Scientific Publishing Dos And Don'ts

A manuscript does not enter a journal as a finished object. It enters as a decision problem. Should this paper be sent to reviewers? Can its contribution be defended? Does it belong in this journal’s conversation? Are the risks manageable? Is there a credible route to publication? Publishing can feel opaque. I want to help make it less so, especially for scholars whose ideas deserve to be heard but who lack access to the informal knowledge that helps work move through the system. Yesterday, I gave a talk at Henley Business School on publishing from the editor’s side, focused on the judgements authors do not always get to see: what creates confidence, what raises concern and why technically competent papers can still struggle to find a publication pathway. If I were to reduce the talk to one piece of advice, it would be this: do not write only for reviewers. Write in a way that helps the editor become confident in the paper! That confidence is not created by polish alone. It comes from the relationship between contribution, theory, method, evidence, fit and audience. The strongest manuscripts do not simply present a study. They make a publishable proposition. That means making the contribution visible early. A gap is not enough. Editors need to see what the paper changes in how readers understand a problem. It means treating journal fit as more than topic fit. A paper can be about something a journal publishes and still not belong there. The stronger question is whether it advances a conversation the journal’s readers care about now. It means using the introduction as the editorial case. By the end of the opening pages, the editor should be able to say what the problem is, why it matters, what conversation the paper enters, what claim it makes and why that claim deserves review. It also means seeing the response letter as an editorial document. A strong revise-and-resubmit does not simply say, “we did everything.” It shows that the authors understood the editorial risk and that the manuscript now has a clearer, stronger and more credible trajectory. I should also say that I am only able to do this work because I have learned, and continue to learn, from many outstanding colleagues across the journals where I serve, whether as an editorial board member, guest editor, associate editor or consulting editor. So let me do some shout-outs! At the International Journal of Human Resource Management, thank you to Michael Dickmann and Emma Parry. At the Journal of Business Research, thank you to Mariano (Pitòsh) Heyden and Mirella Kleijnen. At the Journal of Occupational and Organizational Psychology, thank you to Prof. Julie Gore, Luke Fletcher and Shaun Pichler, MSHR, Ph.D. And at Long Range Planning, thank you to Roberto Vassolo and Thomas C. Lawton. Thank you also to Bernd Vogel, Anastasiya Saraeva, Alex Baker and the LOBR community for the invitation and for such a great discussion.

𝐈𝐬 #AI 𝐢𝐧 #Engineering 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐢𝐧𝐠 𝐚 𝐫𝐞𝐩𝐫𝐨𝐝𝐮𝐜𝐢𝐛𝐢𝐥𝐢𝐭𝐲 𝐜𝐫𝐢𝐬𝐢𝐬 𝐚𝐧𝐝 𝐚𝐧 𝐨𝐯𝐞𝐫𝐨𝐩𝐭𝐢𝐦𝐢𝐬𝐭𝐢𝐜 𝐚𝐬𝐬𝐞𝐬𝐬𝐦𝐞𝐧𝐭 𝐨𝐟 𝐫𝐞𝐬𝐮𝐥𝐭𝐬? 🤔 As in many scientific fields, there’s increasing concern about the reproducibility of results in #MachineLearning (#ML) and ML-based science. A recent study by Nick McGreivy and Ammar Hakim titled 𝘞𝘦𝘢𝘬 𝘣𝘢𝘴𝘦𝘭𝘪𝘯𝘦𝘴 𝘢𝘯𝘥 𝘳𝘦𝘱𝘰𝘳𝘵𝘪𝘯𝘨 𝘣𝘪𝘢𝘴𝘦𝘴 𝘭𝘦𝘢𝘥 𝘵𝘰 𝘰𝘷𝘦𝘳𝘰𝘱𝘵𝘪𝘮𝘪𝘴𝘮 𝘪𝘯 𝘮𝘢𝘤𝘩𝘪𝘯𝘦 𝘭𝘦𝘢𝘳𝘯𝘪𝘯𝘨 𝘧𝘰𝘳 𝘧𝘭𝘶𝘪𝘥-𝘳𝘦𝘭𝘢𝘵𝘦𝘥 𝘱𝘢𝘳𝘵𝘪𝘢𝘭 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵𝘪𝘢𝘭 𝘦𝘲𝘶𝘢𝘵𝘪𝘰𝘯𝘴 sheds light on this issue: https://lnkd.in/e4nZ_fru 📝 After reviewing over 70 papers, the authors caution that current scientific literature may not reliably assess the success of ML in solving partial differential equations (PDEs). Key Issues Identified: 1️⃣ 𝐖𝐞𝐚𝐤 𝐁𝐚𝐬𝐞𝐥𝐢𝐧𝐞𝐬: 🚩 Accuracy vs. Efficiency: Standard numerical methods often balance accuracy and computational efficiency. However, some studies compare highly accurate, traditional solvers with less accurate ML-based solvers. To ensure fair comparisons, it’s essential to match methods on either equal accuracy or equal runtime. ⚖️ 🚩 Inadequate Benchmarks: Some comparisons are made against outdated or inefficient numerical methods, making ML look better than it might be. Comparisons should instead involve state-of-the-art methods, although this requires significant expertise. 🎯 2️⃣ 𝐑𝐞𝐩𝐨𝐫𝐭𝐢𝐧𝐠 𝐁𝐢𝐚𝐬𝐞𝐬: 🚩 Reporting Biases: The analysis, reporting, or interpretation of research findings seems often to be influenced by the nature and direction of the results. 🚩 Publication and Outcome Reporting Biases: The authors found evidence of biases where negative or null results are underreported, creating an overly positive view of ML’s effectiveness in solving PDEs. 📉 🎯 𝐂𝐨𝐧𝐜𝐥𝐮𝐬𝐢𝐨𝐧: In summary, while #ML shows great promise in engineering, particularly for solving complex PDEs, the field must address reproducibility issues and avoid overoptimistic assessments to ensure genuine progress. 🚀 A great example being Cost vs. Accuracy plots which can provide a clearer picture of an algorithm’s performance. 📊 (more in a next post). Last but not least, as the authors point out, this will not be achieved without cultural changes, including the Computational Science and engineering (#CSE) and #NumericalAnalysis community providing more benchmarking cases. 𝑳𝒊𝒎𝒊𝒕𝒂𝒕𝒊𝒐𝒏𝒔: As pointed out by the authors, the study mainly focuses on forward computational fluid dynamics (CFD) problems, and while it’s evidence-based, it’s not conclusive—some uncertainties remain.

Excited that our methods piece on the state of hypothesis-testing research in International Business (IB) has just been published in the Journal of International Business Studies (JIBS). In 2017, the journal published an influential editorial that introduced guidelines to promote more transparency and rigor in the conducting and reporting of statistical analyses. The guidelines aimed to reduce common problems such as p-hacking (manipulating analyses to obtain significant results) and publication bias (favoring positive findings over null or negative ones). 🤔 My colleagues Jelena Cerar, Phillip C. Nell and I wondered how much the field has progressed since the guidelines were introduced. So we analyzed nearly 800 empirical articles published between 2012 and 2024 in two leading IB journals. The verdict? Our results show rather mixed progress. ➕ On the positive side, researchers are more transparent than before: the reporting of effect sizes, confidence intervals, and robustness checks has increased over time. ➖ However, this improvement is uneven and has slowed in recent years. Many studies still omit important details, such as how outliers are handled or the full results of robustness tests ➖ A more disconcerting finding is that there is little to no improvement in addressing deeper issues: Most published results continue to confirm researchers’ hypotheses, and there is ongoing evidence of p-hacking and publication bias. In some cases, these issues may even be getting worse, which suggests that better reporting alone has not solved underlying incentives that discourage the publication of null or negative findings. ✔️ Based on our results, we propose a three-pronged framework and practical steps for further improving research practices in IB, aiming to (1) improve transparency in reporting, (2) strengthen the evidence behind results, and—most critically—(3) encourage the reporting and investigation of null and negative findings. Here is the link to the Open Access article: https://lnkd.in/eHqpGYbJ The article also forms part of the JIBS collection on Quantitative Methodology in IB aimed at advancing quantitative methodology to improve theory and practice (see link in the comments 👇 ) A special thanks to Jelena for leading the effort! #InternationalBusiness #Rigor #Transparency #ResearchImpact

🧪 Breaking the Bias: Publishing Null or Negative Results in Science Negative or inconclusive results (Null) often struggle to find space in scientific literature, yet they are essential for: 📉 Preventing the repetition of failed experiments. 💡 Highlighting gaps in hypotheses. 🔄 Promoting transparency in research processes. 🌟 Why It Matters Studies like Natalie Pilakouta's work on fish behavior challenge assumptions and reshape understanding, even without groundbreaking discoveries. Yet, barriers remain: 🖋️ Journals prioritize positive results over null findings. 📚 Limited platforms for sharing “non-significant” studies hinder scientific progress. 🧠 How can we improve? Let’s normalize publishing null results to: Encourage honest science. Save resources for future researchers. Strengthen trust in research integrity. How do we incentivize it? ➡️ Should journals adopt dedicated sections for null results? Let me know your thoughts! 📖 Source: Max Kozlov, Nature (July 2024).

The most important data is often the data we don't publish! Scientific progress relies on the dissemination of knowledge so that others can build upon it. As Newton said, we see a little further by standing on the shoulders of giants. But we all know that not all data is published. Negative data is brushed under the carpet resulting in a publication bias. This is not a new phenomenon. It's been raised as a problem for decades, maybe centuries. We've almost all wasted time and resources on projects doomed for failure unaware of previous failed attempts by others - sometimes even from within the same organisation! The failure to publish negative data distorts the literature, biases our efforts and probably leads to clinical failures. Moreover, we are now training AI models on datasets missing these critical null findings. It turns out there is a dedicated group trying to address this challenge and they've just published a road map following their Preventing Publication Bias Workshop last year. I'll leave a link in the first comment. There is no easy solution though. Publication bias is deeply routed in scientific culture. Negative data is perceived as less valuable leading to harsher reviews and lower citations. A recent analysis found that 180 out of 215 neuroscience journals do not accept null studies! Open research platforms have tried to shift this balance. BioRxiv has a dedicated contradictory results section but these platforms still have a major underrepresentation of null findings. Industry is no better than academia in this respect. In pharma I worked on several projects that were dead long before I joined the project team yet carried on for years. Drug development projects are like a run away train, once they get going it takes a mountain of negative data to stop them. We had clear toxicity problems on one project, which we believed was fundamental to the target antigen. We were running a very novel assay on our early stage project but the company had a later stage project to the same target about to go into the clinic. My team was ignored, warned not to interfere and threats were made. We stepped back before anyone got sacked. 9 months later two competitors that were far more advanced stopped their clinical trials due to tox. It took another 3-6 months before our projects were halted and another few years for it to be published but nowhere close to the full story. Negative data should be seen as a key learning and rewarded as equally as positive data but in my experience it's rarely the case. This new road map might be the start but it's a long journey to get academics, industry, publishers, funders and good old reviewer number 2 on board with the idea! ----- I'm Ian, I post about antibody engineering, recombinant proteins and my journey to bootstrap Gamma Proteins into a leading supplier of Fc receptors. If you like my content please reshare with your network and follow me to see more.

Responding to Reviewer 2 (or how to keep cool and impress the editor). Everyone has received a review package, where an editor invites a revision ... BUT ... one reviewer has a dark soul and issued punishing comments. Given the ultimate decision will hinge on convincing the editor, while addressing those comments, who do you do it? Impress the editor? And not inspire even more punishing comments? Here are a few suggestions. 1. Take a Step Back Before Responding Read the review objectively—try to find valid points amid the negativity. Avoid an immediate emotional response; let it sit for a day. Focus on what will help improve your paper, not the review's tone. 2. Acknowledge & Appreciate (Even If It Hurts) Start with a neutral, professional tone: "We appreciate the reviewer’s time and effort in providing feedback. Their insights have helped us refine our work." Even if the reviewer is unfair, acknowledging their effort sets a constructive tone. 3. Address Criticism Without Being Defensive Instead of: [-] "The reviewer misunderstood our point." Try: [+] "We appreciate the reviewer’s perspective. To clarify, in Section X, we address this by..." Instead of: [-] "The reviewer is wrong about X." Try: [+] "We acknowledge the reviewer’s concern. To address this, we have now included additional explanations in Section Y." If the comment is unfair, stick to facts and evidence. Remember: NO EMOTION! 4. Push Back Politely If the reviewer is factually incorrect, you can push back without being combative: "We understand the reviewer’s concern but believe there may be a misunderstanding. The relevant literature (Author, Year) supports our approach, as we clarify in Section Z." Offer a small concession if possible: "We recognize that our explanation may not have been clear enough. We have now expanded Section Y to better articulate our argument." 5. Keep Your Responses Clear & Concise Avoid long-winded justifications—get to the point quickly - offer a simple measured response: "We respectfully disagree and believe our revision in Section X clarifies this issue." 6. Show That You Took Action Mention specific improvements you made in response to the review: "Following the reviewer’s suggestion, we have revised Section X to include a more detailed discussion of..." If you disagree with a suggestion, explain why: "While we understand the reviewer’s concern, altering this aspect would conflict with established research in [field]. Instead, we have clarified our reasoning in Section Y." 7. End on a Positive & Polite Note Close with gratitude: "We appreciate the reviewer’s rigorous feedback, which has helped us improve our paper. We hope our revisions address their concerns." This shows the editor that you are receptive to advice —even when dealing with #Reviewer2. Final Thought: A measured, professional response helps demonstrate that you can handle criticism constructively - which means you might get the last laugh - an accept!

A snippet of collaborative editing from last week: “Maybe remove this sentence because it prompts this question in the reader’s mind and leaves them hanging since you don’t answer it until much further.” My client hesitated. Then he said: “I actually want them to have that question in mind. I want to keep it and, if possible, move in the other direction—like reinforce it.” That was the breakthrough, and it changed my feedback. Now instead of removing the sentence, I suggested signposting. Acknowledge the reader's question rather than leaving them wondering whether you've forgotten about it. That way, you build anticipation while reassuring them that the answer is coming at exactly the right moment. You transform uncertainty into anticipation by saying, "I hear you, and I've got you." Authors, always speak up to your editor. Question edits. If your gut tells you something isn't sitting right, open that conversation up. A good editor will never get defensive. Only curious. Good editing isn't about editors proving their suggestion is right. It's about understanding what the author is trying to achieve and then uncovering the strongest way to get there together. Good editing is collaboration, not combat.

A trainee in my lab spent a year on a study and came to our meeting visibly discouraged. The dramatic result he started with had dissolved into a careful negative finding. I told him the truth. That is not failure. That is what good science looks like when you stop cutting corners. We lose physician-scientists not because they lack talent but because the system rewards speed over rigor. A clean positive result published fast looks better on paper than a year spent tightening a design until the inflated effect disappears. The incentives push the wrong way, and trainees feel it. What protects them is not a slogan about resilience. It is structure. Protected time so they can run every sensitivity analysis a reviewer asks for. A senior person who sits with the data and explains why the early effect was the bias, not the biology. A lab culture where you are praised, not penalized, for refusing to publish something you do not believe. The basic science world is even less forgiving. One mistake costs tens of thousands of dollars. In database work the cost is mostly time, but the lesson is the same. The connective tissue between protected time, mentorship, and honest standards is what holds a young scientist through the years when the work is hardest and the wins are quiet. That connective tissue is what we are trying to build, one trainee and one careful negative result at a time.

Some freelancers just get it. They follow instructions the first time, consistently meet deadlines, and deliver clean, compelling copy across categories, from commerce and culinary to travel, design, and the outdoors. I work with these pros again and again because they make my life (and job) easier. Whether I’m editing for a major commerce outlet or a buzzy travel startup, they show up, adapt, and knock it out of the park. To the freelancers who are reliable, versatile, and low-drama: I see you. I appreciate you. I’ll always advocate for you. Tips for Freelancers Who Want to Get Rehired 1. Nail the assignment the first time Read the brief thoroughly. Ask smart clarifying questions if needed, but not ones already answered in pitching guidelines and freelancer/style guides. 2. Respect deadlines (early > late) An editor’s dream? A freelancer who files early, or at least on time without reminders. 3. Stay versatile The more you can write—SEO roundups, narrative features, product copy, interviews—the more opportunities you’ll land. (However, it's also beneficial to be an expert and niche down in certain categories, but that's another post.) 4. Be communicative but concise Keep editors in the loop if something’s off track, but respect their time. Solution-oriented updates go a long way. 5. Make your work easy to work with Clear file names, correct formatting, proper sourcing—all of it matters. 6. Be someone editors want to work with again Professionalism + consistency = long-term work and referrals.

If your results aren’t significant, they don’t count. If your findings don’t fit the narrative, they won’t get published. If your p-value isn’t below 0.05, it’s not worth a journal’s time. This isn’t science. This is gatekeeping. The replication crisis didn’t happen in a vacuum. Psychology has spent decades cherry-picking findings that confirm what we already believe, discarding studies that don’t. Non-significant results don’t get published. Replication attempts that fail get ignored. Research that challenges established theories gets buried under layers of peer review bias. Null results are still results. But the system doesn’t care. Journals reward significance. They reward novelty. They reward what looks good in print. This is why entire fields are built on shaky ground. Take the Stanford Prison Experiment. It shaped how we think about authority and abuse, but it was never a real experiment. Zimbardo coached participants to play roles, manipulated conditions, and still, decades later, textbooks present it as fact (Le Texier, 2019). P-hacking is rampant because the system forces it. When journals won’t accept null results, researchers tweak analyses, run extra tests, or selectively report findings just to meet statistical significance (Simmons, Nelson, & Simonsohn, 2011). Not because they’re fraudulent, but because their careers depend on it. The file drawer problem ensures entire areas of research disappear. For every published study, countless others sit in archives because they didn’t produce the “right” results (Rosenthal, 1979). Psychology and psychiatry publish more positive results than almost any other field (Fanelli, 2010). Not because the research is stronger, but because everything else is thrown out. This isn’t just an academic issue. It distorts real-world psychology, mental health policies, and clinical practices. When only significant results are published, when contradictory findings are silenced, when research is curated instead of conducted, we don’t just lose accuracy. We lose truth. Psychology has a replication crisis because it has a publication bias crisis. Research isn’t failing. It’s being manipulated into success.