# Documents this run could not read, and where they are Written on Ben's ruling of 2026-09-16 on QUESTIONS.md Q5: > "Not admissible for the agent. Do not read figures off a rendering. […] Also write > RENDERED_ONLY.md listing every such document with URL and page, so Ben can read them > himself; any figure he enters later carries 'read by Ben from rendering, [date]', never > presented as byte-verified." Every document below is published, reachable, and returns HTTP 200. None of them carries a figure this run could byte-compare, so **no figure from any of them appears in the CSV or in any source log** — including in prose, including in a footnote, including as a range or an approximation. **If you enter a figure from one of these**, the cell it goes in carries `read by Ben from rendering, [date]` and the row's `source_form` stays as it is. A figure read by eye and a figure compared against the bytes the server sent are different kinds of thing and the dataset has to keep them apart; nothing read from a rendering is ever presented as byte-verified. Covers schools 1–13 of 19. Appended to as later schools are read. --- ## 1. Cleveland University-Kansas City — the cost of attendance **This is the one that blocks fields.** `f1_state` and `f2_state` are `unverified` for Cleveland solely because of this document, with `source_form = image-only`. | | | |---|---| | document | College Financing Plan (the U.S. Department of Education consumer template) | | URL | | | linked from | | | pages | **1 and 2 — both** | | why unreadable | no text layer at all. 0 text objects and 0 embedded fonts on each page; 2,401 and 4,637 vector drawing operations. Every character is a filled outline. Metadata: producer `Microsoft: Print To PDF`, title `print.ss.pdf`. | **What a reading would answer.** Field 1: whether any row of the estimated-cost section names a board or licensing examination, and if so its label and figure. Field 2: the two columns are split by residence type, so the figures that differ between them are field 2's answer. **What is known about the layout without reading any figure**, from the rendering and recorded here only as structure: a section headed in the position where `Estimated Cost of Attendance` appears; a two-column split by residence type; rows for tuition and fees, housing and food, books and supplies, transportation, and other education costs; then Student Aid Index, gift aid, VA benefits, net price, loan and work options; a glossary on page 2. **What Cleveland publishes about the document, as text, and what field 6 rests on:** > The College Finance Plan is a calculation of two terms and does not reflect the full cost of > a program. It is a consumer tool developed by the U.S. Department of Education and is > designed to simplify information about costs and financial aid for prospective students. So whatever the figures are, they are two terms and not a program. --- ## 2. Cleveland University-Kansas City — the NBCE pass rates Blocks nothing. Field 5 asks for the route and the route is published; `f5_state = published`, `f5_source_form = image-only`. Listed so the destination is not mistaken for readable if anyone returns to it. | | | |---|---| | document | `NBCE Exam Completion Rates` | | URL | | | linked from | | | page | 1 | | why unreadable | the pass-rate table is a pasted raster image, 1047 × 389, `FlateDecode`. The page's text layer holds 26 characters, all whitespace. | A second document on the same page is a bare image file: | | | |---|---| | link text | `2021-2024 NBCE Licensing Success Rate By College` | | URL | | --- ## 3. Logan University — the NBCE pass rates and degree completion rates Blocks nothing. `f5_state = published`, `f5_source_form = image-only`. Page: — it contains **no `` element**. Both sets of rates are raster images on Logan's own CDN with no text alternative. | under the heading | image | size | attributes | |---|---|---|---| | NBCE licensing exam success | `NBCE_Exam_Success_2022-2025.png` | 720 × 203 | `alt=""`, `title=""` | | DC degree completion | `DC_Degree_Complete_Rate_2022-2025.png` | 556 × 325 | `alt=""` | Full `src`, as the bytes hold it — the `&` is published as `&`: ``` https://i0.wp.com/www.logan.edu/wp-content/uploads/2026/07/NBCE_Exam_Success_2022-2025.png?fit=720%2C203&ssl=1 https://i0.wp.com/www.logan.edu/wp-content/uploads/2026/07/DC_Degree_Complete_Rate_2022-2025.png?fit=556%2C325&ssl=1 ``` The same page publishes a goal as text, and that is quoted in the row: > Goal: The College of Chiropractic will achieve 80% or higher for the four (4) most recent > years' NBCE licensing exam rates in accordance with CCE Policy 56. --- ## 4. Life University — the NBCE pass rates Blocks nothing. `f5_state = published`, `f5_source_form = external-embed`. Page: — it contains **no `
` element anywhere in the chiropractic section and no rate in the page's own bytes**. The rates exist only inside a Power BI view on a third-party host. Five Power BI embeds sit on the page. The chiropractic licensing one is distinguished by its `title`, which occurs once: ``` ``` The other four, in document order, are `DC_Retention`, `DC_Completion21Qtr`, `UG_Retention` and `UG_Graduation`. The last two are undergraduate and are excluded by the brief; nothing from them is recorded. --- ## 5. Not rendered-only, but not readable where the college points either These do not belong in the list above — the documents behind them carry real text layers and could be read. They are here because **the URL the college publishes does not return them**, so anyone retracing the path finds a wrapper rather than a document and may conclude the figures are unreadable when they are not. **Life Chiropractic College West.** Both the academic catalog and the Policy 56 report are published as Google Drive **viewer** URLs. A viewer URL returns an HTML page — 86,218 and 86,007 bytes — whose `` is the file name and which contains no document. | what | URL Life West publishes | what it returns | |---|---|---| | Policy 56 report | <https://drive.google.com/file/d/1DVBTnOw3V7cIOsEpjzQSLioyhpjXCTrR/view> | wrapper, 86,218 bytes | | Time to Complete | <https://drive.google.com/file/d/1lEj3MpgsG1pEWWpjtv2JXqhVHuVOyGG3/view> | wrapper, 86,007 bytes | | academic catalog | <https://drive.google.com/file/d/1wLsEV4WlK96KJ_9UXjmt8G4mr-BvdTps/view?usp=sharing> | wrapper | The documents were reached instead at `drive.google.com/uc?export=download&id=…` with the same file ids. **That is a URL we constructed and not one Life West publishes**, and the CSV source cells say so rather than showing it as though Life West had linked it. The catalog was read this way and fields 1, 3 and 4 rest on it; its text layer is sound. The Policy 56 report and Time to Complete were fetched, confirmed to carry text — 2 pages, 104 and 121 words, producer `Skia/PDF m128 Google Docs Renderer`; and 1 page, 103 words, producer `www.ilovepdf.com` — and **not read**, because the brief asks for the route to the rates and not the rates. `f5_source_form` is `external-embed` for Life West for this reason: at the URL the college points to, the figures are not there. --- ## 6. Northwestern Health Sciences University — Student Consumer Information Blocks nothing. Every field in NWHSU's row is answered from documents that carry a text layer, and none of them is this one. Listed because it sits in the homepage's About Us menu directly beside the Fact Sheet the pass-rate route ends at, so it is the document most likely to be opened next by anyone retracing that route. | | | |---|---| | document | `Student Consumer Information` | | URL | <https://www.nwhealth.edu/wp-content/uploads/NWHSU-Student-Consumer-Information-v7.pdf> | | linked from | <https://www.nwhealth.edu/> — About Us menu, immediately after `Student Body and Outcomes Fact Sheet` | | pages | **1, 2, 3 and 4 — all of them** | | why unreadable | no text layer on any page: 0 characters each, against 72, 69, 49 and 44 embedded images per page. Metadata says creator and producer `Microsoft® Word for Microsoft 365`, so the document began as text and was flattened to images somewhere after export. The metadata also names an individual as author; recorded as existing and not reproduced. | **What a reading might answer.** Unknown — the title is a federal consumer-disclosure category, which at other colleges in this set has carried cost-of-attendance figures, board pass rates, or neither. Nothing is assumed about its contents from its filename. **Nothing in NWHSU's row rests on it.** Field 1 rests on the DC tuition page, field 4 on the catalog's Financial Policies page, and field 5 on the Fact Sheet PDF, whose three pages all carry text. --- ## 7. Parker University — the chiropractic National Board exam success rate Blocks nothing in the CSV — `f5_state` is `published` because the route is published — but it sets `f5_source_form = image-only` and is the subject of **Q9** in `QUESTIONS.md`. | | | |---|---| | document | the graphic under the heading `CHIROPRACTIC NATIONAL BOARD EXAM SUCCESS RATE` | | URL | <https://www.parker.edu/wp-content/uploads/2026/07/f7238152-0842-4b15-ac76-b89f7b0ab21c.png> | | displayed at | 600 × 269, via `...-600x269.png` | | linked from | <https://www.parker.edu/national-board-scores-completion-rates/> | | why unreadable | it is a PNG. There is no text layer to read and none to recover. The page carries **no `<table>` element at all**, so nothing on it restates the figures as markup. | **Also image-only on the same page**, listed so the pair is not mistaken for one item: the graphic under `CHIROPRACTIC PROGRAM COMPLETION RATES` (<https://www.parker.edu/wp-content/uploads/2026/07/image-600x345.png>, displayed 600 × 345). Completion rates are not this brief's subject, but the same reader meets both. **A readable alternative exists at the same destination** and is why Q9 is open: the `STUDENT ACHIEVEMENT` PDF (<https://www.parker.edu/wp-content/uploads/2026/03/PUBLIC-DISCLOSURE-STATEMENT-Student-Achievement-Goals-2025-v2.pdf>) carries a text layer with an `NBCE` row. **No figure from either the images or the PDF appears in the CSV** — field 5 asks for the route, not the rates.