A large total needs a time frame
An open textbook can serve more than one class and more than one academic year. That makes a growing programme total plausible, but it also makes the number easy to misread. A cumulative estimate can include several years of an earlier grant cohort, while a new cohort has only just entered the calculation. Comparing either number with one year’s grant budget would answer a different question again.
This reference reconstructs one published table from the Colorado Department of Higher Education’s 2025 Open Educational Resources Report. A cohort is the group of grants in one programme year. It follows the six grant cohorts shown for 2019–2020 through 2024–2025 and asks a precise question: how much of each change in the displayed total comes from changes in earlier cohorts, and how much comes from the cohort newly appearing in that row? It does not estimate the programme’s causal effect.
Three results, with their denominators
All figures below come from the same 2025 report edition. Dollars are US dollars as displayed in the source; percentages are our calculations from the integer cells. These are shares of reported monetary estimates, not shares of students or courses.
- The latest displayed total increases from $42,167,701 to $59,791,983: a difference of $17,624,282. Of this difference, $13,473,635, or 76.45%, comes from changes in cohorts 1–5; $4,150,647, or 23.55%, is the first displayed amount for cohort 6.
- In the final cumulative row itself, cohorts 1–5 total $55,641,336, or 93.06% of $59,791,983. Cohort 6 supplies the remaining 6.94%. The 93.06% cumulative share and 76.45% change share have different denominators and should not be substituted for each other.
- Across the five transitions that can be reconstructed, earlier cohorts supply between 61.94% and 86.09% of the change in the reported total. The pattern is not a monotonic growth rate, a measure of adoption retention, or an estimate of students’ actual spending.
What the report’s estimates mean
The source calculates textbook-cost estimates from course enrollment multiplied by the previous textbook’s cost and assumes continued use of the open materials. Its Figure 1 expressly describes compounded estimates under that assumption. Table 1 is captioned as annual savings, but its first cohort’s cells increase from $3,925,488 in the first row to six times that amount in the last row. The final row’s sum also corresponds to the report’s cumulative headline.
We therefore describe the rows as reported continued-use estimates and treat their sums as a cumulative accounting series. That interpretation is stated rather than silently replacing the source’s caption. The calculation does not establish that every earlier course was offered again, that the same enrollment returned, or that the assumed purchase would otherwise have occurred. The report’s model supplies those estimates; this study decomposes their displayed changes.
The latest change, one cohort at a time
Start with the 2023–2024 row total of $42,167,701. Compare each of the five existing cohort cells with its counterpart in the next row. Those five differences sum to $13,473,635. Then add the new sixth cohort’s first displayed amount, $4,150,647. The result is $59,791,983, exactly the sum of the six cells in the final row.
This identity is useful because it keeps two sources of a rising reported total apart. The earlier-cohort portion comes from changes in estimates already represented in the table. The new-cohort portion is a newly introduced component. Neither portion should be called a count of continuing courses, a count of newly reached people, or a verified annual cash benefit. The decomposition follows the monetary cells, not individual learners.
The same bridge across five transitions
The calculation is repeated for every transition after the first displayed row. Earlier cohorts account for 61.94% of the change into 2020–2021, 63.65% into 2021–2022, 86.09% into 2022–2023, 82.24% into 2023–2024 and 76.45% into 2024–2025. The downloadable table retains the dollar components beside those shares so that a small newly introduced component is not confused with a small programme overall.
There is no transition calculation for the first row. The report does not supply a preceding row in this table, so we show the initial displayed total without inventing a measured zero baseline. Cohorts not yet introduced are structural blanks. They are not missing survey responses, failed projects or reported zero outcomes. The method sums only cohorts actually present in each comparison.
Preserve the source cells, including irregular ones
The source table reports whole-dollar amounts. We keep those amounts unchanged, perform arithmetic in integer dollars and round derived shares to two decimal places. That display precision is not an uncertainty interval. It does not quantify error in enrollment, textbook-price assumptions or continued use, none of which can be recovered from these cells alone.
We do not regularize cells into a smooth pattern. For example, the cohort 2 amount in 2022–2023 is $7,235,075, ten dollars below three times that cohort’s first amount. Cohort 3 also has changing increments. These features remain in the reconstruction without an invented explanation or an allegation of an error. The narrative’s approximate $59.7 million headline is not used to rescale the table-derived sum of $59,791,983.
Keep programme cohorts, institutions and enrollments apart
The report contains other useful measures that cannot simply be joined to this bridge. Its separate institutional table presents the report’s estimated annual textbook-cost savings for 2024–2025 across all zero-textbook-cost courses, including courses outside the grant programme and courses using library resources. It is a different reporting universe from the six-cohort continued-use series.
The enrollment table also warns that its counts are duplicated: one person may take several courses using open resources. We do not divide the cohort estimates by those counts to produce a per-student benefit. Such a ratio would require a verified common population, time period and purchase counterfactual. Nor do we calculate a programme return on investment from one year’s grants and a multi-year estimate; institutional work and other costs would need their own accounting.
What this can—and cannot—tell a student about a bill
For a student, the practical question is the charge attached to the chosen course and the materials actually needed. A statewide estimate is not an individual invoice. An available open textbook can remove the need to buy that text, while optional print copies, other required materials or an institution’s billing arrangements may still matter. The programme figures do not identify those choices for a particular student.
The report itself discusses flat-fee material programmes as a reason to distinguish open content from the final charge. We use that historical discussion to explain the distinction, not as a statement about every institution’s current policy. To assess an actual bill, check the institution’s current course-material listing, the required items and the applicable billing or opt-out rules. Nothing in this accounting reference predicts the amount any reader will pay.
What is new here, and what was already public
Colorado had already published the cohort table, a cumulative chart, cost estimates and descriptions of its grant programme. Continuing use was already part of the official explanation. We do not present that idea, the reported total or the source’s programme achievements as discoveries of this study.
The addition is an exact, reproducible bridge for all five transitions, a direct contrast between change shares and cumulative shares, and a traceable long-format reconstruction of the 21 source cells. Readers can inspect every addition and subtraction instead of inferring the size of a contribution from stacked areas. The method also records why differing table universes, structural blanks and source irregularities cannot be silently simplified.
How to reproduce the accounting
The offline recipe takes the original PDF downloaded separately from the Colorado Department of Higher Education. It checks the exact file’s SHA-256 fingerprint before reading Table 1 on PDF page 12. A different report edition stops the calculation and requires a fresh source review. The recipe makes no network request, does not fetch textbooks and does not redistribute the official report.
For each displayed year, sum the cohort cells present to obtain the reported total. For each following row, subtract the previous total. Separately subtract each earlier cohort’s previous cell from its current cell, then add those differences. The newly introduced cohort’s first cell is the other component. The earlier-cohort component plus the new-cohort component must equal the change in the total. The five identities are checked exactly, with no balancing adjustment.
The files include the 21-cell reconstruction with page references, six row totals, five change bridges, six latest-transition cohort components and the two latest-row denominators. These 40 table rows combine source cells and derived summaries; they are not 40 independently observed outcomes. The first row is identified as a starting display rather than an observed change. Percentages use their named denominators and ordinary decimal rounding; no confidence intervals, causal estimates, student-level records or extrapolated 2026 outcomes are generated.
A citation that keeps the claim intact
A suitable short description is: “A reconstruction of Colorado’s 2025 OER report attributes 76.45% of the change between its 2023–2024 and 2024–2025 cumulative continued-use estimates to changes in earlier grant cohorts. This is reported-model accounting, not observed student cash savings or adoption retention.” Include the report edition, academic-year rows and the denominator when reusing the result.
Credit the Colorado Department of Higher Education for the underlying report and Alisson Bento for this reconstruction. The original report, its prose, layout, images, names and marks are not relicensed here. The original analysis and graphics have their own stated reuse terms, and the reproduction code has a separate code licence. Public availability of this reference does not establish institutional endorsement or independent verification of the report’s underlying course data.
Earlier and new cohorts in five reported-total changes
Within reported continued-use estimates, five stacked bars split each displayed-total change into earlier-cohort cell changes and the first amount of the newly introduced cohort. Earlier shares range from 61.94% to 86.09%; the latest is 76.45%. These are not observed annual cash savings.
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Row entered | Earlier-cohort component (USD) | New-cohort component (USD) | Total change (USD) | Earlier share of change |
|---|---|---|---|---|
| 2020–2021 | 3,925,488 | 2,411,695 | 6,337,183 | 61.94% |
| 2021–2022 | 6,337,183 | 3,619,867 | 9,957,050 | 63.65% |
| 2022–2023 | 8,526,995 | 1,377,384 | 9,904,379 | 86.09% |
| 2023–2024 | 9,904,399 | 2,139,202 | 12,043,601 | 82.24% |
| 2024–2025 | 13,473,635 | 4,150,647 | 17,624,282 | 76.45% |
For each transition: changes in earlier cohort cells plus the newly introduced cohort’s first amount equal the change in the displayed total. These are reported-model accounting changes.
How the last reported total is reconstructed
Reported continued-use estimates: the prior total of $42,167,701 plus $13,473,635 in changes in earlier cohorts and $4,150,647 for the newly introduced cohort reconciles exactly to $59,791,983. This is modeled accounting, not observed student cash savings.
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Grant cohort | Prior displayed cell (USD) | Final displayed cell (USD) | Component of change (USD) |
|---|---|---|---|
| Cohort 1 | 19,627,440 | 23,552,928 | 3,925,488 |
| Cohort 2 | 9,646,780 | 12,058,475 | 2,411,695 |
| Cohort 3 | 7,999,511 | 11,619,377 | 3,619,866 |
| Cohort 4 | 2,754,768 | 4,132,152 | 1,377,384 |
| Cohort 5 | 2,139,202 | 4,278,404 | 2,139,202 |
| Cohort 6 | — | 4,150,647 | 4,150,647 |
Compare 2023–2024 with 2024–2025 within the same report. A dash marks a cohort not yet present; it does not describe a measured zero outcome.
The denominator changes the percentage
Within reported continued-use estimates, earlier cohorts account for 76.45% of the latest $17,624,282 change but 93.06% of the final cumulative $59,791,983 total. The new cohort’s corresponding shares are 23.55% and 6.94%. These are accounting shares, not shares of students or cash savings.
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Question | Denominator (USD) | Earlier-cohort amount (USD) | Earlier share | New-cohort amount (USD) | New share |
|---|---|---|---|---|---|
| Composition of the latest change | 17,624,282 | 13,473,635 | 76.45% | 4,150,647 | 23.55% |
| Composition of the final cumulative row | 59,791,983 | 55,641,336 | 93.06% | 4,150,647 | 6.94% |
The $4,150,647 cohort 6 cell is 23.55% of the change and 6.94% of the final cumulative total. Neither percentage is a share of students, courses or cash spending.
Earlier cohorts’ share of the reconstructed increase
Colorado reported-model accounting: earlier cohorts contribute 76.45%, or $13,473,635 of the $17,624,282 increase between the final two rows. Their cumulative share is instead 93.06%. Not observed student cash savings or adoption retention.
Full chart and source
Scroll horizontally to inspect the chart. ↔
Chart and values
Scroll horizontally to see all columns.
| Question | Denominator (USD) | Earlier-cohort amount (USD) | Earlier share | New-cohort amount (USD) | New share |
|---|---|---|---|---|---|
| Composition of the latest change | 17,624,282 | 13,473,635 | 76.45% | 4,150,647 | 23.55% |
| Composition of the final cumulative row | 59,791,983 | 55,641,336 | 93.06% | 4,150,647 | 6.94% |
The $4,150,647 cohort 6 cell is 23.55% of the change and 6.94% of the final cumulative total. Neither percentage is a share of students, courses or cash spending.
Data & downloads
Scroll horizontally to see all columns.
| Question | Denominator (USD) | Earlier-cohort amount (USD) | Earlier share | New-cohort amount (USD) | New share |
|---|---|---|---|---|---|
| Composition of the latest change | 17,624,282 | 13,473,635 | 76.45% | 4,150,647 | 23.55% |
| Composition of the final cumulative row | 59,791,983 | 55,641,336 | 93.06% | 4,150,647 | 6.94% |
The $4,150,647 cohort 6 cell is 23.55% of the change and 6.94% of the final cumulative total. Neither percentage is a share of students, courses or cash spending.
Scroll horizontally to see all columns.
| Row entered | Earlier-cohort component (USD) | New-cohort component (USD) | Total change (USD) | Earlier share of change |
|---|---|---|---|---|
| 2020–2021 | 3,925,488 | 2,411,695 | 6,337,183 | 61.94% |
| 2021–2022 | 6,337,183 | 3,619,867 | 9,957,050 | 63.65% |
| 2022–2023 | 8,526,995 | 1,377,384 | 9,904,379 | 86.09% |
| 2023–2024 | 9,904,399 | 2,139,202 | 12,043,601 | 82.24% |
| 2024–2025 | 13,473,635 | 4,150,647 | 17,624,282 | 76.45% |
For each transition: changes in earlier cohort cells plus the newly introduced cohort’s first amount equal the change in the displayed total. These are reported-model accounting changes.
Scroll horizontally to see all columns.
| Grant cohort | Prior displayed cell (USD) | Final displayed cell (USD) | Component of change (USD) |
|---|---|---|---|
| Cohort 1 | 19,627,440 | 23,552,928 | 3,925,488 |
| Cohort 2 | 9,646,780 | 12,058,475 | 2,411,695 |
| Cohort 3 | 7,999,511 | 11,619,377 | 3,619,866 |
| Cohort 4 | 2,754,768 | 4,132,152 | 1,377,384 |
| Cohort 5 | 2,139,202 | 4,278,404 | 2,139,202 |
| Cohort 6 | — | 4,150,647 | 4,150,647 |
Compare 2023–2024 with 2024–2025 within the same report. A dash marks a cohort not yet present; it does not describe a measured zero outcome.
Scroll horizontally to see all columns.
| Academic-year row | Cohorts present | Sum of reported estimates (USD) |
|---|---|---|
| 2019–2020 | 1 | 3,925,488 |
| 2020–2021 | 2 | 10,262,671 |
| 2021–2022 | 3 | 20,219,721 |
| 2022–2023 | 4 | 30,124,100 |
| 2023–2024 | 5 | 42,167,701 |
| 2024–2025 | 6 | 59,791,983 |
Sums of the source’s cohort cells under its continued-use assumption. Not measured annual cash savings. The first displayed total is a starting row, not a change from an observed zero.
Scroll horizontally to see all columns.
| Academic-year row | Grant cohort | Reported estimate (USD) | Source PDF page |
|---|---|---|---|
| 2019–2020 | Cohort 1 | 3,925,488 | 12 |
| 2020–2021 | Cohort 1 | 7,850,976 | 12 |
| 2020–2021 | Cohort 2 | 2,411,695 | 12 |
| 2021–2022 | Cohort 1 | 11,776,464 | 12 |
| 2021–2022 | Cohort 2 | 4,823,390 | 12 |
| 2021–2022 | Cohort 3 | 3,619,867 | 12 |
| 2022–2023 | Cohort 1 | 15,701,952 | 12 |
| 2022–2023 | Cohort 2 | 7,235,075 | 12 |
| 2022–2023 | Cohort 3 | 5,809,689 | 12 |
| 2022–2023 | Cohort 4 | 1,377,384 | 12 |
| 2023–2024 | Cohort 1 | 19,627,440 | 12 |
| 2023–2024 | Cohort 2 | 9,646,780 | 12 |
| 2023–2024 | Cohort 3 | 7,999,511 | 12 |
| 2023–2024 | Cohort 4 | 2,754,768 | 12 |
| 2023–2024 | Cohort 5 | 2,139,202 | 12 |
| 2024–2025 | Cohort 1 | 23,552,928 | 12 |
| 2024–2025 | Cohort 2 | 12,058,475 | 12 |
| 2024–2025 | Cohort 3 | 11,619,377 | 12 |
| 2024–2025 | Cohort 4 | 4,132,152 | 12 |
| 2024–2025 | Cohort 5 | 4,278,404 | 12 |
| 2024–2025 | Cohort 6 | 4,150,647 | 12 |
Factual reconstruction of the cohort cells on PDF page 12; original report text, layout, images and funding column are not redistributed. Structural blanks are not converted into observed zeros.
Evidence & method
How we calculated it
- One exact 2025 source PDF; six displayed academic-year rows and 21 cohort cells; original values unchanged.
- Each transition equals changes in earlier cohort cells plus the newly introduced cohort’s first cell.
- Whole-dollar arithmetic, named percentage denominators and two-decimal display; structural blanks remain distinct from measured zeros.
What the data cannot tell us
- The report assumes continued use. The reconstruction does not observe adoption persistence or individual purchases.
- Institutional annual estimates and duplicated course enrollments are different universes; no per-student benefit or return on investment is calculated.
- The annual table caption, compounded figure and cumulative narrative are documented rather than silently harmonized. Source irregularities remain unchanged.
- No current programme-policy, 2026 outcome, causal learning or household cash-saving claim is made.
Original sources
- Colorado Department of Higher Education: 2025 Open Educational Resources Report
Source: Colorado Department of Higher Education, 2025 Open Educational Resources Report, acquired 1 November 2026; file last modified 2 December 2025. Table 1 on PDF page 12, Figure 1 and its continued-use note on page 11, institutional scope on pages 12–13 and duplicated enrollment note on pages 15–16. This is one historical report edition, not a current 2026 outcome series.
Accessed November 1, 2026
Cite & reuse this study
Alisson Bento. How earlier grant cohorts shape Colorado’s modeled OER savings. Reconstruction of the Colorado Department of Higher Education’s 2025 Open Educational Resources Report, Table 1, PDF page 12. https://alissonbento.com/research/colorado-oer-reported-cohort-accounting-2019-2025/
Original reconstruction, prose and graphics: CC BY 4.0 with attribution to Alisson Bento and separate credit to the Colorado Department of Higher Education as source. Original code: MIT. These terms do not license or redistribute the official source report, source graphics, textbooks, third-party works or marks.
Corrections or editorial questions