What Is Science Writing

The actual work behind explaining research to people who didn't study it

Science writing is the practice of translating technical research into language that non-specialists can read and act on. It covers everything from lab reports and grant proposals to magazine features and technical documentation. The core skill is not simplification. It is preservation of accuracy while removing the barriers that keep a paper locked inside its own field. I learned this the hard way when I was assigned to cover a new paper on CRISPR-based gene therapy. The authors had published in Nature Medicine. Their methods section was three pages of viral vector specifications, off-target analysis, and dosing schedules that meant nothing to a general reader. I spent four hours calling the lead author's lab. They were polite but frustrated. Every time I asked a clarifying question, they assumed I already knew the baseline terminology. I had to ask them to walk me through the concept from scratch, like I was a smart graduate student in a different department, not a journalist who had read the abstract. That took two more phone calls. The final piece ran at about 1,200 words and explained the mechanism without a single unexplained acronym. It also required me to verify every claim against the supplemental data because the press release had overstated the therapeutic window by about forty percent. That discrepancy showed up only when I checked the raw numbers myself.

What Is Science Writing

At its foundation, science writing means taking information that exists in a highly structured, jargon-heavy format and producing a version that preserves the factual content while changing the structure entirely. That is different from simplification. Simplification removes details. Science writing reorganizes them so the remaining details become legible. A reader should finish the piece with the same factual understanding they would have if they read the original paper, minus the technical scaffolding they do not need. The discipline splits into several branches, and each has different rules and different audiences.

Popular science

This is the magazine and newspaper category. The audience reads for interest, not for training. The writer must choose narrative entry points, find a human angle when one exists, and still not fabricate stakes that are not already in the data. Common failure mode: turning a statistically marginal finding into a breakthrough story because the press release said "groundbreaking." I stopped doing that after a cancer screening paper came out with a relative risk reduction of three percent. The outlet wanted "game-changing test." I wrote "modest benefit with known limitations." The editor pushed back for forty minutes. We compromised on a headline that was honest without being dramatic. This includes documentation, SOPs, white papers, and instructional materials. Accuracy is mandatory. Readability matters, but so does reproducibility. If someone cannot follow your instructions and get the same result, the document failed. Technical writers often work inside companies or universities. They use controlled vocabularies, version histories, and strict style guides. The audience may be engineers, clinicians, or regulators. You adjust for them, but you do not alter the underlying facts. This sits between research and decision-making. The goal is to give policymakers, journalists, or institutional leaders enough accurate context to make informed choices. The work is less about narrative and more about framing uncertainty correctly. Scientists often hate this part because it requires admitting what is unknown. That admission is not weakness. It is the difference between sound policy and expensive mistakes.

One thing beginners get wrong is the assumption that lay audiences cannot handle complexity. They can. They just need signposting. A paragraph that says, "Here is what we know, here is what we do not, and here is why it matters right now" is worth more than three pages of undifferentiated detail. Readers tolerate complexity when they understand the hierarchy of importance. They bounce when they cannot find the main point among the supporting evidence. Another counter-intuitive insight: the best science writers usually spend most of their time reading, not writing. A single feature on a neuroscience paper might require reviewing six to eight related studies, checking the original methodology, and confirming that the conclusions in the discussion section are actually supported by the results. If you skip that step, you will repeat a myth that the primary paper itself got wrong. I have rewritten three stories after discovering that the key finding depended on a cell line that does not behave the same way in vivo. The paper was correct for the model it used. The press release implied otherwise. That gap is where bad science writing lives. There is also a practical workflow that most people do not talk about because it is boring and unglamorous but it is what keeps the work from falling apart.

First, you get the source material. That is the paper, the dataset, the press release, or the expert interview. Second, you read the methods section before the abstract. The abstract is a sales pitch. The methods tell you what actually happened. Third, you take notes in your own words, not quotes. Quoting researchers verbatim is tempting but dangerous. People say things in interviews that sound clear and then collapse under scrutiny. Paraphrase immediately so you are working with your understanding, not their wording. Fourth, you verify the numbers against the original figures. A single percentage point shift in the supplementary table can change the entire interpretation. Fifth, you draft the piece from those notes, not from the paper itself. That forces you to internalize the content before you transmit it. Sixth, you send the draft to an independent expert who did not work on the study. Not the authors. Someone else in the field who can catch misinterpretations without the conflict of interest that comes with defending their own work. Seventh, you incorporate feedback, revise, and publish. This process usually takes four to six hours for a standard feature. A deep investigative piece on contested research can take two to three days. Time is the real constraint. Outlets want speed. Speed trades against verification. If you cannot slow down, you will make mistakes that require corrections. Corrections damage credibility faster than anything else in this field. The tools are straightforward. Zotero for reference management. Notion or a simple text file for drafting. A spreadsheet for tracking claims against source locations. A checklist for verifying numbers, names, and dates. The checklist is the unsexy part that prevents the most embarrassing errors. I keep one that asks: Is this number exact or approximate? Does the graph support the sentence? Did I confuse correlation with causation? Is the sample size stated? Did I mention the limitation that the authors themselves flagged?

There are legitimate downsides to this kind of work, and they are not minor. Science writing pays poorly compared to technical writing in industry. Feature writing at magazines often uses freelance rates that have not kept pace with inflation. The emotional load is higher than most people expect. You are repeatedly exposed to health crises, environmental disasters, and institutional failures. You see how quickly narrative gets weaponized by stakeholders who want a favorable outcome. You also see how easy it is to get fooled by a clean abstract and a well-worded press release. That disillusionment is normal. It does not mean the work is worthless. It means you need boundaries and a verification habit. If you want to start, do not begin by trying to write for major publications. Start by rewriting one press release from a university or research institute into a plain-language summary. Post it somewhere public. Get feedback from actual non-specialists. See where they get confused. Fix those spots. Repeat. That exercise alone will teach you more than any course on science communication because it shows you exactly where the gap is between what the writer thinks they said and what the reader actually understood. The field needs people who treat accuracy as a discipline, not as an afterthought. The rest follows from that baseline.