How to Write a Research Paper: Structure from Abstract to Conclusion

A research paper is not a diary of everything you did. It is a carefully built argument that moves the reader from a problem to an answer. Most journals in science, engineering, social science and medicine use the IMRaD structure (Introduction, Methods, Results and Discussion), and once you understand what each section must do, writing becomes far less intimidating. This guide walks through every section and uses one small worked dataset to show how a finding travels through the paper.

The Overall Shape

Think of a research paper as an hourglass:

• Wide at the top: the Introduction starts with the broad problem and narrows to your specific question.
• Narrow in the middle: the Methods and Results are focused and factual.
• Wide at the bottom: the Discussion widens back out to what your findings mean for the field.

The full order is usually Title, Abstract, Keywords, Introduction, Literature Review (sometimes merged with the Introduction), Methods, Results, Discussion, Conclusion, References and Appendices. Always check your target journal's author guidelines, or your university's thesis manual, because word limits and headings vary.

Writing order is not reading order. Most experienced researchers write the Methods and Results first, then the Discussion, then the Introduction, and the Abstract and Title last.

Title and Abstract

A good title is specific and searchable. It names the variables, the population and sometimes the design. Compare "A Study of Students" with "Weekly Study Hours and Mathematics Achievement Among FSc Students in Punjab". The second tells readers and search engines exactly what you studied.

The abstract (usually 150–250 words) is a miniature paper. A structured abstract answers five questions:

1. Background: why does this matter?
2. Objective: what did you investigate?
3. Method: who took part, how many and with what design and analysis?
4. Results: what did you find? Give the key numbers.
5. Conclusion: what does it mean?

Never include citations, undefined abbreviations or results that do not appear in the paper. Add 4–6 keywords that do not simply repeat the title.

Introduction

The Introduction justifies your study. A reliable three-move pattern, sometimes called "Create a Research Space", works like this:

1. Establish the territory: show that the topic is important, with evidence and citations.
2. Identify the gap: explain what is unknown, contradictory or untested. For example, "Most studies have been conducted in Western contexts; evidence from South Asia is limited."
3. Fill the gap: state your aim, research questions or hypotheses and, briefly, your contribution.

End with clear, testable hypotheses, such as "H1: Weekly study hours will be positively associated with exam scores."

Methods

The Methods section must let another researcher replicate your study. Typical subsections are:

• Design: experimental, quasi-experimental, cross-sectional survey, longitudinal and so on.
• Participants: the population, sampling technique, sample size and its justification, inclusion criteria and demographics.
• Instruments: the questionnaires or apparatus used, their sources, and reliability (for example Cronbach's α) and validity.
• Procedure: what happened, in what order, and ethical approval and informed consent.
• Data analysis: the tests used, assumption checks, the significance level and the software (for example SPSS 29, R 4.4, Stata 18 or Python).

Write in the past tense, and use enough detail that nothing important is left to guesswork.

Results: A Worked Example

The Results section reports findings without interpretation. Here is a small made-up example of how one finding is produced and reported.

A researcher records weekly study hours (x) and exam scores (y) for six FSc students:

Step 1: Means. \( \bar{x} = 36/6 = 6 \) and \( \bar{y} = 390/6 = 65 \).

Step 2: Deviations. For x they are −4, −2, −1, 1, 2, 4, and for y they are −11, −5, −7, 5, 7, 11.

Step 3: Sums of squares and cross-products.

$$ S_{xy} = 44 + 10 + 7 + 5 + 14 + 44 = 124 $$

$$ S_{xx} = 16 + 4 + 1 + 1 + 4 + 16 = 42, \qquad S_{yy} = 121 + 25 + 49 + 25 + 49 + 121 = 390 $$

Step 4: Pearson correlation.

$$ r = \frac{S_{xy}}{\sqrt{S_{xx}S_{yy}}} = \frac{124}{\sqrt{42 \times 390}} = \frac{124}{127.98} = 0.969 $$

Step 5: Regression line.

$$ b = \frac{S_{xy}}{S_{xx}} = \frac{124}{42} = 2.95, \qquad a = \bar{y} - b\bar{x} = 65 - 2.952 \times 6 = 47.29 $$

So \( \hat{y} = 47.29 + 2.95x \), and \( R^2 = r^2 = 0.94 \).

Step 6: Significance test.

$$ t = \frac{r\sqrt{n-2}}{\sqrt{1-r^2}} = \frac{0.9689 \times 2}{\sqrt{1 - 0.9387}} = \frac{1.9377}{0.2476} = 7.83, \qquad df = 4, \quad p = .001 $$

Sanity check: the slope is positive, matching the clear upward pattern in the table. At the mean of 6 hours, the line predicts \( 47.29 + 2.95 \times 6 \approx 65 \), which is the mean score, as it should be.

How it appears in Results (APA style):

Weekly study hours were strongly and positively correlated with exam scores, r(4) = .97, p = .001. A simple linear regression showed that each additional hour of study per week was associated with a 2.95-point increase in exam score, b = 2.95, R² = .94.

Use tables and figures for detailed numbers, and use the text to highlight the key patterns. Report non-significant results too.

Discussion

The Discussion answers the question "so what?" A strong structure is:

1. Restate the main finding in plain words, without repeating all the numbers.
2. Compare it with previous studies: does it agree or disagree, and why might that be?
3. Explain the mechanisms or theory behind the finding.
4. Discuss implications for practice, policy and theory.
5. Acknowledge limitations honestly. In our example, a sample of six cannot be generalised, and correlation does not prove causation.
6. Suggest future research that addresses those limitations.

Avoid overclaiming. Write "suggests" or "is associated with" for correlational data, never "proves" or "causes".

Conclusion

The Conclusion (often 1–3 paragraphs) gives a crisp answer to your research question, summarises the contribution and may offer recommendations. It must not introduce new data or new citations. For our example: "Study time appears to be a strong predictor of exam performance among FSc students, supporting structured study-hour programmes, although larger samples are needed to confirm this."

References and Final Checks

Use one referencing style consistently (APA, Harvard, IEEE, Vancouver or whatever your journal specifies). A reference manager such as Zotero, Mendeley or EndNote prevents mismatched in-text citations. Before you submit:

• Check that every in-text citation appears in the reference list, and the other way round.
• Make sure the numbers in the abstract, text and tables agree exactly.
• Number all tables and figures, and refer to each one in the text.
• Run a similarity check and proofread aloud.

Did You Know?

• The IMRaD format spread through scientific journals mainly after the 1940s, and by the 1980s it dominated medical research papers.
• Many readers decide from the title and abstract alone whether to read a paper, so these short sections do a huge amount of work.
• A Results section that reports only significant findings contributes to "publication bias", which distorts the evidence in meta-analyses.

Key Takeaways

• Follow IMRaD: Introduction (why), Methods (how), Results (what) and Discussion (so what).
• Write the Methods and Results first and the Abstract last. Make the abstract a complete mini-paper with key numbers.
• Report results factually in APA style, with test statistics, p-values and effect sizes, and save interpretation for the Discussion.
• Acknowledge limitations, avoid causal language for correlational data, and never add new material in the Conclusion.

If you need help structuring your paper, analysing your data or writing up your results, Amathify's subject experts and PhD scholars can provide step-by-step solutions and assignment writing support. Contact us on WhatsApp or through the Ask a Question form.