Science Research Paper Help IMRaD Format, Structure & Writing Guidance

A science research paper has a specific and demanding intellectual job: it presents original research you've conducted, or critically evaluates existing research, and convinces a reader examiner, supervisor, or peer reviewer that the work is valid, significant, and connected to prior findings in the field. That's a different task from simply describing what you did and what happened. "We mixed compounds X and Y and observed reaction Z" is a lab note. "We investigated whether X catalyses Z in the presence of Y under controlled temperature conditions, because prior studies showing Z under different thermal conditions leave open whether X or temperature is the primary driver" is the beginning of a scientific argument.

The convincing happens through structure as much as through content. The IMRaD format Introduction, Methods, Results, and Discussion is not bureaucratic convention. Each section does a specific rhetorical job: the Introduction explains why the question matters and what gap the research addresses; the Methods section demonstrates that the approach was rigorous enough to produce trustworthy results; the Results section presents evidence without interpretation; and the Discussion interprets that evidence in the context of the original question and the existing literature. A paper that blurs these boundaries that begins interpreting in the Results section, or that restates Results in the Discussion without adding interpretive depth is failing at the structural level regardless of how strong the underlying science is.

Natural sciences and social science's same structure, different conventions

The IMRaD structure applies across both natural sciences (physics, chemistry, biology, earth science, space science) and social sciences (psychology, sociology, economics, political science, anthropology) but the specific conventions within each section differ considerably. A Materials and Methods section in a chemistry paper specifies reagents, concentrations, reaction conditions, and instrumentation to a level of precision that would allow another chemist to replicate the experiment exactly. A Methods section in a qualitative sociology paper describes the sampling strategy, interview protocol, theoretical framework for analysis, and reflexivity considerations rather than physical procedures. Both are "methods sections" in the IMRaD sense, but they're doing different work for different epistemological traditions and the conventions for each need to be understood before either can be written well.

The introduction's job is to convince the reader that the research question is worth asking and that the paper's approach is the right one to answer it. A weak introduction either starts too broadly (spending the first page on general background that any reader in the field already knows) or too narrowly (jumping to the specific research question without establishing why it matters). The most reliable structure for a scientific paper introduction works through four stages in sequence.

The introduction ends with the research question or hypothesis stated precisely not in vague exploratory language ("this paper examines the relationship between X and Y") but as a testable, specific claim ("we tested whether X produces a statistically significant increase in Y under conditions Z, measured by [specific instrument or method]"). The precision of the research question at the end of the introduction is what allows the reader to assess whether the Methods section is appropriate for answering it.

The Results section presents what the study found objectively, completely, and without interpretation. The boundary between Results and Discussion is one of the most frequently violated structural conventions in student science papers, and it matters because conflating the two undermines the reader's ability to assess the evidence independently of the researcher's interpretation. Results tell the reader what was observed; Discussion tells the reader what those observations mean. Keeping this separate is not pedantic formalism it's what allows the reader to evaluate whether the Discussion's interpretation is actually supported by the Results that precede it.

Quantitative results should be presented with the appropriate statistical summary means and standard deviations (or medians and interquartile ranges for non-normal distributions), test statistics (F, t, χ², r), degrees of freedom, p values, and effect sizes. Effect sizes (Cohen's d, η², r) are increasingly required alongside p values in psychology and social science research following the American Psychological Association's 2019 Statistical Guidelines, which noted that p values alone are insufficient for assessing practical significance. For studies comparing multiple groups, the correction method for multiple comparisons (Bonferroni, Benjamini Hochberg false discovery rate) should be stated in the Methods and applied to the p values reported in the Results.

Tables, graphs, and figures when they help and when they don't

Graphs, charts, and tables are appropriate in the Results section when they genuinely aid clarity when a visual pattern in the data is more immediately comprehensible than the same information in prose or tabular form, or when the volume of numerical results would make prose presentation unwieldy. They're not appropriate as decoration or as a way to make the Results section look more substantial. Every figure and table needs a complete, self-contained caption (a reader should be able to understand what the figure shows without having to read the surrounding text), axes labelled with units, appropriate statistical error bars (with the type of error bar specified in the caption standard deviation, standard error, or 95% confidence interval), and an explicit reference in the text ("Figure 2 shows...") so the reader knows where each visual element fits in the overall Results narrative.

The abstract and conclusion are both best written after the main body sections are complete because both depend on knowing exactly what the paper argues and establishes, rather than what you planned for it to argue and establish before you began writing. This is the most common structural sequencing mistake in science paper writing: drafting the abstract first, then finding that the paper's actual findings and conclusions differ from what the abstract set up, without going back to revise the abstract to match.

The conclusion is not a second Discussion. Its job is to bring the paper to a definitive endpoint summarizing the key findings in one or two sentences, restating the answer to the research question, noting the main implications, and pointing to the most important future research directions the current study identifies. It should not introduce new evidence; revisit debates already settled in the Discussion or pad the paper's length. A conclusion that can be summarized as "we found X, which means Y, and future research should investigate Z" is doing its job well. One that covers three additional paragraphs of general reflection on the field is not.

The abstract, finally, is the section most students write first and should write last. A well written abstract tells the reader in 150–250 words exactly what question was asked, how it was investigated, what the key finding was (with a specific result rather than "findings were significant"), and what the finding means. It uses the same tense conventions as the main paper (past tense for methods and results, present tense for established knowledge and conclusions) and contains no information that doesn't appear in the main text. Many examiners and peer reviewers form their initial assessment of a paper's quality from the abstract alone it's worth the extra revision time it requires.

For PhD candidates and researchers working toward publication, the requirements go considerably beyond a standard coursework science paper and treating them as the same task produces manuscripts that get desk-rejected before peer review, which wastes significant time and delays the publication timeline. Journal manuscript preparation is a distinct service from coursework paper support, and the differences matter.

If you're preparing a science research paper for journal submission rather than for coursework assessment, flag this when you get in touch the support approach, timeline requirements, and level of detail needed are substantially different from a coursework paper, and accurate scoping from the start produces a better outcome than treating it as an extended coursework project.