DOCX or PDF Before Submission: What to Check and What Can Break During Conversion

Before choosing DOCX or PDF, identify what your university, faculty, department, journal, supervisor and submission system currently require. Requirements vary, and one workflow may ask for an editable DOCX during supervision, another for a final PDF, and another for both. A portal may accept a source file and convert it automatically. Signed documents, archival copies or accessible submissions may require a particular kind of PDF rather than any file with a .pdf extension.

Check each deliverable separately. The main thesis, appendices, datasets, questionnaires, media and supporting material may have different format or naming rules. Confirm whether the instructions specify page size, file-size limits, embedded fonts, bookmarks, accessibility tags, signatures or a prescribed archival profile. If the portal generates a preview or converted copy, determine which version becomes the official submission.

A DOCX may retain a rich editable document model: heading styles and outline levels, automatic table-of-contents fields, footnote and endnote anchors, cross-references, section breaks, page-number settings, headers and footers, captions, tracked changes, comments, citation fields, hyperlinks, language metadata and document properties. Tables and equations can also remain editable. This structure can make a DOCX useful for technical analysis because a heading can be recognised as a heading rather than inferred from its appearance, and a note can remain linked to its anchor.

That richer model also helps correction. You can update a field, repair a heading level, revise a caption or change a section setting at its source. However, no DOCX exposes every element reliably in every application. Features created by add-ins, unusual equation editors, embedded objects or application-specific fields may be interpreted differently elsewhere.

• Layout can change between Word versions, operating systems and alternative office applications.
• Missing fonts may alter line wrapping, symbols and pagination; linked images or other assets may not travel with the file.
• Field values can be stale, while comments, tracked changes or private notes can remain unexpectedly.
• The recipient can edit the document, and its visible layout is not completely frozen.

A well-produced PDF generally preserves the visible result more predictably: page dimensions, line and page breaks, placement of text and images, final pagination, visible headers and footers, and the position of tables and figures. This is why you must inspect the exact PDF intended for submission. Reviewing the DOCX alone does not prove that the export contains the same pages or that every object still fits.

PDFs are not all equivalent. Headings may cease to be semantically tagged; footnote relationships may become merely visual; and extracted text can follow an order different from the page a reader sees. A scanned PDF may have no usable text layer. Hyperlinks or bookmarks may be missing, fonts may be substituted or poorly embedded, and accessibility tags may be incomplete. Equations and special characters can become images or malformed extracted text.

The distinction is therefore practical, not a contest in which one format always wins. DOCX often carries more source structure for analysis and correction. PDF is often better for checking the final visual state. Your required submission format and the qualities of the actual files determine which checks matter.

Conversion failures depend on the software, installed fonts, export method and complexity of the document. Even when most pages look correct, a small change in line wrapping can shift later page breaks, strand a heading at the foot of a page, change table-of-contents page numbers or create an unexpected blank page. Section numbering can restart, and headers or footers can inherit the wrong content.

• Fonts may be substituted; bold or italic styling, diacritics and special characters may disappear or change.
• Equations may be damaged, tables clipped, images softened, transparency altered, captions separated or footnotes moved.
• Internal links and bookmarks may be lost, while long DOI or URL strings may overflow their column or page.
• Tracked changes or comments may become visible even though they were easy to overlook in the working view.
• Appendices, selected page ranges or final pages may be omitted through an incorrect export setting.

Also look for missing captions, displaced figures and incorrect page-number transitions. The broader guide to common thesis formatting mistakes explains how conversion problems interact with heading structure, numbering and manual formatting.

Save a separate submission copy so that routine edits do not get mixed with the version being prepared for hand-in. Review tracked changes and accept or reject them as appropriate; remove resolved comments and private notes. Do not delete material that your institution requires, such as declarations, accessibility descriptions or review evidence. Then update the entire table of contents, cross-references and citation fields rather than updating only the visible page numbers.

• ✓Check heading hierarchy, section breaks, page-number settings, headers and footers.
• ✓Review footnotes and endnotes at their anchors and at section or page boundaries.
• ✓Confirm table, image, caption and equation placement, including wide or landscape pages.
• ✓Ensure the fonts in use are available and suitable for the required output.
• ✓Remove accidental blank paragraphs, repeated spaces and manual spacing used only to push content.
• ✓Verify document language settings, diacritics and special characters.

Save, close and reopen the final DOCX. This simple step can reveal a broken link, an unsaved field update or layout that changes after the application recalculates the document. Recheck the first and last page after reopening before you export.

Use the application’s proper Export or Save as PDF function where possible. It is usually safer than printing through an unrelated virtual printer, copying the document into a different application, sending unpublished work to an unknown online converter, taking screenshots or rebuilding pages manually. This is not a claim that one application always creates a correct PDF; it is a way to reduce uncontrolled transformations.

1. Record or consistently use the application and version, the export command and the intended page range.
2. Confirm page size, orientation and any landscape sections before completing the export.
3. Choose bookmark and accessibility or tagging options where they are relevant or required.
4. Review image-quality or compression settings and font embedding where the software supports it.
5. Give the output an unambiguous final filename and do not overwrite it with an experimental conversion.

If the submission portal performs its own conversion, upload only after preparing the source carefully and then inspect the portal’s preview or downloadable result. Your locally exported PDF cannot prove that a separate server-side conversion produced the same output.

Random scrolling is not enough. First compare the total page count, first and last page, title page and table of contents. Then inspect the beginning and end of every major chapter, all page-number transitions and every section boundary. These checkpoints reveal many cascading changes without pretending that a sample replaces a complete review.

• ✓Compare pages containing wide tables, figures, equations or many footnotes.
• ✓Review the bibliography, appendices, final declarations and any unusually formatted pages.
• ✓Compare headers, footers, diacritics, mathematical symbols and other special characters.
• ✓Test representative internal links, external links and bookmarks where they are required.
• ✓Check that chapter titles and page numbers agree with the table of contents.

After checking representative high-risk pages side by side, read or visually review the final PDF from beginning to end. Use the citation, footnote and bibliography workflow for the source details that need a separate two-way check. Finally reopen the exact file selected for upload; do not assume a similarly named copy is identical.

Aggressive compression can blur charts, damage small text inside images, make screenshots unreadable, introduce visual artefacts or remove metadata and bookmarks. An unnecessarily large file can also exceed a portal limit, time out during upload or be difficult to open on a slower connection. Aim for a file that satisfies the stated limit while keeping required content legible; do not optimise only for the smallest number of megabytes.

Treat online conversion services as a data-handling decision. A thesis may contain confidential interviews, unpublished research, personal data, restricted appendices or commercially sensitive material. Before uploading, understand who operates the service, where the file is processed, how long it is retained and whether deletion controls are available. University or employer rules may restrict third-party processing.

ThesisReady can help surface technical, formatting, structural and citation-related signals in DOCX and PDF files. DOCX usually preserves richer structure for analysis, while PDF remains useful for reviewing the final submitted appearance. Extraction quality varies: a PDF with a weak or missing text layer may produce limited analysis, and visual conversion defects can still require direct human inspection. A finding is a prompt for review, not certification. The thesis format checker provides a concise overview of the product workflow.

ThesisReady does not check plagiarism or originality. It does not determine whether work is original. It does not certify institutional compliance, guarantee that conversion preserved every element, judge whether the academic argument is correct or certify citation-style compliance. It does not replace your supervisor, university review, inspection of the exact final submitted file or the current requirements for your programme.

ThesisReady is not CRZP, an official university system or a guaranteed compliance service. Human review remains necessary for academic meaning, confidential material, the suitability of sources, accessibility requirements and any rule that depends on institutional interpretation.

• ✓The current institutional format requirement is confirmed for every deliverable.
• ✓The final DOCX copy is isolated from working drafts.
• ✓Comments and tracked changes have been reviewed appropriately.
• ✓Fields, cross-references and the complete table of contents are updated.
• ✓Headings, section breaks, page numbering, headers and footers are checked.
• ✓Footnotes and endnotes are checked at their anchors and destinations.
• ✓Tables, figures, captions and equations are checked.
• ✓The PDF was exported using the intended controlled method.
• ✓PDF page count, title page and table of contents are checked.
• ✓Chapter boundaries, bibliography, appendices and final pages are checked.
• ✓Fonts, diacritics, symbols, links and required bookmarks are checked.
• ✓File size and the readability of images and screenshots are checked.
• ✓The exact final uploaded file has been reopened.
• ✓The upload preview or converted copy has been reviewed where available.
• ✓The correct final version is stored locally with a clear filename.
• ✓A human has reviewed institutional and academic questions.

Complete the checklist against the files you will actually submit, not an earlier source and not a preview from memory. If a final correction changes the DOCX, export again and repeat the affected comparison plus the complete final PDF review.