Writing Clear and Concise Research Papers

Last week, I reviewed 3 papers in a row that all had the same problem: Good data. Solid methods. No visible novelty. Not because the work wasn’t original, but because the authors assumed the originality would somehow “speak for itself”. It never does. If reviewers and editors need 20 minutes to guess what is new about your paper, they will almost always conclude: “Lack of novelty. Reject.” Here is a simple structure you can use to fix this in your next manuscript: 1. One-sentence contribution (yes, just one) If you cannot explain your contribution in one sentence, the reviewer will not do it for you. Ask yourself: “What does this paper do that no published paper has already done?” Write that sentence. Put a version of it in the abstract and in the last paragraph of the introduction. 2. Make the gap painfully clear Don’t write: “Few studies have examined X.” Write something like: What we think we know. What we don’t know (exactly what is missing, wrong, or unclear). Why this gap is a problem for the field. If the gap is vague, your contribution will look vague. 3. Name the type of novelty Most early-career researchers actually have one of these: Contextual: Testing known theory in a new context or population. Methodological: Using a new data source or technique that reveals what others could not see. Conceptual: Clarifying, extending, or slightly challenging an existing idea. Say which one you are doing and show how. 4. Use contribution language, not “what we did” language Weak: “We analyzed 500 surveys and ran regressions.” Stronger: “We show that the X–Y relationship reverses in setting Z, which existing theory does not predict. This refines how we understand X in volatile environments.” Same work. Different framing. Completely different response from reviewers. 5. Echo the novelty again in the Discussion The Discussion is not just “here are the results again”. It is where you say, clearly: What changes for the field because of your findings. Which assumptions need updating. Where the next person should pick up the conversation. If your final section could have been written before you ran the study, you are not explaining novelty. Your research can be novel, but invisible. Your job is to make the originality impossible to miss. #science #research #scientist #publishing #academia #professor #highereducation #researchservices #novelty #thesis #phd

𝟴𝟬 𝗣𝗮𝗴𝗲𝘀 𝗼𝗳 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵. 𝟮 𝗣𝗮𝗴𝗲𝘀 𝗼𝗳 𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻 In public policy, most reports are 60–80 pages long. But here’s the uncomfortable truth: Most decision-makers only read the first 2. And sometimes? Just the executive summary. As a research analyst, that realization changed everything about how I structure my work. Here’s what I’ve learned about making sure your research drives action, not just collects dust: ✅ Write for the reader, not for the writer. Don’t write to show how much you know, write to show what they need to decide. ✅ Lead with what matters. Start with the “So what?” before the “What.” Policy leaders want outcomes, not background theory. ✅ Use a “3-30-3” format. Your report should offer: → 3 seconds of clarity (title/executive summary) → 30 seconds of insight (key charts/headlines) → 3 minutes of direction (recommendations & next steps) ✅ Assume scanning, not reading. Use bolded insights, clear section headers, and takeaway boxes. They’re not cosmetic, they’re functional. ✅ One page = one message. If a page has three ideas, it has no anchor. Keep it focused. Make it memorable. 🧠 Research doesn’t create impact. Readable research does. We’re not in the business of writing reports. We’re in the business of helping people make better decisions, faster. 💬 Tag a peer who’s ever had to condense 6 weeks of work into 6 bullet points. And if you want more behind-the-scenes frameworks on how research drives real-world change, follow for more. LinkedIn LinkedIn News India #PublicPolicy #ResearchToImpact #ResearchCommunication #ExecutiveSummaries #PolicyDesign #DecisionSupport #LinkedInForAnalysts

The night before submission, 2 a.m. Your model has finally stopped exploding; the results look decent, but something still feels off. You stare at the loss curve the way you once stared at a half-sent text, wondering if you should hit “undo.” That moment—the gut-check between “works” and “trusts”—is exactly what this year’s three Outstanding Papers nailed: - Safety Alignment Should be Made More Than Just a Few Tokens Deep. - Learning Dynamics of LLM Finetuning. - AlphaEdit: Null-Space Constrained Model Editing for Language Models. They didn’t race for bigger GPUs; they cracked open the engine to ask why it stalls. They proved that the frontier problem in 2025 isn’t scale but self-diagnosis: can we align, fine-tune, and edit without erasing a model’s soul? The Honorable Mentions—Data-Shapley-in-One-Run, SAM 2, Faster Cascades—still chase speed and efficiency, and that’s awesome too. But the trophy went to researchers who treated deep nets like relationships: listen first, fix what hurts, then move fast. Want your paper on next year’s list? 1. Start with the uncomfortable question everyone sighs past. 2. Design one experiment that forces a “wait, what?!” in the review thread. 3. Offer a minimalist remedy—one regularizer, one projection, one line of code. 4. Release the repo before the applause; humility travels farther than hype. In the figure: Left (b): standard edit shifts representations (pink). Right (e): AlphaEdit keeps them in place (blue)

I recently received a rejection from a journal. It was not just a rejection but a desk rejection—the kind where your manuscript does not even make it past the editor's desk. They mentioned that the work was not particularly novel and did not quite align with the journal. They had a point. I have been working on this material for a while, and perhaps the novelty become diluted in the process. Every scientist, at some point, faces rejection. It is a rite of passage in academia. Rejections happen all the time, and how we handle them matters. It is about resilience and recalibration. So, what can we do when faced with such setbacks? One way is to look for another journal. Sure, it is tempting to be discouraged, but the truth is, there are many journals out there, each competing for the best papers in their field. What might not be a good fit for one journal could be exactly what another is looking for. The key is knowing the landscape. So, take some time to research journals that are better aligned with your study’s focus, and give it another shot. Another important step is to revisit your literature survey. Journals often reject papers when they feel the work does not clearly state its contribution. We might think we have made a novel point, but it is easy to forget that novelty must be evident, not just to us but to reviewers who might not be familiar with the nuances of your research area. A strong, well-written literature review can be the bridge that takes your work from "not novel" to "groundbreaking." On data. Is your data presented in a way that highlights the core contributions of your work? It is easy to overlook the storytelling aspect of scientific research. Every figure, table, and graph should support the narrative of innovation. If you are not making your data sing, it is easy for reviewers to skim past your most important findings. Go back and ensure the way you present your data draws attention to the novelty. What, really, is novelty? It is a question every scientist grapples with at some point. Novelty is not just about being first; it is about doing something in a new way, solving a problem differently, or opening up a new avenue of research. It is the ability to see connections others do not. If we cannot articulate that innovation in a way that is compelling, the novelty is lost on others. Frame your research in a way that speaks to the current challenges in the field. Why is your work necessary? What questions does it answer that have not been addressed before? Handling rejection is part of the scientific process. Every rejection is an opportunity to refine your work and sharpen your message. Instead of dwelling on the "no," think of how you can make your next submission stronger. Every setback is a step toward making your research even better. So, embrace the process, learn from it, and let it guide you to eventual success. Good luck to everyone and especially to myself on the next submission!

Reviewers agreed my research was rigorous. Then they rejected the paper anyway. The science wasn't the problem. It usually isn't to be honest. But the structure was. Here's what I learned after publishing 300+ papers: Many rejected papers fail for this single reason: Reviewers never make it mentally past paragraph 3. They found nothing that grabbed them. Sorry, but they couldn't find your story or weren't interested in it. I've watched this happen nonstop as an Associate Chair. Solid methodology. Meaningful results. Genuine contribution. But still not passing the bar. None of the data you collected matters if you can't hold your readers attention past the first few sentences. Think of your paper as a pile of LEGO bricks. Raw data? That's the chaotic heap on the floor. Every kid dumps the box out. Every researcher collects findings. Yet nobody ever got famous by just playing around with LEGOs. (Sadly.) But here's where papers really perish for good: Most academics stop at SORTED. Great, you got your themes colour-coded, buddy, but you still gotta build the house. The papers that get cited for decades? They build the house. Brick by brick. Thinking. SORTED → ARRANGED → PRESENTED → EXPLAINED (W/ STORY) That's the journey your reader needs. From chaos to meaning. I now structure every paper as a 5-act story: Act 1: Introduction Create a curiosity gap. Make reviewers think: "Hey, I've never thought about that." Act 2: Literature Review Set the scene. Show how everyone's been circling a problem like sharks that your work now fills. Act 3: Methods Build trust. Write like Betty Crocker. Put down a recipe another researcher can follow. Act 4: Results Deliver surprise. Lead with your most counterintuitive finding. Yes, you can report results in a meaningful sequence. Act 5: Discussion Provide meaning. Connect your data to the bigger picture. Explain the: "So what?" One test I use for every section: → Why would a smart reader keep going? If you can't answer that, rewrite the transition. I spent years treating structure as an afterthought. The science came first. The writing came last. That's backwards. Rigour and readability aren't opposites. The papers that get read for a decade usually have both. Playing with LEGO bricks is fun and all, but have you ever built a house? Save this for your next data session. One tactical system per week. 13k+ researchers. Zero fluff. → https://lnkd.in/e4HfhmrH

PhD Issues: I've read many florid and rambling PhD theses. I understand the desire to showcase your vocabulary and experiment with literary forms. However, I recommend a different approach: Simplify your writing. Present your findings in an easy-to-follow way. This isn't "dumbing down" or being simplistic. It's about communicating effectively. You need to get your message across, reach a wide audience, and make a bigger impact. It takes time to think deeply about your message and express it clearly. Your first drafts won't be perfect - expect to spend time editing and refining. 🟢 For each paragraph, ask: What point is it making? How does it advance the chapter? If a paragraph diverts from these goals, it's misplaced. Readers assume everything you tell them is important - if it isn't, either cut it or move it to footnotes. 🟢 Include clear breaks between paragraphs. Look at your page—is it a wall of text or broken into manageable pieces? 🟢 Is your chapter self-contained within the broader thesis? Avoid frequent references like "See page 587" or "refer to chapter 37" - these interrupt the reading flow. 🟢 Keep chapters to around 6,000–8,000 words. Beyond this length, readers struggle to follow your argument. Either trim unnecessary content or split into two chapters. 🟢 Begin each chapter by outlining what you'll cover and how it connects to your broader research themes. This roadmap helps readers process what follows. 🟢 Vary your sentence length for impact and rhythm. Short sentences can be powerful. Mix them with longer ones to create flow. While you're not writing prose, you are communicating - make it engaging. Complex topics deserve clear expression. Your subject may be inherently complicated - don't add unnecessary complexity to it.

I revisited a paper on how to write a research paper well, and the message was blunt: Many rejections come from avoidable mistakes. Here are a few that matter most: 1️⃣ A weak title weakens the paper before it begins If the title is: → Too vague → Too long → Trying too hard you are already losing the reader. The title should match the study clearly. 2️⃣ Most people will judge your paper by the abstract That is just reality. The abstract is what reviewers see first. It is what shows up online. It is what people find when searching. If it is careless, too long, or poorly written, the rest of the paper may never get a fair chance. 3️⃣ Your introduction should not feel like punishment The paper makes an important point: You are not only reporting a study. You are telling a scientific story. If the introduction is dull, unfocused, or missing a clear hypothesis, you lose momentum early. 4️⃣ Methods are where credibility becomes visible This section must contain enough detail for someone to repeat your study if they want to. If your methods are unclear, the reader starts questioning everything else. 5️⃣ Results are not the place to show off Results should be results. Not discussion. Not interpretation. Not decoration. The paper is clear on this too: → avoid repeating yourself → keep figures useful → use as few images as necessary 6️⃣ The discussion is where many writers become careless This is where people often overstate findings. The paper warns against assigning greater significance to results than they deserve. That one mistake can make a serious paper feel unserious. 7️⃣ Good English is not a luxury It is part of the science being understood. One of the sharpest lessons in the paper is this: If a reviewer missed your point, the problem may not be the reviewer. It may be your writing. That line should humble every researcher. Because clarity is part of the work. ⸻ Good papers are built on: → clear titles → sharp abstracts → focused introductions → reproducible methods → honest discussions → clean writing A paper does not need to sound complicated to sound academic. That is what strong scientific writing looks like. 💬 What do you think damages a paper faster: a weak abstract, weak methods, or an overconfident discussion? ——— Source: Villar R. (2020). How to write that paper. Journal of Hip Preservation Surgery. #AcademicWriting #Research #PhDLife

My students and I are working on a couple of manuscripts, and I am reminded that there is an art to writing research papers that others actually read and, most importantly, appreciate. First and foremost, the paper should tell a story. The paper should not be a brain dump or a chronological description of the experiments you have conducted. It should be a carefully crafted narrative with 1-2 major points. Ask yourself: What new story am I telling? What will readers learn that they did not know before? The answers should appear as early as possible and be clear. The title is a hook. I think of it as a newspaper headline. It should attract potential readers. The paper's title should be specific, brief, and grab attention immediately. One should be able to summarize the paper's contribution in one compelling phrase. It is important to set the stage in the introduction. Motivate your audience by clearly establishing why your work matters, the current state of knowledge, and how you are advancing it. Review prior work not as a literature dump but as context for your unique contribution. This is like writing an expository opening song to a musical where all the characters and the theme are introduced. (I think about the first song in Hamilton.) Educate without overwhelming. Anticipate what your readers may not know and may need to be reminded. This is very hard. You do not want your manuscript to have too much textbook knowledge. On the other hand, most readable papers anticipate the audience and have just enough material and references so that the manuscript is understandable. From your point of view, you want them to know enough to appreciate your work. Each paragraph needs a clear topic sentence that advances your main argument and exposes your idea. The arguments must be crisp. The key is to avoid wandering thoughts or side points that may only interest a few (perhaps only you). Support key claims with evidence or references. Invest serious time in your figures and prepare them with the right software. They should tell your story visually and help organize your narrative flow. They must be appealing, and the message should be easy to grasp. In our group, we prepare the figures and captions first to storyboard the paper. We sweat the details in my group. When I was a grad student, I got myself a copy of "Strunk & White, The Elements of Style," and I would review my papers applying the numbered rules. I would read the paper only looking where I can apply a subset (2 to 4) of the rules. I would pick another set and do it again. These days, companions like Grammarly essentially make this easier. The most important rule is to use clear, definite language with as few words as possible. Proofread again and again. Check grammar. Having clean figures, text, and references with no errors is also a credibility builder. The best papers teach, persuade, and advance our collective understanding while reporting the results of an investigation. 

On renewing my commitment to write well in the New Year. I find the end of the year a good time to take stock, read, and sharpen my thinking about what I want to do in the next year. My goal for this next year is to write fewer and better papers. This goal was stimulated by reading an old blog article in the American Scientist (https://lnkd.in/e5Z6eX2T). It details how many scholars simply aren't great writers. In a world where more is needed, I worry that my writing has degraded in quality over the past couple of years. So, I'm going to revisit my papers in progress relative to a few simple principles outlined in the blog and, going forward, keep these next to my desk as a reminder. Principle One: Scientific writing should be interpretable. Key Problem: Scientific writing is often difficult to read due to the complexity of scientific concepts. The Truth: Complexity of thought does not mandate impenetrability of expression. Poor writing reflects not just stylistic flaws but also issues in the clarity of thought. Principle Two: Understand Reader Expectations Key Problem: Readers interpret articles based on structural cues, not just content. The Truth: Writers should meet these expectations to ensure clarity & uniform interpretation. Use tables & structure to offer a more intuitive organization of ideas(context first, key information second). Similar principles apply to prose. Principle Three: Adhere to Simple Core Rhetorical Principles Key Problem: Many scholars have poor rhetorical skills. Solutions: * Subject-Verb Separation: Readers expect subjects to be followed closely by verbs. * Fix: Position significant material appropriately to minimize interruptions. * Stress Position: Place key information at the end of sentences where the emphasis is natural. * Topic Position: Begin sentences with familiar (old) information to provide context & linkage. * Locating the Action: Verbs should articulate the action to avoid ambiguity about relationships between ideas. Principle Four: Constantly Revise your work. * Simplify Structure: Avoid long, convoluted sentences with misplaced emphasis. * Maintain Logical Flow: Ensure each sentence links back to earlier ideas while guiding the reader forward. * Balance Old & New Information: Place old information in the topic position and new, emphasis-worthy information in the stress position. * Address Gaps: Identify & fill logical or conceptual gaps revealed during revisions. Principle Five: Never forget the Interplay of Writing & Science * Writing is not just a means of recording data but a tool for sharing, interpreting & refining scientific arguments. By consciously applying these principles, I aim to make it easier for readers, to minimize conceptual flaws or omissions and write better papers. And perhaps have fewer rejections! Stay tuned. I'll report back in a year! #academicwriting

[Storytelling in Academia] The core: academic papers are not mere records of what happened, but carefully constructed arguments. Before you say, “No”, let me clarify. I’m not saying we (academics) invent stories. Nope. The science must remain objective and rigorous. But the way we communicate that science is a carefully designed narrative. Some excellent students write their first draft like a documentary. They try to record everything they did because they believe “that’s science”. The result is often a long manuscript with lots of (unnecessary) details. But when I ask, “what is the one message you want readers to remember? They often have difficulty summarizing the gist. A good paper tells a clear story. That also means academic writing is "strategic", not chronological. A manuscript is not a diary of your research process or results. It’s an argument built to help readers understand your contribution. Sometimes, journal editors may recommend you swapping Experiment 1 and Experiment 2 because it creates a more logical narrative. Sometimes, reviewers will ask you to move a figure to the appendix, shorten a methods description, or remove analyses that don’t directly relate to your research question. None of these suggestions change the science. They improve the clarity of the story. Finally, like any story, you have an audience. The same research can be written differently depending on your audience. In Psychology, reporting statistical details in between sentences in the Results section is common practice. But in some HCI journals, reviewers often recommend summarizing statistical details in tables or moving them to an appendix to improve readability. Nether is wrong. They simply reflect different expectations from different audiences. Once you have research that discovers new knowledge, good writing helps others understand why that knowledge matters (and helps your manuscript get accepted, lol). #AcademicWriting #PhDLife #Storytelling #NoOperationDescription #Gist #StrategicWriting #ConsiderYourAudience #ElevatorSpeech