Digest · Method

Preparing a Research or Project Document

Most class assessments and project reports are marked with a rubric that has the same six parts: an introduction, a review of past work, a note on your data, some statistics, a few charts, and a conclusion with next steps. This is a simple, step-by-step guide to writing each part well, plus a ready-to-use outline and a short list of things a beginner can actually read.

The six parts, in short

Introduction — say what the topic is and why it matters

Review of Literature — sum up what others found, and cite them

Variables & Data — list what you measured and where it came from

Descriptive Statistics — work out the numbers, then explain them

Charts — a few clear ones, each tied to the text

Conclusion — sum up, then give clear next steps

Do this before you start writing

An hour of setup here saves you days later. Do these five things first, in order:

Turn your topic into a question.A topic (“foreign investment in India”) has no answer. A question (“How did foreign investors’ holdings in India change between 2013 and 2025, and what does that mean for the banks that hold those assets?”) does. Write it in one sentence and keep it in front of you.
Get the data before you commit.Open the real source, download one file, and check it has the numbers, the years, and the units you need. Most projects go wrong because the data is not there in a usable form — find that out on day one, not in week three.
Set up a reference manager. Install Zotero and its browser button. Save every source the moment you read it. Never keep citations in your head or a rough note.
Pick one citation style and one template. APA 7th is a safe choice unless you are told otherwise. Set the page up — margins, headings, page numbers, a table style — before you write any content.
Make the outline first. Create the six headings now as empty sections. You are filling a known shape, not inventing one.

Step 1 — Introduction

What the marker wants:a short, clear introduction that says what the topic is, why you are doing it, and what it covers — and gives the reader a reason to care.

Write it in four small moves:

Background— two or three sentences of context. Explain any term the reader needs right here.
Why it matters— who is affected, how big the thing is, what decision depends on the answer. This is the sentence that earns the “holds the reader’s interest” marks.
The question— one sentence that names exactly what you are trying to find out.
Aims and scope— a numbered list of two to four aims, then a sentence on what the study covers and, just as important, what it leaves out.

Common mistakes: a page of general history before the question shows up; no clear aims; no scope, so the reader cannot tell what a full answer looks like. Keep the whole thing to about one page.

Step 2 — Review of Literature

What the marker wants:a focused summary of past work — the main themes, how the studies connect, where they disagree — with correct, consistent citations, ending in a clear statement of the gap you fill.

Group by theme, not by paper.Three to five short parts, each one pulling a few sources together. If one paragraph starts “Author (year) found …” and the next starts “Another author (year) found …”, that is a list, not a review.
Say what it adds up to. What does past work agree on, argue about, and never look at?
Cite everything. Every fact gets an in-text citation. Use one style. Let Zotero build the list so it stays consistent.
End with the gap. The last paragraph must say, in plain words, what is missing and how your study fills it. That one sentence links this section to your work.
A small table helps.Theme, main sources, main point — it shows the summary at a glance and is easy to mark.

How much to read: for a class assessment, about 8 to 15 real sources. Prefer journal articles, working papers from bodies like NIPFP or the RBI, and official statistics over blogs and news.

Step 3 — Variables and Data

What the marker wants: clear, relevant variables and a well-described dataset, with a reason for each variable.

Describe the data

What each row stands for (a country in a year? a firm in a month? a person?).
The time period and how often it is measured, and why that window.
The source, named exactly.
How you cleaned or changed the data.
Known limits — missing values, estimates, gaps in the series.

Put the variables in a table

One row per variable, with columns for name, meaning, unit, and source. Then, below the table, one short paragraph per variable that answers a single question: why is this variable here? Tie it to an aim or to the past work. A variable with no reason should be removed.

Common mistakes:vague meanings (“market performance” instead of “NIFTY 50 closing level on 31 March”); variables that appear in the analysis but not the table; no source; no reason.

Step 4 — Descriptive Statistics

What the marker wants: a full set of basic statistics anda clear, useful explanation of them — not just a table.

Work out a full set

Middle: mean, median (and mode for category data).
Spread:standard deviation, variance, range, and the coefficient of variation (SD ÷ mean) for comparing series of different sizes.
Shape: skewness (is it lopsided?), and kurtosis if it matters.
Extremes: the smallest and largest values, each with the year it happened.
Growth, for data over time: year-on-year growth, the compound growth rate over the whole period, and how many years fell.
Links: a correlation table if you have several number variables.

Put these in one clean summary table. Round the same way everywhere. State the units in the caption.

Then explain the numbers

This is where most marks are won or lost. For each number that stands out, write a sentence saying what it means: “a coefficient of variation near 0.5 means the series swings a lot, so income built on it will jump around from year to year.” Point out the trend, the turning points, and the odd values, and say what they mean for your question. Number these points so the reader can follow them.

Common mistakes:a table with no words around it; stating the number (“the mean is 37”) without saying what it tells us; ignoring an obvious outlier.

Step 5 — Charts

What the marker wants: clear, correct charts that make a point better than words alone, and that connect to the text.

Match the chart to the job.Change over time → line chart. Make-up of a whole → stacked bar. Comparing groups → bar chart. Link between two things → scatter. Skip pie charts with more than four slices, and skip 3-D effects.
Label everything. Axis names with units, a caption that states the point, and a source line.
Mention each chart in the text. “As Figure 2 shows, the debt share speeds up after 2023.” A chart the text never names looks like decoration.
Fewer, better. Three or four charts that each make a point beat ten that repeat the table. If dropping a chart costs nothing, drop it.
Keep it honest. Start bar-chart axes at zero. Do not cut the axis to make a small change look big.

Step 6 — Conclusion and Recommendations

What the marker wants: a strong conclusion that pulls the findings together and answers your aims, then clear, doable steps based on what you found.

Sum up, don’t repeat. Go aim by aim: for each one from the introduction, say what you found. No new data or citations here.
Base each step on your own numbers. Every recommendation should point back to a result in Section 4 or 5, not to general opinion. “Because debt holdings grew fastest and stay put longer (Section 4, point 3), the bank should build its bond-handling service now” — specific, and earned.
Be concrete.Who does what, by when, measured how. Vague advice (“firms should be more efficient”) scores low.
Say what you could not do. A short, honest paragraph on the limits of the study and what a follow-up would add.

Match the rubric to your document

Before you hand it in, check each rubric part against the section that covers it, and run the quick check.

How each rubric part maps to a section of the document
Rubric part	Where it goes in your document	Quick check
(a) Introduction	Section 1 — background, why it matters, the question, your aims, what you cover	After one page, can a stranger say what your question is?
(b) Review of Literature	Section 2 — 3 to 5 themes, each summed up; the gap you fill; a reference list	Is every claim cited? Does the section end by naming the gap?
(c) Variables & Data	Section 3 — a table (name, meaning, unit, source) and a reason for each one	For every variable, can you say in one line why it is there?
(d) Descriptive Statistics	Section 4 — a summary table (mean, median, spread, min, max, growth) and a written read-out	Does each number have a sentence saying what it means?
(e) Charts	Section 5 — 3 or 4 charts, each with a caption and a mention in the text	Take a chart out. Does the point weaken? If not, drop it.
(f) Conclusion & Recommendations	Section 6 — sum up against your aims, then clear, doable steps	Do the steps come from your own numbers, not general opinion?

Formatting and handing it in

Order: title page, abstract, numbered sections in rubric order, references, appendix. Add a contents page once it runs past about 8 pages.
Tables and charts: number them (Table 1, Figure 1), give each one a caption and a source line, and mention each in the text.
Citations: one style all the way through, in the text and in the list. Let the reference manager format it.
Length: follow the brief. If there is none, this kind of assessment is usually 2,500 to 4,000 words plus tables.
Honesty:quote rarely and always with credit; put things in your own words; run a plagiarism check; keep your raw data and working file — you may be asked to redo a number. AI has its own rules — see the next section.
File: save as PDF, name it Course_Topic_RollNo.pdf, and open it once on another device to check nothing moved.
Proofread last: read it out loud, check every number against your working file, and make sure the aims in Section 1 match the findings in Section 6.

AI, copying, and using AI the right way

These are two different problems, and a document can fail on either one.

Plagiarism (copying)means showing someone else’s words, ideas, data, or structure as your own with no credit — whether it comes from a classmate, a website, a paper, or an AI tool. Copying a paragraph from a journal without quote marks and a citation is copying. Pasting an AI paragraph and handing it in as your writing is also copying, because the thinking is not yours and there is no credit.
Misusing AIis wider: handing in analysis you did not do, letting a tool make up data or citations, or using AI in a way your school has not allowed. The real test is authorship — can you explain, defend, and redo every part of this document?

Where AI really helps

As long as your school allows it, these uses are usually fine, because you stay the author and can check the output:

Turning a vague topic into a clear question.
Explaining an idea or a statistic that you then confirm from a textbook.
Suggesting an outline that you fill in yourself.
Fixing grammar and tightening writing that you wrote.
Helping fix your analysis code, or explaining an error message.
Summing up a paper you have actually read, to help you study.

Where AI gets you into trouble

Fake citations.Tools invent real-looking references — real names, real journals, a link that goes nowhere. Every reference must be one you opened yourself.
Made-up data or results. Never let a tool create numbers, a dataset, or the statistics. Work them out yourself from the real source and keep the file.
Handing off the explanation. Reading your results and giving the next steps are the parts being marked. If a tool wrote them, there is nothing of yours to mark.
A review of things you did not read. A summary of sources you never opened will get them wrong and will cite things that do not say what you claim.
Word-for-word text. Pasting tool output into the document, even after small edits, with no quote marks and no citation.

A simple checklist for using AI

Read the rules first.Schools differ — some ban AI, some allow it if you say so, some allow it only for certain tasks. If it is unclear, ask the teacher in writing and keep the reply.
Stay the author. Do not hand in a sentence you could not have written and cannot defend if asked.
Check everything. Test every fact, number, and citation against the real source before it goes in.
Do your own analysis. The data, the statistics, and the charts are yours. Keep the raw data and the working file.
Say so if required.Add a short note — for example: “I used [tool, version] on [date] to [outline the introduction / fix grammar]. All analysis, explanation, and final text are my own.”
Cite AI if you quote or closely copy it. APA treats it like software, for example OpenAI. (2026). ChatGPT (Aug 2026 version) [Large language model]. https://chat.openai.com, with the prompt described in the text or an appendix.
Keep a record. Save your drafts and, if you can, your prompts. Then you can show how the document was made.
Run the plagiarism / AI check your course gives you, and read the report, not just the score.

One quick test before you hand in any paragraph: Could I redo this, say it out loud, and point to the source behind every claim? If not, it is not ready.

A ready-to-use outline

Start from this. Each line is a heading and what goes under it. A full example fills every section with real data and a short explanation.

Title page

Title, your name, roll number, course, date.

Abstract (150–200 words)

One line each: the problem, the data, the method, the main result, what it means.

1. Introduction

Background → why it matters → the question → aims (numbered) → what you cover and what you don't.

2. Review of Literature

3 to 5 short themed parts, each summing up a few sources; end by naming the gap you fill.

3. Variables & Data

Unit of study, time period, a variable table, one short reason per variable, data sources, limits.

4. Descriptive Statistics

The data table, the summary table, then 4 to 6 numbered points explaining the numbers.

5. Charts

A trend chart, a make-up chart, a compare chart; a short line on how they help the argument.

6. Conclusion & Recommendations

Answer each aim; give 4 to 6 clear steps; say what the study could not do and what comes next.

References

Every source you cited, in one style (APA 7th), A to Z.

Appendix

Raw data, notes on how you worked out the numbers, a checklist to refresh the data.

A good topic to practise on: “How foreign investors’ holdings in India changed, and what shifted between shares and bonds, from 2013 to 2025.” It has a clean public dataset (NSDL FPI Monitor), an easy shares-vs-bonds angle, and enough years for statistics and three or four charts.

Things a beginner can actually read

Learn how to write research

The Craft of Research — Booth, Colomb, Williams and others (Univ. of Chicago Press)
The classic beginner book on turning a topic into a question and then a draft.
How to Write a Great Research Paper — Simon Peyton Jones (talk + slides)
About 45 minutes, free. The best start if you only have one evening.
How to Write a Publishable Paper as a Class Assignment — Gary King
Written for coursework: how to keep the scope small, how to lay it out, common mistakes.
Economical Writing — Deirdre McCloskey
Short, plain advice on clear writing. Pairs well with Strunk & White's Elements of Style.
Purdue OWL — APA Style (7th edition) guide
Free and full of examples. Use it for in-text citations and the reference list.

Reference managers and style

Zotero — free reference manager
Add the browser button on day one. It saves your sources and builds the reference list for you.
Overleaf — write in LaTeX in the browser (optional)
Only if your course asks for LaTeX. The "Learn" pages are a good tutorial. Word is fine otherwise.
APA Style — How to cite ChatGPT and other AI
The proper way to cite and declare AI output. Check your own school's AI rules first.

Papers and reports to copy the style of

Read these for shapeas much as content — notice how short the introduction is, how the data section is set out, how charts sit next to the text.

RBI Bulletin — monthly articles
A good model for clean tables and captioned charts. Copy the layout, not the length.
SEBI — Handbook of Statistics on the Indian Securities Market
Shows how to present a dataset: meanings, tables, and time charts side by side.
FII Flows to India: Nature and Causes — R. Chakrabarti (2001)
A short, readable study: question, data, simple analysis, conclusion.
Determinants of Foreign Institutional Investment in India — Rai & Bhanumurthy (2004)
Another compact example of the exact shape a rubric rewards.

Free datasets to practise on

About this digest

A plain, working method for research and project documents, built around the six-part rubric most class assessments use. Put together while doing coursework and kept here as a checklist to come back to.