Statewide report · 2015–2025 · OpenAlex data as of November 17, 2026

Two-thirds of Connecticut’s research is free to read. The rest is behind a paywall.

An analysis of 174,990 research articles by authors at 17 Connecticut colleges and universities, published from 2015 to 2025, using the open database OpenAlex.

Key findings

  1. 67.3% of articles are free to read (117,742 of 174,990). Without Yale, the figure is 59.6%. Yale is the only Connecticut institution on 64% of these articles, so it shapes the statewide number.
  2. The share went from 59.6% in 2015 to a peak of 74.0% in 2023, and was 70.8% in 2025. The recent dip is mostly in bronze articles, which may still become free.
  3. Institutions range from 44.2% (Sacred Heart University) to 78.3% (University of Bridgeport).
  4. Elsevier published the most closed articles: 15,664 of 42,089 (37%).
  5. Authors may have paid about $60.1 million in publishing fees, based on journals’ listed prices Estimate
  6. About 43,141 closed or bronze articles could likely be made free by depositing the author’s accepted version in a repository Proxy

How these figures are made, and their limits

Ways an article can be free

There are several routes to free reading. They matter to libraries because they cost different amounts and last for different lengths of time. These are categories, not rankings. Together they cover every article.

Diamond
Free to read in a journal that charges authors nothing (fee confirmed as $0).
Open journal, fee not confirmed
Free to read in an open access journal, but the data does not confirm whether authors pay a fee.
Gold
Free to read in a fully open access journal that usually charges authors a fee.
Hybrid
Free to read in a subscription journal because the author paid a fee for this article.
Bronze
Free to read on the publisher’s site without an open license. The publisher could take it back.
Green
The journal version is paywalled, but a free copy is in a repository (an online archive).
Closed
Behind a paywall, and no free copy could be found.

How it has changed since 2015

Figure 1Share of articles free to read, by year of publication

Statewide, the share of articles free to read rose from 59.6% in 2015 to a peak of 74.0% in 2023, then fell to 70.8% in 2025. Without Yale it went from 50.3% in 2015 to 65.1% in 2025.

The drop after 2023 is mostly in bronze articles, which went from 15.9% to 7.9% of the total. Some publishers make articles free to read a year or more after publication, so recent years are likely to rise as the data updates. Later data refreshes will show whether that happens.

2026 is shown on its own because the year is not over and indexing lags: 66.8% of the 17,435 articles counted so far are free to read.

Show the numbers as a table for Figure 1
Share of articles free to read, by year
YearStatewide: free / totalStatewide: shareWithout Yale: free / totalWithout Yale: share
20157,670 / 12,87859.6%2,414 / 4,79750.3%
20168,189 / 13,37161.2%2,541 / 5,00450.8%
20178,413 / 13,71761.3%2,693 / 5,08553.0%
20188,808 / 14,04962.7%2,852 / 5,18755.0%
20199,619 / 14,87664.7%3,127 / 5,42657.6%
202011,568 / 17,21767.2%3,650 / 6,08660.0%
202111,980 / 17,43768.7%3,787 / 6,17761.3%
202212,025 / 16,92871.0%3,674 / 5,84062.9%
202312,611 / 17,04474.0%3,928 / 5,79267.8%
202413,191 / 18,16772.6%4,116 / 6,20466.3%
202513,668 / 19,30670.8%4,261 / 6,55065.1%
2026 (incomplete)11,640 / 17,43566.8%3,668 / 5,89262.3%

Source: OpenAlex, via the project pipeline (free_to_read_by_year.csv).

Figure 2How articles were free to read, by year (all 17 institutions)

The mix has shifted toward routes where authors usually pay. Gold went from 10.4% in 2015 to 26.9% in 2025, and hybrid from 5.5% to 16.6%. Green (a free copy in a repository) went from 22.1% to 13.0%, and closed from 40.4% to 29.2%.

  • Diamond
  • Fee not confirmed
  • Gold
  • Hybrid
  • Bronze
  • Green
  • Closed
Show the numbers as a table for Figure 2
Share of articles in each category, by year (all 17 institutions)
YearDiamondFee not confirmedGoldHybridBronzeGreenClosed
20151.3%3.2%10.4%5.5%17.0%22.1%40.4%
20161.1%3.8%11.7%5.5%16.9%22.3%38.8%
20171.4%3.7%13.1%5.6%17.2%20.3%38.7%
20181.1%4.3%14.5%6.9%16.4%19.5%37.3%
20190.8%4.7%16.6%6.5%15.9%20.2%35.3%
20200.7%5.0%18.6%7.8%15.1%19.9%32.8%
20210.6%4.8%22.1%9.1%13.8%18.3%31.3%
20220.7%5.1%25.9%10.9%12.6%15.9%29.0%
20231.0%5.4%25.4%13.6%15.9%12.7%26.0%
20240.8%5.9%25.4%14.7%12.7%13.1%27.4%
20250.8%5.5%26.9%16.6%7.9%13.0%29.2%
2026 (incomplete)0.8%4.9%27.3%18.4%4.0%11.4%33.2%

Source: OpenAlex, via the project pipeline (oa_category_by_year.csv).

By institution

Figure 3Share of articles free to read, 2015–2025, by institution

Shares range from 44.2% at Sacred Heart University to 78.3% at University of Bridgeport. 3 of the 17 institutions are above the statewide figure of 67.3%. An article with authors at two institutions counts for both. Smaller institutions have fewer articles, so their shares move more from year to year.

Show the numbers as a table for Figure 3
Articles free to read, 2015–2025, by institution
InstitutionArticlesFree to readShare
University of Bridgeport75559178.3%
Yale University116,14983,02371.5%
Wesleyan University2,7161,85868.4%
University of New Haven1,8841,16661.9%
University of Connecticut (incl. UConn Health)45,28427,59260.9%
United States Coast Guard Academy19411760.3%
Trinity College85151059.9%
Southern Connecticut State University1,66190554.5%
Fairfield University1,75294754.1%
Connecticut College68536252.8%
University of Saint Joseph50926852.7%
Western Connecticut State University42222052.1%
Eastern Connecticut State University57529351.0%
Quinnipiac University2,4401,23350.5%
University of Hartford1,20559149.0%
Central Connecticut State University1,39466447.6%
Sacred Heart University1,34659544.2%
Statewide total (each article once)174,990117,74267.3%
Without Yale (each article once)62,14837,04359.6%

Source: OpenAlex, via the project pipeline (free_to_read_2015_2025.csv).

Figure 4How each institution’s articles were free to read, 2015–2025

Across all 17 institutions, the largest free route is gold (19.9%), followed by green (17.6%). 32.7% of articles are closed, and 40.4% without Yale.

  • Diamond
  • Fee not confirmed
  • Gold
  • Hybrid
  • Bronze
  • Green
  • Closed
Show the numbers as a table for Figure 4
Share of articles in each category, 2015–2025
InstitutionDiamondFee not confirmedGoldHybridBronzeGreenClosed
Connecticut (all 17 institutions)0.9%4.8%19.9%9.8%14.4%17.6%32.7%
Connecticut without Yale0.6%6.5%17.9%8.2%12.4%14.1%40.4%
University of Bridgeport0.4%21.5%28.1%8.9%13.6%5.8%21.7%
Yale University1.1%3.8%21.0%10.6%15.4%19.5%28.5%
Wesleyan University0.2%5.4%16.4%10.9%17.5%18.0%31.6%
University of New Haven0.6%14.5%16.6%12.0%10.5%7.7%38.1%
University of Connecticut (incl. UConn Health)0.4%5.1%19.1%8.1%13.0%15.2%39.1%
United States Coast Guard Academy0.5%17.5%11.9%6.2%11.3%12.9%39.7%
Trinity College0.0%6.7%17.9%7.9%9.5%18.0%40.1%
Southern Connecticut State University1.0%11.1%11.0%6.9%11.6%12.9%45.5%
Fairfield University7.6%6.7%9.1%10.0%9.9%10.6%45.9%
Connecticut College0.1%9.5%12.4%8.2%10.5%12.1%47.2%
University of Saint Joseph0.4%13.2%14.7%5.5%8.6%10.2%47.3%
Western Connecticut State University0.5%15.6%12.1%2.6%10.4%10.9%47.9%
Eastern Connecticut State University0.5%9.9%13.7%9.4%8.9%8.5%49.0%
Quinnipiac University0.1%7.0%17.4%7.1%10.0%8.9%49.5%
University of Hartford0.3%9.5%13.4%5.9%9.1%10.9%51.0%
Central Connecticut State University0.4%12.8%14.3%4.4%7.2%8.7%52.4%
Sacred Heart University0.1%12.3%11.5%6.5%7.1%6.6%55.8%

Source: OpenAlex, via the project pipeline (oa_category_2015_2025.csv).

Where the closed articles are

Figure 5Publishers with the most closed articles, 2015–2025

Elsevier published the most closed articles: 15,664 of 42,089 (37%). Next are Springer Nature (6,239) and Wiley (5,670). Imprints are grouped under their parent company.

  • Elsevier15,664 closed of 42,089 (37%)
  • Springer Nature6,239 closed of 21,946 (28%)
  • Wiley5,670 closed of 16,031 (35%)
  • Wolters Kluwer5,240 closed of 9,517 (55%)
  • Taylor & Francis4,472 closed of 7,161 (62%)
  • SAGE Publishing3,478 closed of 6,739 (52%)
  • Oxford University Press1,833 closed of 7,959 (23%)
  • American Chemical Society1,733 closed of 3,228 (54%)
  • Publisher not recorded1,342 closed of 5,193 (26%)
  • IEEE872 closed of 1,586 (55%)
Show the numbers as a table for Figure 5
Publishers with the most closed articles, all 17 institutions, 2015–2025
PublisherArticlesClosedClosed share
Elsevier42,08915,66437.2%
Springer Nature21,9466,23928.4%
Wiley16,0315,67035.4%
Wolters Kluwer9,5175,24055.1%
Taylor & Francis7,1614,47262.4%
SAGE Publishing6,7393,47851.6%
Oxford University Press7,9591,83323.0%
American Chemical Society3,2281,73353.7%
Publisher not recorded5,1931,34225.8%
IEEE1,58687255.0%

Source: OpenAlex, via the project pipeline (publishers_2015_2025.csv).

Figure 6Fields with the most closed articles, 2015–2025

Medicine has the most closed articles: 21,128 of 65,177 (32%). Fields are OpenAlex’s broad subject groups, assigned to each article’s main topic.

  • Medicine21,128 closed of 65,177 (32%)
  • Social Sciences7,732 closed of 16,317 (47%)
  • Psychology4,078 closed of 10,025 (41%)
  • Arts and Humanities2,463 closed of 4,412 (56%)
  • Engineering2,423 closed of 5,276 (46%)
  • Environmental Science2,107 closed of 6,300 (33%)
  • Biochemistry, Genetics and Molecular Biology2,052 closed of 14,178 (14%)
  • Economics, Econometrics and Finance1,465 closed of 3,985 (37%)
  • Materials Science1,377 closed of 3,049 (45%)
  • Neuroscience1,351 closed of 6,468 (21%)
Show the numbers as a table for Figure 6
Fields with the most closed articles, all 17 institutions, 2015–2025
FieldArticlesClosedClosed share
Medicine65,17721,12832.4%
Social Sciences16,3177,73247.4%
Psychology10,0254,07840.7%
Arts and Humanities4,4122,46355.8%
Engineering5,2762,42345.9%
Environmental Science6,3002,10733.4%
Biochemistry, Genetics and Molecular Biology14,1782,05214.5%
Economics, Econometrics and Finance3,9851,46536.8%
Materials Science3,0491,37745.2%
Neuroscience6,4681,35120.9%

Source: OpenAlex, via the project pipeline (fields_2015_2025.csv).

Articles that could be made free in a repository Proxy

Many publishers let authors put the accepted version of a paywalled article (the final text before the publisher’s formatting) in a repository, the route called green open access. This is one of the cheapest ways a library can make research free to read.

Likely candidates, all 17 institutions Proxy

43,141

24.6% of all 174,990 articles, 2015–2025.

Closed articles among them

37,594

of 57,248 closed articles.

Bronze articles among them

5,547

of 25,165 bronze articles, which are free to read now but could be withdrawn.

This is a proxy, not a policy check.

An article counts if it is closed or bronze, has no free copy anywhere except the publisher’s page, and its journal “shows deposit evidence”: at least 3 of the journal’s articles worldwide are free only as an accepted manuscript in a repository other than PubMed Central. Publisher policies and embargoes still need checking article by article. With a looser rule (1 deposit) the count would be 55,803; with a stricter one (10 deposits), 25,656. How this works.

Librarians can download the article list for planning outreach: repository candidates (CSV). It names no authors.

What authors may be paying to publish Estimate

Many open access journals charge authors an article processing charge (APC). OpenAlex records each journal’s listed price, not what anyone actually paid, so these figures are estimates. They count gold and hybrid articles from 2015 to 2025, each article’s fee once.

Estimate based on corresponding authors Estimate

$60.1 million

22,375 gold and hybrid articles where a corresponding author (the one who usually arranges payment) is at one of the 17 institutions. 2,595 of them have no listed price and are not included. This is the more realistic figure.

Upper bound, any Connecticut author Estimate

$145.4 million

51,896 gold and hybrid articles with any author at the 17 institutions, even when the fee was likely paid elsewhere. 6,230 have no listed price.

Without Yale, corresponding authors Estimate

$16.7 million

7,237 articles where a corresponding author is at one of the other 16 institutions.

Why these are estimates, and which way they could be off

Listed prices are not payments. Discounts, waivers and library agreements lower what is actually paid, and prices rise over time, so earlier years are likely overstated. When two Connecticut institutions are both corresponding, the fee is credited to each in their own figures, but counted once here. Only 82.5% of articles have a corresponding author’s institution recorded. 8,329 articles are in open access journals whose fee is not confirmed and are not counted, which may understate the total.

Fees that other universities report actually paying (to OpenAPC) for the same articles average well under today’s listed prices. See the comparison on the Methods page.