To write an abstract, give the problem and the gap, what you did, what you found with the key numbers, and what that means. All of it has to fit inside the word limit your journal or university gives. Most people who come across your work will read the abstract and many of them will read nothing else. The editor looks at it and decides whether to go further. A reviewer reads it before agreeing to review. Someone searching on Scopus decides from it whether your paper is worth downloading. And examiner forms first view of your thesis from the abstract itself.
We cover abstracts for journal papers, conference submissions and PhD theses here. You will see the four moves every abstract makes and the word limits that come up most. Structured and unstructured abstracts are explained, and a weak and a strong version of one abstract are placed side by side. At the end there is a full thesis abstract as an example. Last reviewed September 2026.
What an abstract has to do
Think of the abstract as your full study in small size. Many first drafts treat it like an introduction instead. Background takes two sentences, then one line says “the results are discussed” and the draft stops there. A reader cannot use anything from that.
From the abstract alone, a reader should be able to cite your main finding. Is the main result in your draft, with a number if you have one? If not, the most important part is missing.
Search also depends on the abstract. Scopus, Web of Science, Google Scholar and Shodhganga all index the title and the abstract. The words somebody would type when looking for a study like yours should come in the abstract, used naturally.
The four moves every abstract makes
In every field, with or without headings from the journal, a good abstract makes the same four moves. The rough share of each one is given below.
- Context and gap (about 20%). Use one or two lines to say what the problem is and what is still not known. Leave out the general lines about how important the field is.
- What you did (about 25%). Mention the design, the sample or material, the setting and the main method of analysis. Put the numbers in, for example 412 respondents, 64 wells, three case companies, or a dataset of 1.2 million transactions.
- What you found (about 35%). Give the main results along with the key figures. Weak abstracts mostly leave this part out, and it should get the biggest share of the words.
- What it means (about 20%). Say what the study means for theory, practice or policy. Keep the claim at the level your data can support and not stronger.
If you are stuck, try one small exercise. Take a blank page and write one sentence per move, style you can fix later. Then expand each sentence till you reach the word limit. Filling out a complete skeleton is much easier than cutting down a long abstract which has no results in it.
Tense and voice
Write the methods and results in past tense (“we surveyed”, “adoption was higher”). For the conclusion and for general statements, present tense is used (“these findings suggest”). Most journals are fine with “we”. In thesis abstracts some universities like the passive voice better, so go as per the practice in your department.
Structured, unstructured and thesis abstracts
| Type | Where you meet it | Typical length | Headings |
|---|---|---|---|
| Unstructured (single paragraph) | Most engineering, science, social science and humanities journals | 150–250 words | None, but it still follows the four moves |
| Structured | Medicine, nursing, public health, many management journals | 200–300 words | Journal-specific, e.g. Background, Methods, Results, Conclusions |
| Conference abstract | Calls for papers, often judged on the abstract alone | 200–500 words | Sometimes set by the organisers |
| Thesis abstract | The front matter of your PhD thesis | Often one to two pages; set by your university | Usually none |
| Synopsis summary | The start of a registration or pre-submission synopsis | Half a page to one page | Usually none |
These limits are typical ranges only, not rules. The real limit comes from your target journal’s author guidelines or from your university’s thesis regulations. Journals follow the limit strictly, and many times the online submission system itself will not accept an abstract that goes over.
Structured abstracts
In a structured abstract you use the headings the journal gives, the exact same headings in the exact same order. Medical journals commonly have Background, Methods, Results and Conclusions. Emerald journals in management have Purpose, Design/methodology/approach, Findings and Originality/value, plus some optional headings for limitations and implications. The four moves stay the same. In a structured abstract the reader simply finds the moves faster because of the headings.
Highlights and graphical abstracts
Many Elsevier journals want highlights also. Highlights are three to five bullet points, and each one can be at most 85 characters with spaces counted. Some journals ask for a graphical abstract too, which is one image summing up the study. Neither of these is part of the abstract. Whether they are required or optional, the guide for authors of that journal mention it.
Thesis abstracts
A thesis abstract has to cover the full study and not just one finding, so it comes out longer. Usually it sums up the objectives, the methods used across the chapters, the main findings for each objective, and the contribution. Each university fixes its own limit. Some want one page, some want 500 words, some allow two pages. If your university has a published thesis format, our thesis structure guide has links to the university format guides we maintain. Examiners mostly read the thesis abstract and the synopsis together, so see that both say the same thing.
Example: a weak abstract and a stronger rewrite
We made up the example below only to show the difference, and its numbers are not from any real study. The weak version is the kind editors keep sending back. The stronger version describes a study of the same kind.
| Weak abstract | Stronger rewrite |
|---|---|
Water quality is a very important issue in the present scenario. In this study, an attempt has been made to study the water quality of groundwater in the study area. Various parameters were analysed. The results are discussed in detail. The study will be useful for policy makers. | Rural Tiruppur district gets most drinking water from groundwater, but since 2012 no study has measured nitrate levels near textile dyeing clusters. We sampled 64 bore wells within 5 km of three clusters in the pre- and post-monsoon seasons of 2024 and tested them for pH, TDS, nitrate, fluoride and heavy metals against BIS 10500. Nitrate exceeded the permissible limit of 45 mg/L in 22 wells (34%) before the monsoon and in 15 wells after it, with the highest values downstream of effluent outlets. Wells within 2 km of outlets should be watched first, and their households need other sources in the dry season. |
See what has changed between the two:
- The first line now talks about one problem in one place, and not about the whole world.
- “An attempt has been made to study” is replaced with what was actually done, giving the sample size, the seasons and the standard used.
- In place of “Various parameters were analysed”, the parameters are listed.
- Where it said “The results are discussed in detail”, the result is now stated, with its numbers.
- “Useful for policy makers” turns into a clear action that the data supports.
The weak version has 48 words and the stronger one has 104. So the trouble with the weak one was its content and not its length.
What does a PhD thesis abstract look like? A worked example
The paper abstract above reports a single finding. A thesis abstract must carry four or five years of work, and so its shape changes. It gives the objectives, then one sentence of method for each strand of the study. After that it answers the objectives one by one and closes with the contribution. Many scholars try to press all this into one paragraph of paper length. The examiner then cannot make out which finding goes with which objective.
The study below is invented. Madurai is a real place, but every number is made up.
Notice a few points in it. Each objective gets its own finding, in the order the objectives were listed, and the examiner can tick them one by one. The methods part names both strands with their sizes and keeps the details for Chapter 3. One limitation is also written plainly at the end. If a thesis abstract claims that nothing went wrong, expect questions about it in the viva-voce. This example comes to roughly 310 words. Open the thesis regulations of your university and see the number there before you cut or add.
Common mistakes in abstracts
- No results. We discussed about this above. It is still the fault we see most commonly.
- Too much background. Half of the words go on explaining why the field is important.
- Copying sentences from the introduction. Once the paper is complete, sit and write the abstract fresh. Sentences lifted from other parts of the paper will also match against your own preprint or thesis in the similarity report.
- Claims the paper does not support. If you ran a regression on survey data, you have not “proved” anything with it. Editors look at the abstract to see whether the author is claiming more than the data shows.
- Abbreviations and citations. Write the terms out in full. References in the abstract most journals will not take, so leave them for the main text.
- Different numbers in the abstract and the results. Mostly this comes when the scholar runs the analysis again after writing the abstract. At the end, sit with the results tables and check each figure once more.
- Phrases that fill space. Remove things like “In this paper”, “it was found that” and “the present study is an attempt to”, and give those words to your results.
When the abstract is clear, the paper has a better chance in the editor’s first reading. For the other points editors check, see the desk rejection guide.
How to choose keywords and check an abstract
Most journals want four to six keywords. Pick the terms people would search with which are not already in your title, because the title gets indexed in any case (see how to write a research title). If your method or context is unusual, add that also (“panel data”, “Kerala”, “finite element analysis”). Some journals use a controlled vocabulary like MeSH in medicine, and in economics some ask for JEL codes. In those journals, use that system.
Go through the points below before you submit:
- Someone who reads the abstract and nothing else can say what the problem was, what you did, what the main result was (with a number) and what it means.
- Each number in the abstract is the same as the number given in the body of the paper or thesis.
- Nothing is said in the abstract that the main text does not also say.
- There are no citations, no abbreviations left unexplained, and no mention of any figure or table.
- You have actually counted the words and they fit the journal’s or university’s limit. Don’t go by guess.
- The keywords a searcher would type come in the title or the first two sentences, and they read naturally there.
- You cannot lift any sentence as it is and put it in the abstract of some other study.
For the other sections of the paper, our guide on how to write a research paper goes through them one by one. If your university wants a style guide like APA for the thesis, the APA 7 guide shows how the abstract is set out on the page.
Want a specialist to help with this in your own work? Research Paper Review Before Submission: a free first consultation, then a written scope and fixed quote.
Sources
FAQ
Questions scholars ask
Should I write the abstract first or last?
Write it at the end. A rough version in the beginning can help you think clearly about the study. The final abstract should come from the finished paper, so that it reports what the paper actually shows.
Can the abstract contain citations?
Mostly no. People many times read an abstract with no reference list near it, so most journals don’t want citations there. In a few fields one citation is let through, mostly when your study directly replicates an earlier paper or replies to it. Check the author guidelines of your journal for this.
Can I reuse my thesis abstract for a journal paper?
No. The thesis abstract summarises the full study, and a paper abstract summarises one finding only. For each paper from the thesis, write a new abstract from scratch. Then your own thesis on Shodhganga will not show up as a text match.
What if my results are not significant?
Still report them, and report them clearly. A line like “No difference was found between the two groups (p = 0.41)” is a result. Many journals publish a well-designed study that has a null result. If you cover it up with vague words, the editor sees a weak abstract and nothing else.
How long is a conference abstract?
The call for papers decides it, but usually it falls somewhere from 200 to 500 words. Conference abstracts are many times reviewed alone, without the paper, so concrete results are needed. If the work is still in progress, state clearly what results will be ready by the time of the conference.
