An annotated bibliography lists your sources, each followed by a short paragraph that summarizes and evaluates it. In STEM, the evaluation focuses on methods, data and how useful the source is for your project.
This page explains how to choose sources, what each annotation should contain, how to judge technical papers and how to format the result. It includes a full example and a checklist you can work through before submitting.
The Four Parts of a Strong STEM Annotation
A strong annotation does four things in about 100 to 200 words: it states what the study did, summarizes the key findings with numbers where useful, evaluates the methods and limitations, and explains how the source relates to your own project.
The citation comes first, formatted in your required style, with the annotation underneath.
Structure of an Annotated Bibliography
| Part | What it contains | Typical length |
|---|---|---|
| Title and optional introduction | Your topic or research question and how you chose sources | One short paragraph, if required |
| Citation | Full reference in APA, IEEE, Harvard or your required style | One entry |
| Summary | Aim, method and main findings | 2 to 3 sentences |
| Evaluation | Strengths, limitations, reliability of data | 2 to 3 sentences |
| Relevance | How the source supports or shapes your project | 1 to 2 sentences |
Order entries the way your citation style orders a reference list. APA and Harvard sort alphabetically by first author; IEEE usually numbers entries, though for an annotated bibliography many tutors accept alphabetical order. Check your brief.
Choosing the Right Sources
Start from your research question, not from whatever the search engine shows first. A focused question makes it easier to judge whether a paper belongs.
- Databases: use subject databases such as IEEE Xplore, PubMed, Scopus or Web of Science rather than general web searches.
- Peer review: prefer peer-reviewed journal articles and major conference papers. Label preprints clearly as not yet peer-reviewed.
- Recency: fast-moving fields such as machine learning need recent work; include older papers only if they are foundational.
- Balance: mix review articles for context with primary studies for detail, and include work that disagrees with your expected answer.
How to Evaluate a Technical Source
Evaluation is what separates an annotated bibliography from a list of abstracts. Ask the questions a reviewer would ask.
| Question | What to look for |
|---|---|
| Is the method sound? | Controls, sample size, calibration, appropriate statistics |
| Are the data available? | Shared datasets or code, enough detail to reproduce |
| Do conclusions follow from results? | Claims matched to evidence, limits acknowledged |
| How general is it? | One lab, one dataset or one material, or broader? |
| Who funded it? | Declared conflicts of interest |
You do not need every question for every source. Pick the one or two points that most affect how much you can rely on it.
Step by Step: Writing the Annotations
- Write your research question at the top of a working document.
- Search, screen titles and abstracts, and shortlist sources.
- Read each shortlisted paper's abstract, methods, results and conclusion closely.
- Note the aim, method, key numbers and limitations in your own words.
- Format the citation, then draft the summary, evaluation and relevance.
- Edit each annotation to the required length and check the citation details against the source.
Write annotations in your own words, in full sentences. Copying the abstract is plagiarism and also misses the point, because the abstract cannot evaluate itself.
Example Annotation
APA 7th edition
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
This review explains how deep learning models learn layered representations of data through backpropagation, and surveys convolutional networks for images and recurrent networks for sequences such as speech and text. The authors connect these methods to rapid gains in image and speech recognition.
As a review by three of the field's founding researchers, it is authoritative on core concepts, but it predates transformer models and so says little about current language-model architectures. It reports no new experiments.
For my project on image classification of plant diseases, it provides the conceptual grounding for using convolutional networks; I will rely on more recent primary studies for performance benchmarks.
The same citation in IEEE
[1] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015, doi: 10.1038/nature14539.
Notice how the evaluation names a specific limitation (its date) and the relevance sentence says exactly how the source will be used.
Common Mistakes
| Mistake | Fix |
|---|---|
| Annotations that only summarize | Add at least two sentences of evaluation |
| Copying or lightly rewording the abstract | Write from your own notes after reading |
| Vague relevance ("this is useful") | Say what the source will support in your project |
| Citation details wrong or inconsistent | Check each entry against the source and the style guide |
| Too many old or non-academic sources | Favor recent, peer-reviewed work unless a classic is needed |
Checklist Before You Submit
- Every source fits the research question.
- Each entry has summary, evaluation and relevance.
- Annotations are within the word limit and in your own words.
- Citations follow one style consistently, with DOIs where available.
- Entries are in the order the style or brief requires.
- Preprints and gray literature are labeled as such.
How STEM Donkey Helps with Annotated Bibliographies
Send your topic or research question, the number of sources, the citation style, any required databases or date range, and the word count per annotation. Deadlines run from 3 hours to 20 days.
A writer in your field prepares a custom annotated bibliography written from scratch, using real, verifiable sources with evaluative annotations. It is plagiarism-checked and meant for study and reference.
Choose a Standard, Master or Elite writer, see the price before you pay, and get free revisions within the scope of your original request. Your identity is never shared with the writer.
Want Sources Found and Annotated for You?
Send your topic, citation style and source count. You get a custom annotated bibliography with evaluative annotations and correctly formatted references.
Order Your Annotated BibliographyFree revisions · Full refund if late · Written from scratch for your order
Frequently Asked Questions
Usually 100 to 200 words, unless your brief says otherwise. Some tutors ask for short descriptive annotations of around 50 words.
A descriptive annotation summarizes the source. An evaluative one also judges its quality and relevance. Most university STEM tasks expect evaluative annotations.
Often yes in the relevance sentence, as in "I will use this to...". Check your brief, as some tutors prefer the third person throughout.
Follow your brief. If none is given, 8 to 15 sources is common for an undergraduate project.
No. An annotated bibliography treats each source separately. A literature review synthesizes sources into themes and an argument.
Yes, if your tutor allows them, but label them as not peer-reviewed and check whether a published version now exists.