A 15th article published by BERA containing a hallucination (in this case misattributing a quote: the wrong article and page number). The writing is also pure AI (see 'not as...but as' etc. in extract)

It is encouraging that people reading this see that John Hattie is being protected. They recognise that it is ludicrous that: Melbourne only looked into 8% of the plagiarism allegations; Elgar somehow explained away a hallucinated reference. Absurd and indefensible.

If in 2026 a hallucinated reference is not treated as indication of AI-authorship then we may as well give up.

"At the time, the university had reviewed 14 of 181 of Vainker’s allegations at random, finding none amounted to misconduct." I have sampling questions.

Hattie: 'I am reluctant to use AI' Also Hattie: 'I use it every day...in fact I've used it twice this morning, there's a new tool out that actually writes the essay for me, and I've constructed one...' youtu.be/lCnSKwz6cQ4?...

The use of [sic] towards the end of the article is 👏🏻👏🏻👏🏻👏🏻 Well done @stephenvainker.bsky.social

The University of Melbourne has opened a formal investigation into the prominent Australia-based education researcher John Hattie, backtracking on a decision months ago that concerns about his work didn’t warrant further scrutiny.

We're reaching 'dog ate my homework' levels of excusing Hattie's misconduct: retractionwatch.com/2026/03/16/u... See my response: open.substack.com/pub/stephenv...

I just learned from Lawrence Rees that trench warfare wouldn’t have happened without tinned food… Is this an obvious thing I should have known?!

Here is how OpenAI is reimagining educational research. It's basically GPT accesses all student records and "over-time signals" for constant tracking. It's total student surveillance for its model improvement.

The British Educational Research Journal have now included in a special issue a paper that they've known for weeks contains a hallucinated reference. If anyone at @beranews.bsky.social cares about integrity, they're doing a good job of hiding it. doi.org/10.1002/berj...

Corwin have apparently removed five hallucinated references from a recent John Hattie book through a stealth edit. In desperately protecting Hattie (jettisoning basic principles of integrity), Corwin as well as the Uni of Melbourne are wrecking their reputation. open.substack.com/pub/stephenv...

Another BERA article with hallucinated reference published today (now 14). BERA continues to show absolutely no signs, 6 months after being told of the problem, to care in the slightest. doi.org/10.1002/curj...

It's completely all over the place. The methods and results are specifically focused on peer evaluation, but peer evaluation is not mentioned in the introduction or discussion. It's crazy. And the silence around the @beranews.bsky.social producing endless AI nonsense is remarkable.

This results section would fail any undergrad course. But it's published in a @beranews.bsky.social journal. It's weird to see the total collapse of editorial standards across the journals of an association of academics, and no one seeming to care. It makes me incredibly pessimistic.

"Among them, 44 students did an online questionnaire and 174 students responded to an online questionnaire." This new BERA paper looks great. Really top quality paper. Lots to offer. And why is marketing italicised?? doi.org/10.1002/curj...

BERA have now published 13 (thirteen) articles with hallucinated references. It's a scandal that will have to be confronted sooner or later. This time It's a PIRLS study (large datasets often targeted for LLM articles). doi.org/10.1002/berj...

BERA has published well over hundred LLM-authored articles. Each one is likely to contain misleading claims and must be retracted. Here's one illustration open.substack.com/pub/stephenv...

More evidence that many of the articles BERA publish are AI, even if they don't have fake references. This Abstract today in the Curriculum Journal is clearly AI-written even though the references all check out. But in the author's last paper, there was a fake reference. doi.org/10.1002/curj...

I've marked up a conclusion to an LLM written journal article, showing all the generic post-2023 phrases. It really clearly demonstrates the pre-packaged approach, that any particular topic or finding can just click into. open.substack.com/pub/stephenv...

BERA have published 12 articles with fake refs. This requires confronting a delicate issue. BERA has objectives to increase diversity. But all articles from countries without robust integrity standards. Has this has led them to being a soft touch on LLM articles? open.substack.com/pub/stephenv...

11th BERA article with fake refs. One to support this incredible 2024 sentence: 'Additionally, we conducted a systematic literature search across two academic databases, ProQuest & Taylor & Francis Online, as they have been recommended in the field of education research (Snyder & Snyder, 2008)'

Impressive, in a way to fit 5 lots of 'not (just) x, but y' into one BERA blog post. This really matters, not just because it fosters genuine irritation, but because it illuminates probable use of diverse AI tools. www.bera.ac.uk/blog/cant-le...

First and last sentence: it's not just x, it's Y (used 4 times overall). Second sentence "fostering active participation and curiosity" (never used before 2024, 7 times since). No fake reference this time, but as Hattie himself says, he loves to get AI to write for him.

Today a BERA journal published a highly misleading article on how AI can develop social & emotional learning. It uses evidence of AI boosting engagement to argue it can boost emotional competence. These are two very separate things; eliding them can only do harm. open.substack.com/pub/stephenv...

New paper in BERJ claims to extend Bourdieu beyond its 'traditional' focus on the individual. 😂 It manages to do this without mentioning the terms 'field' or 'habitus' once. 🤣🤣 Purest AI slop. doi.org/10.1002/berj...

John Hattie included a fake reference in a foreword to a handbook. Since I pointed this out it's been removed through a stealth edit. (Credit - @tommahoneyedu.bsky.social) open.substack.com/pub/stephenv...

Interessant 🤔, per pensar la frase que te @stephenvainker.bsky.social al seu perfil: “Interested in how the child came to be managed as an employee in our schools”. #eduCat

I've just discovered that 2 months after John Hattie threatened to sue me for calling him a plagiarist, he published a book containing this (and 13 other) instances of egregious plagiarism. It would be funny if he didn't remain such a respected & protected figure. open.substack.com/pub/stephenv...

Hi @berj-2025.bsky.social! The article "Afro-Amerindian cosmopolitics in environmental education" at doi.org/10.1002/berj... cites an article "Development of intercultural relationships at universities" in @jsinted.bsky.social. I'm not finding it. Could you let me know the correct reference please?

BERA continue to ignore flood of AI-authored articles. They've failed to retract 5 articles with multiple fake references almost 6 months after being informed. They continue to publish AI-authored articles, even with fake references. It's a scandal. stephenvainker.substack.com/p/bera-fails...

Despite the business-as-usual #BERA2025 conference, there's clearly a crisis. In the current Review of Education issue, 7 out of 8 articles appear LLM-written. Even worse: all four editors appear to have published LLM-written work (including a fake reference). open.substack.com/pub/stephenv...

John Hattie explaining in his own words how AI writes his papers: "there's a new tool out that actually writes the essay for me, and I constructed one, read it... the first one I though 'ah', then I realised I was at fault... so hone, hone, focus, quality control." www.youtube.com/watch?si=YpW...

I've read some absurd AI-written claims but this one from Hattie takes some beating: 'Online survey tools such as Kahoot! and Quizizz are effective tools to assist students in writing assignments, collecting data, and searching for research [19]' Love that it cites a source!

There are now two more AI-written articles co-authored by John Hattie. It makes sense a plagiarist would turn to LLMs, but it is still shocking that the most prominent education researcher of the 21st century is turning out AI slop, together with fake references. open.substack.com/pub/stephenvai…

Before AI, John Hattie plagiarised. Now, it appears he uses LLMs. A recent contribution to a handbook, published by Elgar, included a fake reference. The evidence could not be clearer... stephenvainker.substack.com/p/john-hattie-used-an-llm-to-write?r=695xb6

My final report on 131 LLM-authored articles in BERA journals: BERJ: 46 BJET: 59 Review: 22 Curriculum: only 4 These are staggering numbers & show that BERA, marginalised in the national discussion about education, has seriously lost its way. Change is required! open.substack.com/pub/stephenv...

Universities have definitions of plagiarism etc. so that it can clearly and fairly hold students & professors to account. So it's scandalous when Melbourne University didn't use it own definition in an investigation of John Hattie, & instead used a far softer one. open.substack.com/pub/stephenv...

Today's education establishment think they're FAR more sophisticated than John Hattie. But they've adopted his obfuscatory delusion of being purely evidence-based & non-ideological. We can't tackle the educational crises we're facing without addressing this. open.substack.com/pub/stephenv...

If you have a passing interest in John Hattie's plagiarism but don't want (for some strange reason) to read an over-long academic article on it, you may be interested in this 👇 open.substack.com/pub/stephenv...

You might think a university would use its own definition of plagiarism when investigating a professor. You might also think a university would look at all allegations of plagiarism in its investigation. You'd be wrong. open.substack.com/pub/stephenv...

You might think a university would use its own definition of plagiarism when investigating a professor. You might also think a university would look at all allegations of plagiarism in its investigation. You'd be wrong. open.substack.com/pub/stephenv...

Al articles are a road to nowhere, involving a great deal of pain. Beginning: shameless platitudes Middle: dense technical description of elaborate relationships between endless variables. End: shameless platitudes

One AI-written article is a tragedy, 57 is a farce. stephenvainker.substack.com/p/57-ai-writ...

One AI-written article is a tragedy, 57 is a farce. stephenvainker.substack.com/p/57-ai-writ...

I've updated the paper on John Hattie to include plagiarism from 2005, 2021, & 2025. I've now shown that he has plagiarised over 19,500 words across 11 different works, from 1981-2025. It's been 5 months & I've received no indication of any errors. researchgate.net/publication/...

This has to be a low point. An article in the BERJ about the experience of playing 'TargetShootingMania' with 21 fake references going into granular detail about the movement of the cardboard boxes that need to be shot. The research questions? Nothing to do with education. At all. This is madness.

If we allow our flagship journals to fill with AI slop (presumably pushing out excellent or at least competent work in the process) then we are making a better case for our marginalisation than any critic ever could. open.substack.com/pub/stephenv...

Any worse writing style than AI florid?