Section 3.7

Abstracts & Titles

Q: How long should a thesis abstract be?

Most abstracts run 150–250 words; check your program's or journal's limit, which is often a hard 200 or 250. But the aim is completeness, not a word count. A tight 180-word abstract that makes all five moves (context, aim, method, a result with numbers, and a calibrated conclusion) is better than a padded 250-word one that never says what was found. If you're over the limit, cut the sentences that make no move: general background and "further research is needed" go first, never the numbers.

Q: What makes a good research title?

Findability. In a database or a search engine your reader meets the title before anything else, so it should contain the words they would actually type: your variables, your population, your design. Two dependable shapes work: state the finding ("More Sleep Predicts Higher Student Wellbeing") or state the question ("Does Napping Improve Memory?"). Avoid vague, clever-but-empty titles like "An Investigation into Certain Factors". They don't contain the words a reader would search for, and they tell the reader nothing. Put the same keywords in the keyword field, too.

Your title and abstract are the most-read part of your work, and often the least revised. Most people who come across your study see only these, in a database search, a supervisor's inbox, or a search engine's results. If the title doesn't say what you found and the abstract doesn't summarize it in one tidy paragraph, most of those people will not read further. Both follow a recipe. Write the paper first, then the abstract, and the title last of all.

The abstract's five moves

A good abstract is not a summary of every section. It makes five specific moves, usually one to three sentences each, in roughly 150–250 words:

Context: the gap or question in one line. Why should anyone care? (Not "sleep is important," but the specific thing we don't yet know.)
Aim: what this study set out to test, stated as a question or hypothesis.
Method: who, how many, what design, what measure, enough to judge the evidence without recruitment minutiae.
Result, with numbers. This is the move students skip, and it's the one readers came for. Give the actual estimate: the effect size, the direction, the interval. "Sleep predicted wellbeing, r = .28" tells the reader more than "sleep was related to wellbeing."
Conclusion: what it means and how far it reaches, calibrated to the design (see Discussion & Limitations).

If a move is missing, a reader has to open the full paper to answer an obvious question, and in a list of search results most readers will move on instead.

Write it last

The abstract is written after the paper is finished, never before. Only then do you know your final numbers, your actual conclusion, and which framing survived. If you draft the abstract first, it will contain claims your results may not support. Treat it as a summary of the finished paper, not a plan, and check every number in it against the Results (the Final Checklist covers this check).

When the abstract comes with labels

Many journals, and some degree programs, ask for a structured abstract: the same content with headings printed in front of it, usually Objective, Method, Results, Conclusions. If you have written the five moves, you have already written it, because the labels map onto them almost one to one (context and aim both fold into Objective). A labeled abstract is also much harder to pad, because a heading with one vague sentence under it looks empty.

The five moves are also close to what APA's reporting standards require of an abstract. One item on that list matters especially: JARS wants findings reported with effect sizes and confidence intervals, not just significance. So the result move needs more than a number. It needs the estimate, its interval, and its direction, written as Reporting Statistics in APA Style sets out. APA also caps the abstract at 250 words, which is the upper end of the range above rather than a target to reach.

Below the abstract comes one more line that most students improvise: the keywords. APA specifies its format. Indent it like a paragraph, put the label Keywords: in italics, then three to five terms on the same line, lowercase except for proper nouns, separated by commas, with no period at the end. Choose them the way you chose the title, by asking what a stranger would type: your constructs, your population, your design. Spelling out an abbreviation you used in the title is a good use of one of the five slots, because a database matches strings and not meanings.

Titles that get found

A title's main job is to be findable by the words your reader would type. Wordplay can work in a newspaper headline, but a research title should be searchable first. Two reliable shapes:

State the finding: "More Sleep Predicts Higher Student Wellbeing" tells the reader the result before they click.
State the question: "Does Napping Improve Memory? A Diary Study of 120 Undergraduates" works when the answer is nuanced.

Either way, put the main keywords (your variables, your population, your design) into the title and the keyword field. Those are the terms a future student types into a database when looking for a study like yours. A title like "An Investigation into Certain Factors" will not turn up in their searches.

Build a better abstract

Below is a draft abstract, sentence by sentence. Some sentences make one of the five moves, and others just add words. Toggle sentences in and out to cover all five moves, put numbers in the result, and cut every sentence that makes no move. The panel grades your draft as you go.

🎯 Abstract grader

Toggle sentences to cover all five moves, put numbers in the result, and waste nothing.

Every sentence makes a move. A finished abstract is five moves and almost nothing else: context, aim, method, a result with numbers, and a calibrated conclusion. Write it last, keep it to ~150–250 words, and title the paper with the keywords your reader would actually search. If a sentence makes no move, cut it.

Why it matters: a thesis is graded in full, but people find it through its abstract and title. A busy supervisor, an examiner skimming a pile, and a stranger three years from now will all read the abstract, and often nothing else. Make it complete, accurate, and specific about the numbers, and give the paper a title that states what you found. Then the people looking for a study like yours can find it.

Common questions

How long should a thesis abstract be?

Most abstracts run 150–250 words; check your program's or journal's limit, which is often a hard 200 or 250. But the aim is completeness, not a word count. A tight 180-word abstract that makes all five moves (context, aim, method, a result with numbers, and a calibrated conclusion) is better than a padded 250-word one that never says what was found. If you're over the limit, cut the sentences that make no move: general background and "further research is needed" go first, never the numbers.

Should my abstract mention limitations?

One clause, inside the conclusion move, and only when the limitation changes what the finding means. In "Small gains in sleep may support student wellbeing, though the correlational design cannot establish causation," the caveat takes eight words. They are needed, because a reader who stops at the abstract would otherwise take away a causal claim you never made. Leave out a separate limitations sentence or a list. An abstract is 250 words at most, and a caveat that does not change the interpretation takes space the result move needs. The full accounting belongs in the Discussion, where you have room to say what each weakness could have altered.

What makes a good research title?

Findability. In a database or a search engine your reader meets the title before anything else, so it should contain the words they would actually type: your variables, your population, your design. Two dependable shapes work: state the finding ("More Sleep Predicts Higher Student Wellbeing") or state the question ("Does Napping Improve Memory?"). Avoid vague, clever-but-empty titles like "An Investigation into Certain Factors". They don't contain the words a reader would search for, and they tell the reader nothing. Put the same keywords in the keyword field, too.