A scientific procedure is a step-by-step written account of exactly what you did (or plan to do) in an experiment, detailed enough that someone else could repeat it and get the same results. That standard, reproducibility, is the single most important test of whether your procedure is well written. If a reader has to guess at a temperature, a timing, or which hand to use the pipette with, the procedure fails. Here’s how to write one that doesn’t.
Start With a Clear Goal Statement
Before listing any steps, state what the procedure is designed to accomplish. This isn’t an abstract or a hypothesis. It’s a single sentence that tells the reader the purpose of the protocol: “This procedure describes how to extract DNA from cheek cells using a salt-based lysis method,” for example. A good goal statement lets someone scanning a dozen procedures quickly decide whether this is the one they need.
List Every Material Before the First Step
Your materials list is a shopping list with specifications. Every item needs enough detail that someone could order the exact same thing. For chemicals and reagents, include the concentration or purity, the quantity needed for a single run of the experiment, and where to get it (manufacturer and catalog number when possible). “Sodium chloride” is not enough. “Sodium chloride, ACS reagent grade, ≥99.0% purity, 5 g” is.
For equipment, note the model or key specifications. “Centrifuge” tells the reader almost nothing. “Centrifuge capable of 10,000 × g” tells them whether the one on their bench will work. Include glassware sizes, filter pore sizes, and any consumables like pipette tips or sample tubes. If you prepared a stock solution in advance, describe how you made it or reference the preparation protocol.
Write Steps in Strict Chronological Order
Number every step. Each step should describe one action. When a single step tries to cover two actions (“Add the buffer and centrifuge for 5 minutes”), readers lose track of where they are, especially when they’re looking back and forth between the bench and the page. Split it: step 4, add 500 µL of lysis buffer to the sample tube. Step 5, centrifuge at 3,000 × g for 5 minutes at room temperature.
Every action that involves a measurement needs three things: the amount, the unit, and the conditions. “Heat the solution” is vague. “Heat the solution to 95 °C on a hot plate for 10 minutes” can be reproduced. Time, temperature, speed, volume, mass, pH, wavelength: if you measured it or controlled it during the experiment, it belongs in the procedure. Use SI units (grams, liters, meters, seconds, degrees Celsius) consistently throughout.
When a step involves waiting, say so explicitly and state how long. “Allow to cool” is ambiguous. “Allow to cool to room temperature (approximately 20 minutes)” sets expectations. If there’s a visual or measurable cue that signals when to move on, such as a color change or a pellet forming, describe what the reader should look for.
Include Safety Information Where It Matters
Safety warnings belong at the point of use, not buried in a separate appendix the reader might skip. If step 7 involves concentrated hydrochloric acid, that step should note the hazard (corrosive, produces harmful fumes), the required protective equipment (chemical fume hood, nitrile gloves, splash goggles), and the exposure route to watch for (skin contact, inhalation). Laboratory safety standards require that procedures involving hazardous chemicals describe the specific engineering controls needed, whether that’s a fume hood, a glove box, or a biological safety cabinet.
For any chemical that poses a serious risk, include a brief note on what to do if something goes wrong: the spill procedure, the first-aid response, or both. Chemical incompatibilities matter too. If your procedure uses an oxidizer in one step and a flammable solvent in another, note the storage and handling separation required between them.
Choose the Right Voice and Tense
There’s a longstanding assumption that scientific writing must be passive (“The solution was heated to 95 °C”). That convention is shifting. The American Chemical Society notes that more editors now encourage active voice because it’s more direct, more concise, and less ambiguous. Journals like Science and Nature explicitly ask authors to write in active voice when suitable. “We heated the solution to 95 °C” is clearer than “The solution was heated to 95 °C,” which leaves the reader wondering: by whom? By what?
That said, passive voice still has a place. It works well when the doer genuinely doesn’t matter and you want to keep the reader’s attention on the method itself. “Samples were stored at −20 °C until analysis” focuses on the storage condition, which is the important part. The best approach is to default to active voice and switch to passive only when it makes a sentence clearer. Write for readability first, and the voice will sort itself out.
For procedures written as standalone protocols (rather than the methods section of a paper), imperative voice often works best. “Add 5 mL of buffer” is the clearest form for someone following instructions at the bench. It reads like a recipe, which is essentially what a bench protocol is.
Add Visual Aids for Complex Steps
Some procedures are hard to follow as text alone. If your protocol involves branching decisions (if the sample is cloudy, do X; if clear, do Y), a flowchart communicates that faster than nested paragraphs. Research in science education confirms that flowcharts help people remember sequences of events and recall interactions in complex processes more reliably than written descriptions alone.
Arrange flowcharts so steps progress from top to bottom, matching the natural reading direction. For physical setups, like assembling a distillation apparatus or positioning electrodes, a labeled diagram saves hundreds of words of description and reduces the chance of error. Photographs of what a correctly prepared sample should look like at key stages can also prevent mistakes that no amount of text would catch.
Specify Your Controls and Variables
A procedure without controls is just a recipe. State what your positive and negative controls are and where in the sequence they’re processed. If you’re running a negative control alongside your experimental samples, it should appear in the numbered steps at the point where it diverges from or parallels the experimental treatment.
Identify your independent variable (what you’re deliberately changing), your dependent variable (what you’re measuring), and any variables you’re holding constant. This information sometimes lives in the introduction or experimental design section rather than the step-by-step instructions, but it needs to be documented somewhere in the procedure document. Without it, a reader can replicate your actions but not your logic.
Document for Reproducibility
Reproducibility is the standard that separates a useful procedure from a personal reminder. Nature Portfolio journals require that authors make protocols available to readers in enough detail to replicate the findings. When editors can’t resolve complaints about missing procedural details, they may publish a formal correction stating that readers were unable to obtain the information needed to reproduce the work. That’s a reputational consequence worth avoiding.
To hit the reproducibility bar, pressure-test your draft with a simple question at every step: could someone in a different lab, with no contact with me, do this correctly? That means specifying the brand of kit you used (not just “a commercial DNA extraction kit”), the software version you ran your analysis in, and the exact settings on any instrument. If you made a judgment call during the experiment, like discarding samples that looked contaminated, describe the criteria you used to make that call.
Version Control and Record Keeping
Procedures change over time. If you optimized a centrifuge speed or switched suppliers for a reagent, the document needs to reflect that without erasing the history. In electronic lab notebooks, corrections are tracked by an audit trail, so every change is timestamped and attributed to a specific person. The NIH recommends that uploaded protocol files include specific metadata: filename, title, description, creator, and creation date.
If you’re working with paper, the convention is to cross out errors with a single line (so the original text remains readable), then sign and date the correction. Never tear out pages or use correction fluid. For formal standard operating procedures in regulated labs, Good Laboratory Practice guidelines from the OECD require documented SOPs with clear management responsibilities and archived raw data. Even if you’re not in a regulated environment, adopting version control habits early saves confusion later when collaborators ask which version of your protocol produced a specific dataset.
Common Mistakes That Break Procedures
- Omitting preparation steps. If a buffer needs to be made fresh the day of the experiment, or a piece of equipment needs a 30-minute warm-up, that belongs in the procedure, not in your head.
- Using vague quantities. “A small amount,” “several drops,” and “briefly” are meaningless to another researcher. Replace them with numbers.
- Assuming shared knowledge. What’s obvious in your lab (which centrifuge, which incubator, which bench) is invisible to an outside reader. Write for the stranger.
- Burying critical details in parentheses or footnotes. If a detail matters enough to include, it belongs in the main text of the step.
- Skipping waste disposal. Every procedure that generates chemical or biological waste should state how to dispose of it. This is both a safety requirement and a practical courtesy to the person who follows your protocol next week.