Showing posts with label referencing. Show all posts
Showing posts with label referencing. Show all posts
Saturday, 14 February 2026
A research article’s digital cul-de-sac. How to fix an academic paper that is hidden and/or unreachable
Written for researchers keen to understand potential obstacles to their articles not being readily searchable via academic indexes and search engines thanks to broken linkages or a metadata mismatch.
Most researchers would seem keen for their works to be well-indexed, quick-to-find, and reliably linked to their academic publisher’s landing page for that paper via search engines. This is especially true for scholars whose performance reviews include consideration of their works’ impact in reaching interested readers. Such impact scores can include citations, amplification and readership, as tracked via ALtmetrics and/or other sources.
Most researchers would seem keen for their works to be well-indexed, quick-to-find, and reliably linked to their academic publisher’s landing page for that paper via search engines. This is especially true for scholars whose performance reviews include consideration of their works’ impact in reaching interested readers. Such impact scores can include citations, amplification and readership, as tracked via ALtmetrics and/or other sources.
This post provides an example to illustrate why a research author should not assume that every article of his or hers will be properly indexed. As result, this work may not be seen by potentially interested parties searching for pertinent work with salient keywords, despite the work's merits. Metadata mismatches and broken linkages can unwittingly emerge during academic publishing, and may persist if not checked and corrected. This post gives helpful advice for correcting a few key problems, sharing the practical example of errors that impacted a South African journal article's availability shortly after its digital publication on the 1st of December, 2025:
'Brandjacked for social media advert fraud'- a ghost in the fog...
The first publication from a research project that began in 2021, Brandjacked for social media advert fraud: Microcelebrities' experiences of digital crime, was published in the Acta Criminologica: African Journal of Criminology & Victimology journal. Ordinarily, new articles should only take a few weeks after being indexed by CrossRef before they appear in the paper's authors academic profiles (such as Google Scholar linked to ORCID), listings under academic social networks and in search engine and AI results. While results for the article were evident under the last, it could not be found under the former. This post explores such challenges and why the article took over two months to be properly indexed for academic searches. While the paper technically existed, it could not readily be shared online by its four authors, while listing results linked to a 'page not found' error message, rather than the publisher's article landing page.
Multiple URLs as expected
A normal part of the academic publishing process sees several URLs being created for a particular publication. In Brandjacked's case, these are from the handle.net domain (pointing to https://hdl.handle.net/10520/ejc-crim_v38_n3_a2) and the journals.co.za one (onto https://journals.co.za/doi/10.10520/ejc-crim_v38_n3_a2). Sabinet African Journals (journals.co.za) is a searchable platform providing a comprehensive, full-text collection for over 600 African-published electronic journals. Handle.net is developed by the Corporation for National Research Initiatives (CNRI) to support a distributed information system for persistent, unique, and actionable identifiers (handles) for digital objects over the internet. The Handle System is the underlying technology used by the Digital Object Identifier (DOI) system to identify content, particularly for digital publishing and commerce.
Brandjacked's DOI value changed post-publication
A DOI is a unique, permanent alphanumeric string that is assigned to digital objects. DOIs include a prefix always starting with 10 and a suffix, separated by a forward slash (/). Prefacing the DOI with doi.org creates an actionable link. Originally, the brandjacked article was allocated the DOI 10.10520/ejc-crim_v2025_nse1_a2. However, this was subsequently changed to 10.10520/ejc-crim_v38_n3_a2, reflecting the article’s publication as the second in the 38th volume’s third issue.
Commonly linked to journal articles, books, and research datasets, each DOI should provide a stable, long-lasting link to a research output’s internet location, regardless of whether its online address [or Uniform Research Location (URL)] gets changed.
Social media shares…
On social media, academics, researchers and scholars typically share the DOI version of URLs, since DOIs usage is the primary identifier for tracking comments and mentions via AltMetrics. The attention for academic publications is sparse, with most not being shared on social media, and a significant portion only being shared once via popular social media platforms. Only a few papers, often with high public interest or "quirky" topics, receive the vast majority of shares. To help Brandjacked's visibility, I wrote a 31 tweet long thread at x.com/travisnoakes/status/2001585342000898177, shared a LinkedIn post at linkedin.com/feed/update/urn:li:activity:7407406366052569089/, and wrote this public Facebook post at facebook.com/share/p/1BtsXaJtQC/.
… but no AltMetrics listing yet
On the Brandjacked article’s landing page, no Altmetrics were shown at the date of this post's publication (see Figure 1 below).
 |
| Figure 1. No AltMetrics listing for Brandjacked article. |
Nor are results for the Brandjacked article are shown on https://app.dimensions.ai/discover/publication?search_mode (see Figure 2).
 | |
|
As a research blogger, I like to share Altmetrics badges for my publications, but none is available for Branjacked via badge.dimensions.ai/details/doi/10.10520/ejc-crim_v38_n3_a2?domain=https://journals.co.za (see Figure 3).
No automatic article updates to the authors’ ORCID profiles
Each of Brandjacked’s four authors have an ORCID address. This is a 16-digit persistent digital identifier (e.g., 0000-0001-9566-8983) that uniquely identifies research-contributors all over the world. Like the DOI value does for research outputs, ORCID resolves name ambiguity. It links all research publications, datasets, and grants to an individual, regardless if their name or institution changes. Authors can set their ORCID profile to be automatically updated with research outputs that are linked to them. As of this post’s date of publication, the Brandjacked article was not linked from any of the authors’ ORCID profiles.
Google Scholar links to a 404 error on the article's landing page
The academic search engine, Google Scholar, is popular due to offering the widest automatic global indexing of scholars' works. In its case, Google Scholar account holders can create manual entries for their articles, or select pre-existing records featuring their author name.
In Brandjacked’s case, I created a manual record in late December, since no record for this paper was available for selection (Figure 3). Since then, Google Scholar has come to list two versions of the article.
 |
| Figure 3. Two versions of Brandjacked on Google Scholar. |
The top version, which can be Saved and Cited, produced a 404 error when clicked through to Acta Criminologica (Figure 4).
 |
| Figure 4. DOI not resolved from Google Scholar |
Thinking that this may have been caused by the manual record I created featuring the initial DOI, I chose to delete the article from my record. Then went into Google Scholar trash and selected ‘Delete Forever’. However, the article available for manual selection (Figure 5) still links to a 404 result.
 |
| Figure 5. Google Scholar manual selection of Brandjacked |
Reporting a failed DOI link via doi.org
A manual entry created on ResearchGate (at https://www.researchgate.net/publication/398814505_Brandjacked_for_social_media_advert_fraud_Microcelebrities'_experiences_of_digital_crime_in_South_Africa) lists the correct DOI (Figure 6), but it also showed a 404 error when linking to the digital publication.
 |
| Figure 6. ResearchGate DOI is correct for Brandjacked |
At least this page presented an option to report the failed DOI. So, I submitted this error report to DOI: "The article 'Brandjacked for social media advert fraud' should be available at this DOI. Related working URLs are: https://journals.co.za/doi/10.10520/ejc-crim_v38_n3_a2 & https://hdl.handle.net/10520/ejc-crim_v38_n3_a2. Thanks in advance for fixing this."
 |
| Figure 7. DOI error report for Brandjacked, 12 Feb 2026 |
As a shortcut for checking-and-reporting DOI errors, authors should open www.doi.org, then scroll down to the 'Try resolving a DOI name' section. There they can enter a DOI value to check it, and report a DOI URL address that fails in throwing a 404 error.
![DOI Prefix [10.10520] Not Found.png](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjDEzqL4beKD8zMIGkhp6iWGUj-JBn6HNu5Kxiuk7YzppaCG5je4xEOMY0xNKm7XroYFTLduiyNrxPPZprQV-PDNvHlEmgVh5_aY4jmLe47-YFoi1U09UxXqurCEexdlvbO6fPvuf-BQ5T4CPmPYtfHufuU5oqAXWrvsZyzt5Tl5uQim1lw1W7pTg-qI1sq/s2108/DOI%20Prefix%2010.10520%20Not%20Found.png "DOI Prefix [10.10520] Not Found.png") |
| Figure 8. DOI Prefix 10.10520 Not Found.png |
A few downloads from restricted access, the holidays... and a ghost listing?
Acta Criminologica’s ‘Open Access’ policy initially provides restricted access via SABINET's archives. These are available via university libraries’ annual SABINET subscription for SA_ePublications (Sabinet African Electronic Publications). There is a 12-month open access embargo from the date of publication/loading on CRIMSA’s website. Perhaps being published before the Christmas holidays in an access restricted format contributed to Brandjacked only being downloaded five times in two months. But even achieving that readership is not been helped by a ghost listing amidst a heavy stream of academic publications. As Trend MD reports, nearly 8,000 research articles are published everyday!
Writing a letter to the Editor
Given that the landing page URL is throwing a 404 error and the DOI is not resolving, it seems likely that the error is a flaw on the publisher's end. I have written to the journal's editor to request that the CrossRef metadata is updated. Hopefully, the journal’s production team can soon ensure that: (i) the landing page URL for the article is correctly mapped to its DOI. And that (ii) the metadata provided to CrossRef and other aggregators is updated to reflect a functional DOI link. I did flag that this issue impacts other articles in the Cybercrime special issue, so it's crucial to fix for all its authors’ benefit, too.
Conclusion
After overcoming the challenges of review, revisions, and achieving a digital publication, academic authors may face a new challenge- following up that their publications are well-indexed, as this should not be assumed. Hopefully this post presents a helpful example of this problem, and ideas that authors can follow to to resolve it.
N.B. This blogpost is a work-in-progress, as the problem is not fixed, yet. I will update the post to reflect new learnings as the issue gets fixed. Do add a comment below if you have any related suggestions or other feedback...
N.B. This blogpost is a work-in-progress, as the problem is not fixed, yet. I will update the post to reflect new learnings as the issue gets fixed. Do add a comment below if you have any related suggestions or other feedback...
Saturday, 6 April 2019
My Cursed Referencing = Google Scholar citation imports + legacy Refworks
A cautionary tale for researchers using legacy Refworks and importing citations via Google Scholar (hint: it's not a "shortcut").
Probably the most helpful change for my purposes was that the Legacy Refworks did not flag incomplete citations. By contrast, ProQuest Refworks did this in yellow for essential and blue for optional information. This proved helpful for Lungile Madela and I as we worked through correcting my Refworks database in one week.
I sincerely hope that there will be no further episodes of 'My Cursed Referencing', but commit to writing a sequel post if they do... Here's to helping others avoid or overcome similar predicaments.
The Academic Referencing Horror Story- a neglected genre?
There seems to be a dearth of blogposts that share academic referencing disasters in detail. This should not be surprising, since all the role-players in such dramas have little to gain from sharing their time-sucking examples: Scholarly search engines and referencing software conceal their flaws within legalese and support departments. Institutions will not spotlight their limitations in offering minimal support. Academics' reputations may suffer for sharing referencing mistakes. Even students who share their frustrations and experiences publicly (i.e. via tweets) seem to lack the motivation for elaborating these into lengthier narratives (e.g. a twitter roll). Further, since the costs are paid by the "free time" and suffering of individuals, there seems to be "no need" for systemic interventions that might address the attendant losses of scholars' time and morale.
On the flipside, students and researchers who lack examples of what to avoid MUST be widely repeating the errors of their peers, surely? This may range from colleagues within a particular university to software users globally whose experiences of a service's functionality differs from its promises. Just, maybe, the downside of sharing academic referencing horror stories SHOULD not trump the importance of providing important learnings for online audiences?!
In the spirit of being transparent about such disasters for others to avoid similar suffering, here's 'My Cursed Referencing'. It compiles key incidents with a cast of fellow researchers, an Apple laptop running Microsoft Word 2010, Google Scholar, Refworks Legacy and ProQuest versions plus support, the University of Cape Town, plus yours-victim-truly.
MY CURSED REFERENCING
The internet was just becoming a thing when I completed my MA thesis using Claris Works in 1997, so I was pretty much an online referencing virgin when I returned to university 12 years later. In terms of this type of software, Refworks seemed the obvious choice. It was offered as a free tool at UCT and seemed easy to access and use. Besides, which PhD candidate can prioritise the time to: (i) research the universe of referencing software and compare their upsides and downsides, (ii) check they will install on Mac and integrate well with his or her university library's back-end, plus (iii) investigate if the support post-purchase will be sound? Doing the PhD reading on 'Diffusion Model of Innovation' and 'Concerns Based Adoption Model' was daunting enough, thanks...
The Unknown Evil of using Google Scholar exports for creating a bibliography
If Google Scholar valued 'Don't Be Evil', its 'import reference' function would be labelled 'import an (in)complete reference' as a fair warning. A bibliography made up of such exports for recent journal articles may well be accurate and complete. BUT should your sources range across old books, scholarly dissertations and online sources, be vigilant. Your citations are probably incomplete AND imported in the wrong format (i.e. a 'book section' can be indexed as a 'journal article' by default).
Evils is Even (bad things come in twos)
I was alerted to this conundrum after presenting a draft of 'Capital meets Capabilities' to the Technology in Education Postgraduate Researchers group in January, 2018. A senior researcher observed, 'You didn't generate your references with software'. Stunned, I answered that indeed I had. After requesting that he highlight the errors in my bibliography, I was mortified to see copious redlining that highlighted many of my references missing information...
This was horrific, because of the time it must take to correct. I was preparing my thesis for submission in early February and could ill afford prioritising additional work. My legacy Refworks database had around 1,600 references in it. If I used 1,200 of those in my thesis and 1,000 required correction with six minutes for finding the missing information, I would need to find 100 hours to fix this giant mistake.
On a related note, my main supervisor flagged another major citation evil. The particular Harvard referencing style I had used (since my BAFA Hons in '94, ahem) might be flagged as 'outdated' by external reviewers. They might prefer me to use one of the other nine Harvard styles. Prof. Marion Walton recommended that I shift to the American Psychological Association, which has just one style. It's widely used in Media Studies, so would also prove useful in developing articles from my PhD. In response, I learnt this new style and restyled my in-line citations.
Enter ProQuest Refworks, enter light (yellow)?
By default, the UCT off-campus login (see Figure 1) points to the old version of Refworks . The option to upgrade was listed in a tiny box on the top right of Legacy Refworks (see Figure 2). I clicked on this link to find out what had changed...
 |
| Figure 1. Screenshot of UCT off campus login at https://login.ezproxy.uct.ac.za/menu shows Refworks' high salience. |
 |
| Figure 2. Screenshot of the Legacy Refworks screen, which shows the low salience of its top left link to ProQuest Refworks. |
Probably the most helpful change for my purposes was that the Legacy Refworks did not flag incomplete citations. By contrast, ProQuest Refworks did this in yellow for essential and blue for optional information. This proved helpful for Lungile Madela and I as we worked through correcting my Refworks database in one week.
Exported references with Missing Parts
We had taken considerable care to ensure each reference was as complete as it could be, but some references became incomplete on export. For example, the PhD and Masters dissertations (in which we tried variations for 'Faculty' and 'Degree' under the 'thesis' type) would not display all their information under the citation view, nor in the exported bibliography. At the time, the only UCT specialist who could advise on a workaround was on leave. I did learn from ProQuest Refworks support that the university and type of degree are not mapped from the old database into the new one. So, I did a *. search for all thesis entries, then manually cut-and-pasted the correct version of each citation into my bibliography. This points again to the importance of understanding the constraints of the referencing software you use, versus what it does.
The Disconnected Bibliography of the Damned
Another important constraint existed in my use of an old version of Microsoft Word that did not support the use a ProQuest Refworks add-in. In particular, it automatically generates a bibliography from in-line citations. By contrast, read-through the bibliography to remove sources that had become irrelevant. I then passed my thesis through Reciteworks, a free APA and Harvard citations checker. It proved super-useful for matching in-line citations with the bibliography and also identifying stylistic errors to fix.
The Summoning of my Legacy Refworks database and its Zombie Citations
I eventually had to update to the latest version of Word for handling my large thesis file (Word 2010 crashed repeatedly while I combined all chapters into a thesis). After submitting it, I started to use the Refworks Citation Manager for manuscripts from my thesis and the 'online academic bullies and mobs' project. This add-in worked well for me until a 'new projects' functionality upgrade was launched. Projects are a better way for organising references at a high-level; rather than showing all of them at once, one can associate a Word document with a particular project's references. I split my projects into two; one for cyberbullies, the other for my 'Inequality in Digital Personas' PhD. |
| Figure 3. The ProQuest Refworks add-in for Microsoft Word shows old folders from my Legacy Refworks database. |
At last on the leading edge of referencing technology, an unexpected downside was that my'PhD' project combined with my UCT proxy access to summon the Legacy Refworks database's return. The project's old database did not reflect the new folders, citations and corrected references in my ProQuest Refworks database. Thankfully, Rich and Jay from its support portal were very helpful in organising my Legacy Refworks database's deletion.
I sincerely hope that there will be no further episodes of 'My Cursed Referencing', but commit to writing a sequel post if they do... Here's to helping others avoid or overcome similar predicaments.
Subscribe to:
Comments
(
Atom
)
Posts
orcid.org/0000-0001-9566-8983