six complete instructions survived a later change. Preserve the condition and the action as distinct records.
Show How a Visual Arrangement Can Hide Interleaving
Suppose the fictional implementation stores the units in this order: P1, S1, P2, S2, P3, S3. It visually positions each unit inside its corresponding column.
The left column appears to read P1, P2, P3. The right appears to read S1, S2, S3. The unstyled text sequence alternates between the two topics.
Read that alternating sequence aloud as an editorial exercise. “Prepare the sample documents. Choose a review window. Obtain approval for the sample set before sending it.” The second heading interrupts the first section's instructions.
This is an invented failure condition, not a report from a screen reader. The example isolates the problem the implementation owner must reproduce in the actual environment.
W3C's informative guidance describes preserving meaningful sequence in alternative presentations. Its multi-column example reads down one column before moving to the next. Understanding Meaningful Sequence.
Do not infer that every multi-column page fails. Compare the actual units, their dependencies, and the order exposed by the implementation. The fictional defect comes from interleaving dependent passages.
Likewise, do not infer failure from a difference between visual placement and source alone. Explain how that difference changes the particular instruction's meaning.
Compare Three Kinds of Evidence
Begin with the visual presentation. Capture the whole checklist and record the viewport, expanded state, page revision, and reading task. Identify which text the visual grouping associates.
Next, inspect the source structure and rendered document structure. Record the actual sequence of the six units. A content template and the rendered document can supply different evidence, so name each inspected version.
Use a text-only or unstyled view as a diagnostic aid. Record the method. Do not treat one extraction as a complete account of every accessibility API or assistive-technology presentation.
Finally, examine the page with an assistive-technology and browser combination suitable for the review's supported environment. Record the product names, full versions, settings, navigation mode, and starting point.
Ask the reviewer to follow the preparation task through its final instruction. Record the words or units encountered, the order, and where a dependency becomes unclear. Keep the observation attached to that environment.
No assistive technology was run for this fictional example. The observation field remains untested. The proposed worksheet asks a real reviewer to supply that evidence rather than treating the invented interleaving as a measured result.
A useful finding compares all three records. It says what the columns suggest, what the inspected structure contains, and what the supported reading attempt exposes.
Copy the Sequence Worksheet
Use one row per content unit. These fields are a proposed editorial method, not a required WCAG form.
| Field | What to record |
|---|---|
| Content unit | Stable identifier, exact words, and page location |
| Dependency | Earlier condition, attached heading, or none identified |
| Intended order | Position within its meaningful sequence |
| Programmatic order | Observed position and inspection method |
| Observed failure | Reproduced meaning change, or untested |
| Review decision | Preserve, investigate, change, or retest, with an owner |
For P3, record P2 as the prerequisite and P1 as its section heading. Its intended local position is third. In the invented interleaved structure, its document position is fifth.
For S3, record S2 as the prerequisite and S1 as its section heading. Its intended local position is third within scheduling. Its position in the combined intended sequence is sixth.
Distinguish local order from whole-page position. “Third” alone is ambiguous when the page contains two instructions with three units each.
Record a programmatic inspection separately from the assistive-technology result. In this walkthrough, the structural mismatch is specified by the example. The actual speech observation remains untested.
Assign the implementation question to an owner. Assign disputed wording or dependencies to the content owner. An unresolved instruction cannot become approved merely because someone rearranged its containers.
Allow Independent Regions to Change Places
The preparation card and scheduling card can have a different overall order if their conditions remain understandable. In this fictional service, scheduling does not require the sample set first.
That assumption matters. If the real service requires accepted samples before scheduling, the content owner must record that dependency. The intended combined sequence would then need to reflect it.
Without that dependency, S1, S2, S3 followed by P1, P2, P3 is another coherent proposed sequence. Each instruction remains intact.
W3C explicitly distinguishes meaningful order within content from flexible order between independent regions. Its guidance does not require a single universal arrangement for every page section. Understanding Meaningful Sequence.
Write the permitted variation into the acceptance record. Otherwise, a reviewer might reject a coherent narrow layout because it begins with the other independent card.
Keep qualifications close to the action they control. If an editor proposes moving a condition into a distant sidebar, ask how the revised sequence preserves the connection. Leave that decision open until the reading evidence answers it.
Retest the Complete Instructions After a Change
Suppose the implementation owner groups P1, P2, P3 together and S1, S2, S3 together in a revised page. Preserve the original evidence before retesting.
Compare all six units with the approved inventory. Check the words, dependencies, order, and section boundaries. A change that removes the approval condition cannot pass this review by producing a cleaner sequence.
Repeat the visual, structure, and supported assistive-technology observations in their recorded environments. Record the revised page identifier and any changed configuration.
Use a bounded after note. It should identify the preparation sequence, the scheduling sequence, and the specific observations now available. Leave untested components and environments visible.
Repeat the separate keyboard and reflow checks when their affected conditions require it. Keep their results in their own records. A repaired instruction order does not establish that every control or narrow layout works.
Close a finding only when its evidence addresses the original meaning change. Keep any missing assistive-technology observation marked as missing.
Keep each instruction together.
Sources
- W3C: WCAG 2.2 SC 1.3.2. Normative Level A requirement. Checked 2026.10.28.
- W3C: Understanding Meaningful Sequence. Informative guidance on dependent sequences, independent regions, and multi-column content. Checked 2026.10.28.