From research to manuscript From research to manuscript From research to manuscript From research to manuscript
Tools and Techniques
A Stereotyped Format
Research papers are the repositories of scientific observations plus the recipes used to make those observations.
Scientific papers have a stereotyped format:
Though exact section headings vary, most scientific papers share a remarkably uniform structure that minimizes novelty in narrative and keeps the focus on the content The framework remains unchanging so the material can be studied without distraction, and the predictable form—with a standard set of sections arranged in a stereotyped order—ensures readers know what to expect and where to locate specific information, enhancing research readability and navigation.
Precise Language
Within this stereotyped format, the language of a scientific paper aims to be clean, clear, and unemotional.
Much of the color of our everyday language derives from ill-defined, emotion- ally charged, ear-tickling images conjured up by sensuous words such as ‘slovenly,’
‘sibilant,’ and ‘sneaky.’ Science, however, avoids colorful words.
Clarity is the defining trait of scientific writing; vague phrases and slippery language create confusion, leaving no room for ambiguity or arcane jargon in the scientific record In science, descriptions must be precise, methods must be complete, data must be exact, logic must be transparent, and conclusions must be stated clearly.
THE STANDARDS OF A SCIENTIFIC PAPER
M.J Katz, From Research to Manuscript, 3 © Springer Science + Business Media B.V 2009
A Single, Clear Direction
Beyond a stereotyped format and transparent language, a scientific paper also needs clarity of direction Your entire paper should point inexorably toward its Conclusion.
Guide your reader along a clear path by removing tangents and digressions Keep a single theme at the fore, with every section supporting it If your conclusion centers on temperature, make temperature the throughline of the paper so it remains ever-present, reinforcing coherence and enhancing SEO by aligning keywords with the central topic.
Temperature in Focus: From Historical Descriptions to Precise Measurement and Its Role in Experimental Outcomes, this piece traces how predecessors described temperature as a key driver of reaction kinetics, enzyme activity, and physiological processes, often relying on coarse readings that left substantial variability and spurred the shift to precise, traceable instrumentation; in Materials and Methods we detail the instruments and operations used to quantify temperature, employing calibrated platinum resistance thermometers (PRTs) and Type K thermocouples connected to a data-logging system with sensors traceable to NIST, placing samples and environments in a programmable climate chamber, performing routine calibration and pre‑equilibration, ensuring insulation to minimize gradients, and recording temperatures at 1‑minute intervals over 24 hours; the Results section reports temperature data with ambient readings ranging from 18.5°C to 26.2°C (mean 22.8°C, SD 2.1°C), sample temperatures closely tracking ambient with peak deviations of about 0.5°C, and 95% of readings within ±0.8°C of the target; the Discussion then connects these measurements to existing scientific literature on temperature regulation and its impact on process rates and biological outcomes, emphasizing how instrument choice, calibration, and chamber design influence observed stability and reinforcing the importance of transparent temperature reporting in future studies.
Reviewed and Made Available to Others
Make your scientific paper accessible to other researchers by submitting it to a peer‑reviewed journal Peer review by experts ensures clarity, credibility, and usability for a broad scientific community Publishing in a public forum preserves the paper for long‑term access, enabling the scientific community to use and build on the work.
Scientifi c Text Needs Exactness and Clarity
In science, the goal is to write papers that are easy to understand, and the art of scientific writing is not about a subtle, underlying message in your prose; instead, scientific writing is judged by how clearly it defines the details of the observations and data you have gathered, ensuring readability and reproducibility Unlike a short story, where readers may marvel at sensual writing and hints of space and time, a scientific paper should make the prose disappear so readers can marvel at a realistic, explicit, and cleanly etched picture that the evidence supports.
Scientific papers follow a standardized format to minimize distractions and emphasize content, and scientific prose should be clear, plain, and formulaic; ultimately, the message matters more than the medium, so in writing a research paper you should make your message precise and keep the presentation unobtrusive.
Writing precisely means removing adornment and focusing on clarity It takes practice to spot fluff and imprecision in your own writing, so train yourself to identify vagueness, emotion, indirectness, and redundancy and to eliminate them For practical examples of simplifying wordy phrases, see Appendix B.
It helps to remember that your goal is to speak plainly, i.e., to write clean straightforward sentences without hedging Say what you mean directly For example:
“It may therefore not be unexpected that …” should be “These results suggest
“An effort was made to …” should be “We tried to …”
“The sorbitol probably acts to increase …” should be “The sorbitol probably
“This gene is of significant interest for understanding commonalities in the
Clarifying the evolutionary history of the microorganisms A and B is clearer, simpler, and more informative when you state exactly what you have in mind For example, "A single mutation in this gene of microorganism A has brought about its new use in microorganism B." This precise articulation directly links a genetic change to a functional outcome, reducing ambiguity and improving readability By naming the specific mutation and its effect, the narrative becomes more engaging for readers and more SEO-friendly, emphasizing the mechanism of cross-species adaptation and the evolutionary relationship between the two microorganisms.
M.J Katz, From Research to Manuscript, 5 © Springer Science + Business Media B.V 2009
“It is our considered opinion that other authorities may have misstated the relative
To address the import of particulate concatenations in the soluble phase of the paradigm, it helps to be specific: In their 1994 paper, Drs Williams and Wilkins say that the drug’s failures are due entirely to the clumping of suspended drug particles In contrast, we propose that the viscosity of the solvent causes 40–50% of the failures This reframing places solvent rheology at the center of drug formulation challenges, linking solubility, suspension stability, and bioavailability to the viscosity of the solvent By prioritizing viscosity control, formulation strategies, such as adjusting solvent composition, improving particle dispersion, and enhancing suspension stability, can dramatically reduce pharmaceutical delivery failures and improve overall efficacy.
Numbers have ideal properties for scientific writing: they are precise, objective, unambiguous, and free from emotional bias They enable rigorous quantitative analysis by providing measurable, reproducible data that supports clear conclusions Moreover, numerical values describe many real-world phenomena; in a variety of ways, numbers can represent shapes and sizes, along with other quantitative attributes, facilitating accurate comparisons and effective communication of findings.
Because quantifiable adjectives are ideal descriptors in science, try to redefine all your adjectives as numbers ‘Tall’ should be defined numerically, for example,
‘greater than 2 m’ or ‘greater than 7 km.’ Likewise, ‘heavy’ should be ‘greater than 10 kg’ or ‘greater than 100 kg’ or, perhaps, ‘greater than 10 5 kg.’ If you use
‘brief,’ tell us whether it means less than a minute, less than a second, or less than a millisecond.
Even the inherently subjective adjective ‘painful’ should be set as a number on a scale quantifying how painful, as is done in most hospitals:
Pain Score worst none mild moderate severe
Writing cannot rely on numbers alone; when quantifiable language isn’t available, use precise and objective vocabulary The clarity of any sentence hinges on the reader’s ability to define every component, so scientific writing demands well-defined terms For example, “The needle vibrated continuously” is acceptable in a scientific context only if you specify which needle, the type of vibration, and the duration of the vibration The core rule of science is to define all words to ensure transparency and reproducibility.
Beyond this rule, several writing habits help ensure high-quality scientific text One essential habit is to weed out or replace vague and subjective terms; for instance, remove ambiguous language, non-specific descriptors, and unsubstantiated claims, and substitute precise data, clear measurements, and quantified statements Emphasizing clarity, accuracy, and objectivity strengthens the readability of methods, results, and discussions, making the writing more suitable for academic audiences and SEO-friendly with keyword-rich phrasing such as "quantified results," "statistical significance," and "reproducible methods."
Expressions with no clear limits, such as
• a lot, fairly, long term, quite, really, short term, slightly, somewhat, sort of, very
Words of personal judgment, such as
• assuredly, beautiful, certainly, disappointing, disturbing, exquisite, fortuitous, hopefully, inconvenient, intriguing, luckily, miraculously, nice, obviously, of course, regrettable, remarkable, sadly, surely, unfortunately
Words that are only fillers, such as
• alright, basically, in a sense, indeed, in effect, in fact, in terms of, it goes with- out saying, one of the things, with regard to
Casual colorful catchwords and phrases, such as
In team discussions, 'agree to disagree' can be the bottom line when stakeholders can't see eye to eye, yet progress emerges when you apply cutting edge strategies rather than brute force methods Easier said than done, turning ideas into action is where plans often fall through the cracks, and when results are few and far between you need food for thought to realign If you witness leaps and bounds, you're on the right track, but stay okay with a touch of discipline and avoid no nonsense shortcuts If someone wants to quibble, stay calm and use seat of the pants adaptability, while watching for sketchy details and a snafu that could derail the timeline Even a tad of effort with a tidbit of insight can reveal a tip of the iceberg, signaling larger opportunities.
Effective scientific writing relies on precise vocabulary and standardized verb tenses In scientific discourse, the present tense conveys general knowledge and enduring principles, while the past tense denotes the results of experiments.
1.2.1 Present Tense Is for Generalities
Use the present tense for general knowledge statements, widely accepted state- ments, and statements for which you could cite textbook references; for example:
• Rudbeckia hirta), a member of the Aster family, is a plant native to North America.”
“Hexoses formed by digestion in the intestinal tract are absorbed through the gut
• wall and reach the various tissues through the blood circulation.”
“The term ‘nuclide’ indicates a species of atom having a specified number of
• protons and neutrons in its nucleus.”
“On a protein-rich diet, the amount of methylhistidine in the urine increases.”
1.2.2 Past Tense Is for Specific Observations
Your results—the particular observations that you made during a research study—are bits of history, so use the past tense when you report your experimental results For example:
“In photographs of Guatemalan tarantulas, we found that the number of dorsal
• stripes ranged from six to nine.”
“During his war-time expedition to Guatemala, Rawski (1943) reported finding
“Eighteen percent of the patients in our study developed a mild rash.”
“The diodes were compared at regular time points during the next 75 h.”
The Paragraph Is the Unit of Exposition
In a research paper, each paragraph should contain one main idea, and the space between paragraphs should be like taking a mental breath Picture the text as, Idea
Most readers absorb ideas in small chunks, and scientific writing works best when paragraphs stay concise You can gauge a paragraph’s absorbability by counting its sentences The ideal length for a paragraph is 3–4 sentences, with five sentences serving as the practical upper limit If you end up with six or more sentences in one paragraph, break the text into smaller, more digestible chunks to enhance clarity and readability.
Consider this paragraph about insulin.
‘To keep all the cells in the body coordinated and working toward the same meta-
Hormones are chemical messengers carried through the bloodstream, delivering the same message to every cell they meet In sugar metabolism, insulin is the key messenger Insulin is a protein produced by beta cells clustered in the pancreas When blood glucose levels rise after a meal, the beta cells release insulin into the bloodstream, signaling tissues to take up glucose and helping restore normal blood sugar levels.
“it’s time to absorb, use, and store glucose.” ’
Hormones are chemical messengers that coordinate activities across the body by signaling tissues and organs They govern essential processes such as metabolism, growth, reproduction, and stress responses, helping the body adapt to changing conditions Understanding how hormones work explains how imbalances can affect energy, mood, and overall health.Insulin is a specific hormone central to regulating blood sugar and metabolism It is produced by the pancreas and promotes glucose uptake into cells, lowering blood glucose levels after meals When insulin signaling is disrupted or insufficient, metabolic disorders such as diabetes can develop, underscoring insulin’s crucial role in energy management and health.
‘The body uses hormones to coordinate the metabolism of its many far-flung
Hormones are chemicals transported through the bloodstream that deliver messages to the cells they contact In sugar metabolism, insulin acts as one of the hormone messengers, and its message is to take up, use, and store glucose.
‘Insulin is a protein that is made in beta cells, which are clustered inside the
The pancreas acts as a critical regulator of blood glucose When blood glucose levels rise, the beta cells in the pancreas secrete insulin into the bloodstream to lower glucose levels After a meal, for example, the pancreas releases a large dose of insulin to help glucose enter cells and restore metabolic balance.
Literary writing thrives on the ebb and flow of words that convey subconscious emotion, and short, choppy paragraphs can disrupt the reader’s experience; by contrast, scientific writing pursues a clear, unemotional presentation A methodical structure—Idea #1, breathe, Idea #2, breathe—offers an effective way to organize research, express each point with precision, and guide readers through the logic of the paper.
Hormones act as long-distance messengers that coordinate the metabolism of the body's many distant cells A clear example is insulin, a protein produced by beta cells clustered in the pancreas, illustrating how a specific hormone governs glucose regulation.
In each paragraph, begin with a lead sentence that identifies the focal point, and use the remaining 2–3 sentences to expand on that point with supporting details; inside the paragraph, these sentences may give concrete examples, data, or brief anecdotes that illustrate the focal point and deepen the reader’s understanding For SEO, weave relevant keywords naturally, use clear, concise language, and maintain coherence by sticking closely to the focal point, ensuring every sentence reinforces the main idea and the paragraph flows smoothly.
Give more details about the focal point.
Remind readers that the focal point is a member of a more general class of similar
Highlight an implication of the focal point.
In our example above, the first lead sentence tells us that the focal point of the paragraph is:
The second sentence gives details of both sides of this equation:
HORMONAL MESSENGERS TRAVEL VIA THE BLOODSTREAM
Finally, the third sentence gives specific examples:
INSULIN’S MESSAGE = “TAKE UP, USE, AND STORE GLUCOSE”
To ensure a scientist can read your paragraphs without pausing, give your writing a natural flow by making each sentence lead into the next Start with a clear subject or object in the first sentence, and carry that same subject or object into the second so the idea continues smoothly By sharing its predecessor’s focus, the second sentence extends the discussion and links new ideas to what has already been established This technique of chaining sentences—where the next sentence inherits the previous one—creates coherence, guides the reader, and supports logical progression From an SEO perspective, this clear, connected structure improves readability, keeps readers engaged, and helps search engines understand the topic and its relationships.
In the opening example, the word 'hormone' serves as the object in the first sentence It is then used as the subject in the second sentence, illustrating a clear shift in focus from object to subject This transition creates coherence by linking ideas across sentences, helping readers follow the narrative with ease Such sentence-to-sentence cohesion also supports SEO-friendly writing, because a key term like hormone remains central and repeated in a natural, contextually relevant way.
• hormones to coordinate the metabolism of its many far-flung cells A hormone is a chemical that is carried in the bloodstream and that gives a message to cells it contacts.”
In sentence two, 'hormone' serves as the subject and 'message' as the object, while in sentence three, 'message' functions as the subject and 'hormone messenger' serves as the object This alternation demonstrates how terminology can shift syntactic roles within the text, clarifying the relationship between hormones and their messages in scientific writing.
A hormone is a chemical messenger carried in the bloodstream to the cells it contacts, delivering signals that regulate bodily functions In sugar metabolism, insulin is one of the key hormone messengers, and its message is to take up, use, and store glucose.
To smooth transitions between paragraphs, craft each paragraph’s lead sentence to refer back to the previous one The flow feels most natural when the lead sentence’s subject or object mirrors something from the last sentence of the preceding paragraph For example, in our illustration, insulin acts as the bridge that links the two paragraphs, creating a seamless progression for readers.
‘The body uses hormones to coordinate the metabolism of its many far-flung
Hormones are chemicals carried in the bloodstream that deliver messages to the cells they contact In sugar metabolism, insulin acts as a hormone messenger, signaling cells to take up, use, and store glucose.
Insulin is a protein produced by beta cells in the pancreas When blood glucose levels rise, these beta cells release insulin into the bloodstream to help tissues absorb glucose After meals, the pancreas releases a larger dose of insulin to manage the glucose surge and keep blood sugar in check.
From sentence to sentence and from paragraph to paragraph, the flow of your argument should be linear:
• implies point B, point B implies point C, point C implies … symbolically:
Begin to Write While You Experiment
At the outset of writing, you often do not have a clean line of reasoning that explains your data, and you may be unsure which conclusions drawn from your observations are the most useful or defensible You might not yet have identified the strongest, clearest, or most thoroughly documented aspects of your data With these uncertainties unresolved, drafting the paper in a strict linear order—from the Introduction through the main sections to the Conclusion—can be difficult, making it hard to present a coherent argument from start to finish.
A great deal of intellectual work must be done before a tightly reasoned research paper can be completed Rather than do all this mental work before you begin writing, however, you can discover the logic while you write The process of writing a research paper can be exploratory, and it can even be a part of the research project.
Writing a research paper is a vital part of the intellectual work behind your study As you draft, you organize your data, develop clear explanations, and discover how your results connect with the findings of other scientists This process also builds the logical structure and scientific context for your experiments, guiding readers through the rationale, methods, and implications of your work and strengthening the paper’s overall coherence and credibility.
Start Broadly, Work on the Details Later
In a research paper, aim to show how your data can be fitted into a pattern that is revealing, satisfying, and perhaps a bit surprising Achieving such a pattern takes time and trial-and-error as you test ideas and refine methods You must wrestle with your data and your arguments, and if you discover your paper while you write, many drafts will come and go before the form of your manuscript becomes stable and solid.
During the initial drafting of your paper, focus on structure rather than polishing every sentence Put your editing tools aside and step back from the details, viewing the manuscript from a distance—as if through the wrong end of a telescope—so you can grasp the big picture, refine your argument, and lay a solid foundation for later refinement.
M.J Katz, From Research to Manuscript, 13 © Springer Science + Business Media B.V 2009 you can see the broad sweep of your research Then, start your work by blocking out thick chunks of ideas and arranging these chunks in a simple linear order. With a string of blocks of ideas, your next step is to sculpt these rough chunks Whittle away excesses and irrelevancies Rearrange ideas, looking for simple patterns and natural connections Carve and remodel, then step back and reassess As you progress, shape finer and finer details, and find and highlight smaller and smaller interconnections Meanwhile, be satisfied working with rough, imperfect sentences Only at the very end, when the manuscript has settled into an organized linear narrative, should you polish the language.
Adopt a global-to-local writing strategy and tackle the work in layer-by-layer steps, breaking tasks into separate sessions In the first session, collect raw material—lists of ideas, notes, and facts In the next session, build logical connections by turning those items into statements In a subsequent session, organize the statements into rough paragraphs Only in the final sessions should you refine the wording to make the paper crisp, readable, and SEO-friendly.
Starting to write often means you don’t have a clear vision of your paper, but that’s normal and not something to fear Even without knowing the final shape, you can dive in fearlessly, because using a global-to-local strategy helps the overall structure emerge naturally, guiding the details and ensuring a coherent, logically organized paper.
A Magnifi ed View of the Writing Process
Illustrating global-to-local writing, this section paraphrases the opening of a classic article, “Intensifier for Bodian Staining of Tissue Sections and Cell Cultures” by Katz and Watson (1985) The study presents a chemical intensification technique for a widely used silver stain, aimed at enhancing stain quality in both tissue sections and cultured cells The intensifier improves visualization of embryonic axons and of axons growing in tissue culture, underscoring its practical value for neuroanatomical staining Later chapters will again draw on portions of this paper as representative examples.
To reveal the full writing process, I will describe each small step in detail You already perform many of these steps automatically as you write, but when they are made visible, they can be analyzed, refined, and improved This deliberate inspection of unconscious actions helps you sharpen your craft, particularly in scientific writing, where examining even the most elementary steps can enhance clarity, rigor, and readability.
Begin writing your paper one section at a time Each section of a scientific paper has a stereotyped internal structure, a skeletal outline, and these skeletons are
An axon is the long, thread-like extension of a nerve cell that carries electrical impulses from one neuron to the next Each axon is microscopically thin, and when hundreds or thousands of axons are bundled together, the resulting cable forms a nerve that serves as the primary communication pathway of the nervous system.
The skeleton of the Introduction section of a scientific paper is:
An Introduction begins by restating a general, well-accepted idea, grounding the reader in established knowledge It then moves from this known information to the specific unknown area—the scientific gap—that the paper aims to explore As a specialized historical essay, the Introduction often opens with the Currently Accepted General Statements subsection, which looks into the past to frame the context and justify the research.
My paper was to be about staining axons, and I decided to divide the initial retrospective subsection into two topics, General and Specific The beginning of my skeletal outline became:
1 Currently Accepted General Statements a General History of Axon Stains b Specific History of the Bodian Stain
Take your outline and fill the empty spaces under each heading by listing all related ideas that come to mind Don’t worry about completeness or logic at this stage, and don’t bother to write full sentences yet Gather these ideas freely, then convert them into a cohesive English paragraph that clearly conveys the core meaning while naturally integrating relevant keywords and a logical flow for better search engine optimization.
Continue brainstorming and jotting down notes for the entire outline of the section you are writing Record every idea and fact that comes to mind, and don’t stop until each heading is followed by at least three words or phrases, ensuring depth and SEO-friendly detail This ongoing ideation creates a coherent structure where core concepts and keywords are integrated naturally, helping readers follow the argument and aiding search engines in understanding the topic.
For the topic entitled “General History of Axon Stains,” I tried to think of words and phrases about the classic work that has been done on axon staining
1 Currently Accepted General Statements a General History of Axon Stains
State of the art is molecule-specific
Individual cell stains a big leap
Began with silver stains of Golgi used by Cajal
3.3 Collect Information from Outside Resources
Next, review your references—the books, articles, and notes that support your manuscript If a section relies largely on outside information, such as the Introduction or the Discussion, use a diverse range of sources, including books, scholarly articles, and reputable databases, to ensure accuracy and credibility.
When drafting sections that are largely based on your experiments, such as Materials and Methods or the Results, rely on your research records For each reference, search for relevant information and incorporate these facts under the corresponding headings of your outline, adding a brief note about the source to ensure proper citation and traceability.
After doing this, my outline looked like:
1 Currently Accepted General Statements a General History of Axon Stains
State of the art is molecule-specific
Individual cell stains a big leap
Began with silver stains of Golgi and Cajal
Ref A – cresyl violet for neuron chromatin
• neurons have strong affinity for weak silver solutions
Santiago Ramon y Cajal, 1890–1911 (Spain), summarized in “Degene- ration and Regeneration of the Nervous System” 1928
Ref B – David Bodian 1936 used silver protein and metallic copper, produced
• clean stains of nerve cells, nuclei, axons, dendrites
Ref C – Camillo Golgi, Italian, late nineteenth century, pretreatment with
• potassium dichromate, followed by silver nitrate, stains only fraction of neurons and neuroglia and blood vessels
Golgi method gives 3-D view, good for cell architecture
Silver stains = best views of individual cells until electron microscopy Ref D – Golgi, 1880, first used photographic processing techniques of Daguerre,
To complete your ideas, take the words or phrases from your lists and turn them into full sentences, adding any information necessary to make each statement precise and informative; writing complete and understandable sentences often requires careful thinking, and you may need to consult your references again to ensure details are accurate When rewriting for SEO, present clear points, integrate relevant keywords naturally, maintain a logical flow, and use concise, scannable sentences; replace bullet items with connected sentences that convey the core message, back up statements with data where possible, and verify facts This approach yields content that is both reader-friendly and search-engine friendly, helping your article rank better while preserving accuracy and usefulness.
Continue writing full sentences for the entire outline of the section on which you are working.
1 Currently Accepted General Statements a General History of Axon Stains
Useful cell stains must give reproducible results.
A good stain will be specific for components of the feature of inter-
A good stain highlights specific parts rather than the entire feature For cells, it emphasizes particular areas such as the membranes and various organelles When staining membranes or organelles, the aim is to reveal specific component molecules.
A good stain of tiny items gives a signal that is strong or that can
For neuroanatomy, a big leap in understanding the architecture of
• the nervous system was the ability to stain an entire cell with all of its fine processes.
The study of individually stained nerve cells began at the end of
• the nineteenth century with the Italian histologist Camillo Golgi and the Spanish histologist and father of neurohistology Santiago Ramon y Cajal.
Cresyl violet proved a good stain for neuron cell bodies, highlight-
Neurons have a strong affinity for the silver in weak salt solutions.
Between 1890 and 1911, Cajal meticulously documented the cellular
• architecture of the nervous system using silver stains Much of his work is summarized in the English tome, “Degeneration and Regeneration of the Nervous System,” published in 1928.
In the United States, Stephen W Ranson began a series of silver
• studies on neural histology in 1914.
In the United States, in 1936, David Bodian introduced a simple and
• reliable silver stain for axons using solutions of silver protein with metallic copper His stain produced clean staining of the nerve cell, its axon, and dendrites.
The use of silver stains for neurons was introduced in the late
Developed in the nineteenth century by the Italian histologist Camillo Golgi, the Golgi staining technique pre-treats fixed tissue with potassium dichromate and is followed by a solution of silver nitrate This idiosyncratic method stains only a fraction of neurons, neuroglia, and neural blood vessels, but the stained cells typically reveal the full three-dimensional cell architecture.
Until the invention of electron microscopy, silver stains gave the
• best views of the three-dimensional structure of individual nerve cells (Ref C Santini, 1975; Parent, 1996).
Golgi introduced his technique in 1880 and based it on Daguerre’s
1839 procedures for processing silver-based photographs.
3.5 Arrange the Sentences into Themes
You now have a list of complete statements Your next task is to organize the statements into paragraphs.
In your finished paper, each paragraph should advance a single point To build focused paragraphs, start by collecting statements that share a common subject or theme Then group related sentences together and assign each group a Temporary Theme Label (TTL) Using TTLs helps organize content clearly, keep each paragraph centered on its main idea, and improve readability and SEO by signaling topic relevance and structure to readers and search engines.
In my rough draft, I divided my list of sentences into three themes:
1 Currently Accepted General Statements a General History of Axon Stains
TTL #1 General Features of a Good Cell Stain
Useful cell stains must give reproducible results.
A good stain will be specific for components of the feature of interest, not
For cells, a good stain highlights specific parts of the cell, the membranes,
For membranes or organelles, a good stain highlights specific component
A good stain of tiny items gives a signal that is strong or that can be easily
Cresyl violet proved a good stain for neuron cell bodies, highlighting the
Neurons have a strong affinity for the silver in weak silver salt solutions.
For neuroanatomy, a big leap in understanding the architecture of the
• nervous system was the ability to stain an entire cell with all its fine processes.
Until the invention of electron microscopy, silver stains gave the best
• views of the three-dimensional structure of individual nerve cells (Santini, 1975; Parent, 1996).
TTL #3 The History of Silver Stains
The use of silver stains for neurons was introduced in the late nineteenth
• century by the Italian histologist Camillo Golgi.
His technique pre-treated the fixed tissues with potassium dichromate and
• followed with a solution of silver nitrate.
The Golgi technique was idiosyncratic, staining only a fraction of the neu-
• rons, neuroglia, and neural blood vessels However, a stained cell usually revealed its full three-dimensional cell architecture.
The study of individually stained nerve cells began at the end of the nine-
• teenth century with the Italian histologist Camillo Golgi and the Spanish histologist and father of neurohistology Santiago Ramon y Cajal.
Golgi introduced his technique in 1880 and based it on Daguerre’s 1839
• procedures for processing silver-based photographs.
Between 1890 and 1911, Santiago Ramon y Cajal meticulously docu-
• mented the cellular architecture of the nervous system using silver stains. Much of his work is summarized in the English tome, “Degeneration and
Regeneration of the Nervous System,” first published in 1928.
In the United States, Stephen W Ranson began a series of silver studies
In the United States, in 1936, David Bodian introduced a simple and reliable
• silver stain for axons using solutions of silver protein with metallic copper. His stain produced clean staining of the nerve cell, its axon, and dendrites.
3.6 Make Your Themed Lists into Rough Paragraphs
To convert each themed group into a coherent paragraph for scientific writing, begin by outlining a rough paragraph for each group Start with a concise summary sentence that states the main point, then follow with sentences that expand that point step by step, adding detail, evidence, and context This structure makes the core idea immediately clear, while the subsequent sentences provide depth and logical progression, helping readers and search engines understand the content.
Group your statements into themed clusters based on common topics, then discard the original Temporary Theme Label Re-examine each cluster to identify its elemental common denominator, and for each set ask: what is the best one-sentence summary of these statements? Write that summary as a simple Lead Sentence that will begin your paragraph, and follow it with the remaining sentences in a logical order that enhances coherence and readability.
For my paper, I turned my three groups of sentences into these three rough paragraphs:
1 Currently Accepted General Statements a General History of Axon Stains
LS #1 An ideal cell stain is detailed, reproducible, and strong.
An ideal cell stain is detailed, reproducible, and robust, enabling precise visualization of the internal structure of the object of interest At the cellular level, the stain should highlight components such as membranes and the various organelles to reveal cellular architecture At the subcellular level, it should illuminate molecules and molecular complexes within organelles, providing insight into their composition and interactions.
A reproducible stain gives the same results in different researchers’ hands A strong stain gives a signal that is easily detected macroscopically or that can be easily amplified.
Advice to Speakers of Other Languages
Scientific logic is universal and unchanged across languages If you are more comfortable writing in a language other than English, begin by drafting your paper in your native language Once your draft is complete, translate it into English yourself or have someone else do the translation to ensure accuracy, then review and refine the English version for clear, publication-ready academic writing.
To make the final translation clearer, try to follow these suggestions when first writing your manuscript in your native language.
Use simple verbs: write ‘use’ not ‘employ.’
Turn adjectives into numbers: write ‘2’ not ‘several.’
Don’t use similes or metaphors, because they do not always translate properly
Use precise, action-based descriptions when recording the behavior of the mixture Phrases such as “the mixture could not be poured” clearly indicate high viscosity, while “beads of the mixture stuck to the sides of the tube” demonstrate adhesion and surface tension Avoid vague, qualitative comparisons like “the mixture was as thick as glue” and instead report observable effects, such as whether the material pours, sticks to container walls, or forms droplets This approach yields a coherent narrative of the mixture’s properties and supports SEO by targeting specific terms like viscosity, pourability, adhesion, and drainage behavior in experimental notes.
Put only one idea into each sentence.
Ignore the sound and the rhythm of the sentence in your native language, and
• don’t try for smooth, flowing speech Simple writing is easier to translate accurately than writing that sounds good to your ear.
In each paragraph, arrange the sentences in direct logical order.
After your paper has been translated, it is important to have it edited by a
• scientist who speaks English comfortably.
When I came to the end of my polishing ability, the first paragraph of my Introduction read
1 Currently Accepted General Statements a General History of Axon Stains
Silver staining of neurons began in the nineteenth century when Camillo Golgi discovered that nerve cells have a strong affinity for silver salts By 1880 he adapted Louis Daguerre’s 1839 methods for developing silver iodide photographs so that silver would stain neurons in fixed tissues Golgi’s silver stain provides a clean, three-dimensional view of the full arborization of axons and dendrites in individual neurons Building on Golgi’s technique, Santiago Ramón y Cajal (1928) mapped the cellular architecture of a wide variety of nervous systems, and his comprehensive silver-stain studies remain the foundation of modern neuroanatomy (Santini, 1975; Parent, 1996).
Scientific papers aim for objectivity, but this is a struggle, in part because words—even a scientist’s words— come from a subjective vocabulary One person’s
Color terms are subjective, with what one person calls “hot” another might call “warm,” and what one person labels “pink” another could see as “red.” By contrast, numbers are widely understood and carry similar meanings for most people, making them especially suitable for objective writing Numbers are not only objective; they also possess non-linguistic power, enabling precise measurement and comparison For instance, numbers can be ordered, providing a clear sequence that transcends individual linguistic variation.
0 1 2 3 4 5 6 7 8 9 and ordering allows you to use the elemental comparisons greater than, equal to, and less than precisely and unambiguously.
For example, suppose that we have three sets of dots:
Do these three patterns have any natural sequence or order?
Currently, dot patterns can be arranged in any order we choose, but by adopting a rule that assigns numerical values to each dot, we can place them in a single, universally accepted sequence A widely used option is the Braille rule, which maps dot patterns to numerical values, enabling consistent ordering across different contexts.
Now our three dot patterns have the numerical values 4, 6, and 8, and a natural order for them is:
M.J Katz, From Research to Manuscript, 35 © Springer Science + Business Media B.V 2009
In addition to giving us the power of ordering, numerical statements can be embedded in an idealized abstract continuum, which is the basis for mathematical induction.
Mathematical induction allows you to slide along an endless continuum in mathematical space.
Mathematical induction is a mode of travel through the landscape of mathematical statements, beginning with a specific assertion and then moving along a continuum by infinite, minuscule, automated steps By using mathematical induction, you start at one point in that continuum and trigger a chain of iterative events, each set off by the previous one, like a line of dominoes falling in sequence.
Once you trigger the sequence, the dominoes cascade along a preordained path that extends far into the distance, falling with an inevitable momentum that keeps every destination connected to the starting point By applying induction, you can travel vast distances and make reasoned numerical predictions about places you haven’t visited and things you haven’t yet seen.
Scientific writing becomes clearer when it leverages the objectivity and broad reach of numerical data By quantifying observations, you anchor claims in evidence, and using numbers instead of vague language throughout your writing enhances precision, transparency, and credibility.
Tables
Real-world phenomena are inherently complex, and experimental data typically exhibit variation and diversity Individual numbers have clear meaning, but a large, varied dataset can be difficult to interpret Effective data analysis requires time and effort to discern usable patterns in a substantial collection of numbers, so organizing your results before presenting them in a report helps readers understand the findings clearly and supports stronger data-driven conclusions.
Begin by organizing your results in a well-constructed table, the essential first step in scientific analysis A thoughtfully designed table consolidates data and underpins everything from qualitative discussions to sophisticated statistical analyses and clear graphical presentations.
A one-dimensional table is simply a list of results In this setup, each observation is equivalent, answering the same question: “What happens when I do Experiment A?” The only distinctions among results are their numerical values, and the natural way to organize a one-dimensional data set is to list the values in numerical order.
A two-dimensional table shows two features for each result in an experiment, providing a compact and comparable view of the data The two features can be the length and weight of each output (for example, “1 cm, 2.5 g”), or the volume of each output and the time it was produced (for example, “50 mL, 12:05”) This data representation makes it easy to compare results across samples and analyze how size, mass, or production timing relate to experimental outcomes.
In a two-dimensional table, each result is a pair of numbers, such as (1, 2.5) or (16, 8)—for example, a pairing like 16 ml and 8 h after the start The natural way to organize these pairs is to list them in numerical order according to the first number of the pair, creating an ascending sequence that makes it easier to compare results.
1.1.3 Each Entry in Your Tables Should Be a Legitimate Experimental
An experimental variable refers to each measured quantity within a single experimental result One-dimensional tables report results tied to a single experimental variable, while two-dimensional tables display results across two experimental variables There is no theoretical limit to the number of variables that can be measured and reported in an experiment; however, there is a restriction on what kinds of things can be considered legitimate scientific variables.
To determine how many petals a new variety of black-eyed Susan (Rudbeckia hirta) produces, I designed a simple experiment In May I plant 500 hybrid seeds of Rudbeckia hirta By the end of August I count, by hand, the petals on each of the 500 flowers that have bloomed To ensure I count every flower, I label each flower with a unique ID, and when I record my counts I write each result as a pair of numbers—the flower’s ID and its petal count In the end I have 500 pairs of numbers, organized in a two-dimensional table that records the link between flower identity and petal production for this Rudbeckia hirta variety.
Characterizing a New Variety of Black-eyed Susan
(a) Flower identification number (b) Number of petals on that flower
Flower id number Number of petals
I report two variables in this experiment, illustrating a focused selection of experimental variables within objective scientific study While I could have reported thousands of variables without violating the principles of scientific rigor, the chosen scope keeps the analysis manageable If I had used 500 sets of 1,000 numbers, the data analysis task would amount to 500,000 numbers to examine, highlighting the scale of data handling required in such research.
However, the two particular variables I chose to record pose a challenge Scientific experiments must be repeatable, and for a scientific report to pass muster, its procedures—its experimental methods—must be explained clearly enough that a reader could follow the written instructions and obtain a similar set of results.
In my experiment, one of the reported variables is the number of petals If you read the procedure described earlier and visit my greenhouse, you should be able to generate a list of petal counts that matches the counts I observed The experimental variable, number of petals, is a measurable trait that can be counted and compared across specimens, making it a central factor in the study.
Another researcher could measure without much additional instruction
On so doing, his results would probably match mine
However, this is not the case for my other variable, ‘flower ID number’ My written explanation was,
“To be certain that I count all the flowers, I label each flower with a unique
• number, and when I record my counts, I write each result as a pair of numbers— the flower’s ID number and the flower’s petal count.”
I won't disclose the method for pairing each flower with its ID number, and the description doesn't provide enough detail for you to visit my greenhouse and map my original IDs to the exact flowers While the overall list of petal counts may be similar, those counts are unlikely to correspond to the same flower ID numbers I used Consequently, our two-dimensional data tables will probably differ.
During my experiment I recorded an experimental variable with an incomplete recipe, and part of the procedure cannot be repeated Consequently, reporting a 'flower ID number' in the data analysis is not possible, and the reportable results are limited to a string of numbers—the number of petals on each flower—and the results can only be presented in a one-dimensional table This illustrates the principle that some variables a scientist might consider recording are not acceptable experimental variables The rule for choosing a legitimate experimental variable is straightforward: it should be well-defined, measurable, and reproducible so that the data support valid analysis and conclusions.
Legitimate experimental variables are the output of repeatable, explicitly
Number of petals is a legitimate variable because it reflects data obtained by following a fully described research recipe in the experiment's protocol; Flower ID number, however, is not a legitimate variable until I provide a reproducible method that would allow other researchers to generate the same ID numbers.
What recipes would generate scientifically legitimate flower ID numbers? Some examples include:
Recipe a: flowers are numbered in the order that they bloom
Recipe b: flowers are numbered in the order of their planting dates
Recipe c: flowers are numbered in the order of their heights
Recipe d: flowers are numbered in the order of their diameters
Recipe e: flowers, which were grown evenly spaced along a North-South line,
• are numbered according to their GPS coordinates
In some experiments, using a 'flower ID number' as an experimental variable is legitimate because there is enough information for another researcher to reproduce the numbering system However, this ID is only legitimate insofar as it serves as a placeholder for a more direct and descriptive variable, such as the date of blooming or the plant's height Therefore, whenever possible, it is more informative to omit the intermediary ID numbers and report the actual variable that was measured.
1.2 Inside Tables Use Numerical Order
When you build your tables, take advantage of the intrinsic order of numbers For one-dimensional tables, list the numbers numerically If your data are, for example:
For some experiments, this list will be quite long In an experiment with only a few results, you have room to report the entire numerically ordered list On the other hand, in an experiment, such as my fictional black-eyed Susan project, with a great many results, the list of results cannot easily be packed into a concise research paper Here, you should report a summary of the list.
For a one-dimensional table, a useful summary is a list of the frequency of occurrence of each of the different numbers For example, if the results are:
• the summary table—the frequency distribution—is
For the entire hypothetical black-eyed Susan experiment, the frequency distri- bution table might be:
Number of petals Number of flowers
Counts of the number of petals on 500 hybrid flowers
The number of petals on each flower
Statistics
Mastering statistics is a skill that goes beyond simply running numbers, and even prepackaged statistical software must be used thoughtfully No matter how much experience you have with numerical analysis, always seek advice from researchers with more expertise If you are fortunate enough to have access to professional statisticians, involve them from the very beginning of your experiments to ensure proper study design, accurate data collection, and rigorous interpretation of results.
In the upcoming sections, I outline the core vocabulary and essential tools you need for statistical analyses, establishing a solid foundation for your work For readers who want to go beyond the basics, I recommend the 2007 fourth edition of Statistics by Freedman, Pisani, and Purves (W W Norton, New York) as a key reference.
When numerical data pour out of your experiment, it helps to reduce the vol- ume to a few characteristic numbers These characteristic numbers are descriptive statistics.
Descriptive statistics go beyond mere summaries of numerical data points, as they characterize the entire dataset They view the data as a single entity with its own distinctive size, shape, and texture, highlighting the overall pattern and structure of the data rather than focusing on individual values.
A statistical analysis takes the data pile, re-orders it, and offers numerical descriptions of the entire ordered set of data.
Certain descriptive statistics are especially helpful and frequently used, because they give you a direct intuitive feel for your data pile; for example:
The size is the total number of data points in the pile Size is often represented by
‘N.’ The size of my pile of flower petal data, from above, is N = 500.
The range is the distance between the smallest and the largest data values The
• range is the data pile’s full width In my flower petal data, the range is minimum
= 15 petals to maximum = 20 petals, or range = 15–20 petals.
There are three commonly used middles.
• The mean is the center of mass, the balancing point of the data The mean is the average data value For my flower petal example, mean = 17.8 petals.
• The mode is the data value that occurs most often The mode is the data pile’s maximum height In the flower petal example, mode = 12 petals.
The median is the value that splits a data set into two equal halves, with half of the data values below and half above it, effectively dividing the data pile in half; for example, in a flower petal count, the median is 12 petals If the data set has an odd number of values, the median is a whole number, while if it has an even number of values, the median is the average of the two middle numbers, which can be a fraction.
Spread describes how compact or dispersed a data set is When values vary little, the data form a tight, compact cluster When the values show greater heterogeneity and variability, the data become more dispersed and spread out.
The standard deviation is a commonly used measure of the spread, or variability, of a data set It is calculated from the deviations of each data point from the mean, by squaring those deviations, averaging them, and taking the square root A small standard deviation indicates that the data points are closely clustered around the mean, whereas a large standard deviation signals greater dispersion The standard deviation is the square root of the variance and, depending on whether you are working with a population or a sample, you divide by N or by N−1 when computing the average of the squared deviations Used in statistics and data analysis, the standard deviation helps quantify uncertainty, compare variability across data sets, and inform decision making.
Standard Deviation is illustrated by a particularly useful statistical diagram that shows the frequency of occurrence of each value laid out as a graph These graphs are called frequency distributions or histograms (see Histograms below).
A histogram illustrates birth weight data from a study The hatched bars indicate how many babies were born in each weight class, showing the distribution across weight ranges The majority of newborns—113 infants—fall in the 3,133–3,393 g (7.0–7.3 lb) weight class, making this interval the mode of the distribution.
(continued) of numbers is compact, while a large standard deviation indicates that the pile of numbers is spread out.
All statistics programs, most spreadsheet programs, and many hand-held calcu- lators can compute standard deviations.
The central 50% provides a more intuitive measure of data spread by defining the limits of the middle half of the data set When this central 50% is narrow, the data set is compact, and when it is wider, the data are more spread out.
Drawn over the histogram is a continuous line showing how an idealized, bell-shaped normal curve (a normal distribution) might be fitted to the actual experimental data.
On the idealized birth-weight curve, the standard deviation is 521 g (1.1 lb) and the mean weight is 3,263 g The bottom line marks the x-axis locations at 1, 2, 3, and 4 standard deviations away from the mean, illustrating the spread of baby weights In a normal distribution, about 68% of birth weights fall within ±1 standard deviation of the mean, and about 95% fall within ±2 standard deviations, reflecting the typical dispersion of weights around the mean.
Birthweight (g) normal distribution curve data histogram
2.1.1 Confirmatory and Exploratory Data Analysis
Some scientific projects are straightforward documentation tasks that rely on an observational method defined by a specific research protocol You run the protocol without interference until you have collected sufficient data to properly document the protocol’s results Usually, you begin with a question that can be reframed as a practical recipe, and you simply follow that recipe to obtain answers This approach emphasizes replicability, transparent reporting, and clear documentation of each step so the study can be independently verified.
“What is the average blood pressure of 65 year old male Caucasians in California in 2009?” might be rephrased as:
“What is the result of following the recipe:
Step (1) Identify a 65 year old male Caucasian in California
To identify the central 50% of a data set, first sort all data points in ascending order The lower bound of this central range is the value at which 25% of the observations are below it (the 25th percentile), and the upper bound is the value at which 25% of the observations are above it (the 75th percentile).
Birth weight data show that the central 50% of newborn weights ranged from 2,909 g (6.4 lb) to 3,617 g (8.0 lb), a width of 708 g (1.6 lb) The mean birth weight in this study was 3,263 g (7.2 lb), with half of all newborns weighing between 2,909 g (6.4 lb) and 3,617 g (8.0 lb).
Step (2) Have him sit and rest for 5 minutes
Step (3) Take the blood pressure in his right arm while he is still sitting
Step (4) Repeat for 1,000 different males.”
This documentation form of scientific work is sometimes called confirmatory data analysis, and it is the classic technique for proving or disproving a hypothesis
In confirmatory data analysis, you turn your hypothesis into a precise, well-defined protocol and execute that protocol exactly as written, with no deviations For example, if you hypothesize that the average blood pressure of 65-year-old male Caucasians in California is 135/85 mm Hg, you design a measurement protocol and then follow it strictly, collecting data until you have enough observations to properly test the strength of the hypothesis.
Hypotheses are tested through confirmatory data analysis, but testing a hypothesis is the second stage of the scientific method, while the first stage is generating the hypothesis Hypotheses can be generated by scientific research, and just as they can be tested by research, they can also be generated by research This form of scientific work is sometimes called exploratory data analysis.
Basic Guidelines
Science often communicates ideas through words and numbers, yet humans are inherently visual beings, and our brains can interpret pictures without words or equations Visual learning accelerates understanding because images convey meaning instantly, enabling quick grasp of concepts that might take longer to explain in text For example, a single picture of a plane flying over the Grand Canyon can convey scale, motion, and context at a glance, illustrating how visuals complement verbal descriptions in science communication.
M.J Katz, From Research to Manuscript, 59 © Springer Science + Business Media B.V 2009 and this picture of Mount Everest: that cannot easily be put into words or numbers.
Images bypass language and numbers by going straight to the brain’s visual centers, where they register in the topology of the brain’s cellular architecture They convey information that words or numbers cannot, making pictures a powerful way to share complex insights By including images in your content, you offer readers richer, more diverse information and boost understanding, engagement, and SEO through visual storytelling.
A wide range of visuals—photographs, diagrams, drawings, and graphs—may serve as figures in a scientific paper, and their content determines where they appear in the text Figures that portray techniques belong in the Materials and Methods section, figures that present data belong in the Results section, and figures that illustrate synthetic concepts, abstractions, theories, and models belong in the Discussion section As with tables, figures must be referenced and explained in the text, and figures should be numbered consecutively in the order of their citations in the text.
Even within the formulaic framework of a scientific paper, you can be creative with your figures as long as the figure remains informative rather than entertaining The substantive content of the figure—whether it is a diagram of a procedure, a photograph of your experimental subject, a graph of your data, or a drawing of your hypothesis—must be clear and prominently presented to communicate the study’s main findings.
As to the form of the figure, there are a few general conventions that you can use as guides:
Try to build your figures within an imaginary rectangle 1 unit high and 1.5 units
Make the information flow from left to right, and array numbers so their magni-
• tudes increase from left to right.
The frame of the figure should be unobtrusive In graphs and similar figures,
• baselines (such as zero levels) and a scale should be indicated modestly at the bottom and at the left of the diagram.
Whenever possible, position words horizontally, not vertically, and make the let-
• tering clear, unadorned, and well spaced.
Number each figure, and give it a title and an explanatory legend The legend
Provide a self-contained, slide-ready figure explanation that defines every symbol and abbreviation, states the units and scales, and explains what each axis, line, bar, or marker represents; then summarize the essence of the figure—the main relationship or trend, the comparisons across groups or conditions, and any reference lines or thresholds that guide interpretation If the figure has multiple panels, describe what each panel shows and how it supports the overall message, ensuring the captions connect logically to the study question Conclude with a brief interpretation and practical takeaway so the slide can stand alone, using clear, concise language suitable for presentation while aligning with SEO goals through terms like figure explanation, symbols and abbreviations, data visualization, and self-contained caption.
Overall, a figure should look simple, clean, and professionally produced.
Figure Legends
Each figure needs a legend, and, together, a figure and its legend should be self- explanatory, a miniature report in themselves.
Write figure legends in concise phrases—complete sentences are not required Use the first phrase as a title that summarizes the figure, followed by essential details in subsequent phrases Include definitions of symbols and abbreviations after the narrative explanation, and end the legend with a scale indication if needed.
Figure 2 An embryonic neuron and neurites in culture The fixed culture was stained using Bodian with 10 min post-staining intensification (DIC microscopy, x 300)
Example of a figure and its legend
Numerical Figures
Graphs reveal more visual detail than tables when your data are numbers Tables organize data in rows and columns, while graphs map data on a full two-dimensional plane, so use two-dimensional graphs whenever possible Two-dimensional visuals are generally best because three-dimensional graphs can make pattern recognition harder for readers.
You can make small charts and graphs using spreadsheet software, such as Microsoft’s Excel More versatile, dedicated software is also available; an excellent example is OriginLab’s Origin.
One simple type of graph is the histogram, also called a frequency-distribution diagram, which shows how many data points have each value The range of data values is laid out along the x-axis, and the numbers of data points having each value are listed along the y-axis In a histogram, the x values—the classes or measurement intervals for the values—must be divided into equally sized intervals Each bar represents the frequency of data points within a specific interval, providing a quick view of the data distribution.
In Chapter 4, I presented a set of flower petal data as a table:
Distribution of Number of Petals
Number of petals Number of flowers
Table 1 presents our new Rudbeckia hirta hybrid, which produces yellow flowers that closely resemble the wild variety In a controlled greenhouse with regulated temperature and humidity, 500 hybrid flowers were evaluated, and the results show that 44% had 18 petals while 85% exhibited 17, 18, or 19 petals Importantly, no flowers had fewer than 15 petals or more than 20, indicating a consistent petal count and uniform bloom form under these conditions.
The data can also be presented as a histogram, where each flower is a data point and the number of petals on the flower is the value of that data point.
Expanding the table into a graph converts a numerical pattern into a spatial pattern, making the relationship easier to see Although the numerical pattern in Table 1 may require some effort to interpret, Graph 1 uses visual logic to reveal the spatial pattern at a glance.
Histograms offer a simple view by illustrating a single variable from your data points When your dataset records two or more variables per observation, you can create separate histograms for each variable For example, by measuring plant height and counting petals, you could draw a first histogram of petal counts and a second histogram that shows how many plants fall into each height range.
Graph 1 presents a histogram of flower petal counts, showing that all observed flowers have between 15 and 20 petals In a sample of 500 flowers, 221 (44%) have 18 petals, and a large majority—about 83%—possess 17, 18, or 19 petals The distribution yields a mean of 18.8 petals and a median of 18 petals, with the histogram displaying an asymmetric shape.
NUMBER OF PETALS ON HYBRID RUDBECKIA FLOWERS number of flowers number of petals 0
Two histograms, one for petal number and one for plant height, present the data as if it were collected from two independent experiments, providing a clear visualization for data analysis This approach helps you view patterns more distinctly, especially when multiple key variables are recorded in the same data set within a single experiment When that happens, creating independent histograms for each variable enables you to examine the distribution and occurrence pattern of that variable in isolation, improving insight and interpretation.
When your experiments record two key variables for each observation, you can use a scatter plot to examine the pattern of their co-occurrences In a scatter plot, each observation is a data point, and its position in the plane is determined by the values of the two variables The data points are distributed across the graph, so this visualization is called a scatter plot or scatter diagram and is an effective way to visualize the relationship between the two variables.
For example, a scatter plot of the co-occurrence of petal number and plant height, in my hybrid plant example, would have 500 points, one for each flower in the study:
A scatter plot invites you to consider whether the two key variables you’ve graphed are correlated In statistics, correlation means that changes in one variable are related to changes in the other, indicating a potential relationship between them The direction and strength of this relationship can be inferred from the pattern of points: an upward trend suggests a positive correlation, a downward trend indicates a negative correlation, and no clear pattern implies little or no correlation Understanding this relationship helps with interpretation, prediction, and data-driven decisions based on your plotted data.
The question hidden underneath the search for a correlation is, “Can the same mechanism be generating both variables?” For instance, suppose that I have measured:
The number of petals on the flower of a plant
The height of the plant
Exploring the correlation between petal number and plant height can reveal whether these traits share a common underlying biological cause If analysis shows that taller plants tend to produce more petals, researchers can begin hunting for a shared mechanism or environmental factor that drives both traits Potential candidates include the amount of sunlight, water, or fertilizer, as well as the time since seed germination or proximity to electromagnetic fields, all of which could influence growth and floral development simultaneously Identifying such factors would explain co-variation between petal number and plant height and guide subsequent experiments to pinpoint the root causes.
It is always possible that two variables show no correlation, especially when the only link is that they were measured at the same time during data collection When data are gathered together, any observed relationship may reflect the measurement process rather than a genuine connection For example, counting the number of petals on a flower while recording the Dow Jones Industrial Average at the same moment is unlikely to involve variables that share the same underlying causes, so finding a meaningful correlation would be surprising.
Graph 2 presents a scatter plot examining the relationship between plant height and flower petal count in 500 hybrid Rudbeckia plants The data reveal a positive trend: taller plants tend to produce flowers with more petals Petal counts range from 15 to 20 petals, while plant heights span from 10.7 cm to 25.9 cm This pattern suggests that flower petal number increases with plant height in this Rudbeckia sample.
Number of Petals on Hybrid Plants of Different Heights
Once you have measured two variables for your data points, the scatter plot of their co-occurrences is an ideal starting point for identifying correlations This visualization helps you observe how the variables relate through the pattern of data points, signaling potential linear or nonlinear relationships Begin the search with your eyes, leveraging your innate ability to detect spatial patterns, because visual inspection often reveals trends before any formal analysis In short, a scatter plot provides the first step in understanding the correlation structure between two variables in your data.
Look at your graph, and ask yourself these questions:
Does it look like the points (or a subset of the points) in the scatter plot have
Are the points more concentrated in certain areas?
Do the points form a shape?
Do the blank areas form a pattern?
To begin pattern recognition, try naming what you see: does it resemble a basic mathematical shape—such as a line, a circle, a parabola, or a wave? Before applying any mathematical tools, put your impressions into words Commit to your notes by writing, “I see a pattern” or “I do not see a pattern,” and if you do see a pattern, fill in the blank in the sentence, “The pattern looks like _,” as unambiguously as possible, using clear phrases to describe the shape.
Dense points on the left grading to sparse points on the right
This sentence is your rough hypothesis.
Scatter plots are a fundamental tool for identifying and describing patterns in experimental data, helping you make sense of your research results The analysis is iterative: you begin with a working hypothesis—this group of data points appears to lie on a straight line—and you test, refine, or discredit it as you continue exploring the data.
Preparation for Submission to a Journal
Since journal figure requirements vary, choose your target journal before polishing the figure Print journals usually require photographs of each figure—either as physical prints or high-quality digital prints Some print journals and all online journals request electronic figure files in a specific format (such as JPEG) and at a sufficiently high resolution so they can be inset directly into a website Photographs can be easily polished and prepared for publication with software like Adobe Photoshop Elements.
Scientifi c Patterns Should Be Reproducible
Patterns in a data set can jump out at us when the data is presented visually, as in this graph:
Visual data visualization is a powerful tool for uncovering patterns in complex data, but it is a double-edged sword: the act of looking at graphs can both reveal and mislead Our brains tend to impose structure where none exists, a phenomenon known as pareidolia, causing us to see mirages—clusters, lines, curves, waves, shapes, and even faces—in data visualizations This perceptual bias is common in dense graphs and can appear even in randomly distributed data, such as scatter plots, where apparent patterns emerge despite the absence of real structure.
This observation leads to a general caution about any pattern you might discover, because the universe is vast and its patterns are effectively uncountable; nevertheless, science focuses on only a small fraction of those patterns.
Science deals in reproducible, verifiable patterns that are stable or repeatable features of the systems under examination Describing today’s pattern of raindrops on the ocean is scientifically useless, because raindrops splash in endlessly unpredictable, unique patterns Therefore, scientific descriptions rely on generalizable patterns drawn from repeated observations, acknowledging that Nature also presents a bottomless well of unique, one-time patterns that contrast with the repeatable features science seeks.
Science relies on repeatable patterns, not fleeting, one-off occurrences A pattern that appears only once is irreproducible and cannot be used to make reliable predictions Without reproducibility, we cannot build stable systems or develop reliable technologies Such unique instances offer little value for understanding, predicting, or controlling our environment.
When you identify a pattern in a graph or any data representation, you must demonstrate that this pattern is a persistent feature of your experimental system In practical terms, you should either demonstrate the same pattern when you repeat your experiments or verify it across multiple trials to establish reproducibility and reliability of your findings.
Cite independent examples of the same pattern being associated with your
Use statistical techniques to argue that this particular pattern is unlikely to have
Research winds and backtracks characterize the scientific journey: experiments often veer into unanticipated directions as irregular, peculiar data emerge, threatening to derail even the best-laid plans Technical complications demand attention, while new questions materialize, ideas evolve, and researchers experience second thoughts and doubts about their hypotheses Disorder continually intrudes on the daily life of an experimental scientist, reshaping the path of inquiry and forcing constant reevaluation.
Drafting your scientific paper while you conduct experiments helps keep day-to-day research orderly and ensures that methods, data, and interpretations stay linked Writing as you go keeps your efforts and ideas connected and clear, preventing gaps that can derail a study later This real-time documentation creates a coherent research narrative, improves traceability of procedures, and reduces revision time by capturing hypotheses, results, and insights when they are freshest By building a solid manuscript foundation during experimentation, you enhance reproducibility, strengthen the final paper’s persuasiveness, and streamline the path from data collection to publication In short, proactive scientific writing supports an efficient research workflow, better clarity, and higher overall quality.
When you put your ideas into sentences, you have to face their logic (or lack of
When you sort your data into tables, you can see the holes that remain in your
When you commit your recipes to the page, you are forced to record
When you look at a skimpy or lopsided reference list writ large in black and
• white, you are embarrassed into doing more background research.
The skeleton of a developing scientific paper serves as an effective blueprint for building a cohesive narrative from the diverse activities in real-world research This guide helps you pull together disparate elements into a single, compelling story By writing as you work, you’ll make faster progress and operate more efficiently, turning ongoing research into publishable results.
M.J Katz, From Research to Manuscript, 75 © Springer Science + Business Media B.V 2009