The quality of your sources is judged on more than whether they exist. Markers ask whether each source is the right kind for your subject, whether it reports research or repeats it, whether it is recent enough, and whether its author has standing on the point. To improve the quality of your sources, test each one on those points and replace the ones that fail with the work behind them.
EdCitation publishes this guide, drawn from published rubrics and marking criteria, the CRAAP test and the SIFT method on their authors' own pages, and six library guides, all read on 2 October 2026. The worked example, a source list on nudging and healthy eating, is ours.
Three steps below are free in EdCitation, with no account. Find sources searches about 300 million published works with filters for kind of source and years, Verify references flags retracted papers in a pasted list, and Cite a source builds each new entry from the publisher's record. Choosing and reading the sources stays with you.
What do markers mean by the quality of your sources?
Markers mean three things: sources of a kind that can support an academic argument, chosen on several criteria rather than one, and handled critically rather than repeated.
What the marking criteria say
In the D band (40 to 49%) of the University of Edinburgh's School of Social and Political Science (n.d.), work "may over-rely on sources" that are not particularly relevant or "of a kind that would not support a credible academic argument"; in the failing E band it may rely on them wholly, including sources "of dubious or suspect quality". At the top, A1 work may be publishable in peer-reviewed journals.
The University of Nottingham's sociology scheme, published as an example in its quality manual (University of Nottingham, n.d.), measures handling. Lower second work is "dependent on the sources"; upper second treats source material critically; first-class work places it "within this broader framework of debate".
What the AAC&U rubrics add
The Association of American Colleges and Universities wrote its VALUE rubrics to assess institutions, not to grade papers. Its information literacy rubric shows what "quality" contains: each level adds criteria for choosing sources (Association of American Colleges and Universities, n.d.-a):
| Level | How sources are chosen |
|---|---|
| 1 | A few sources, on limited criteria such as relevance |
| 2 | A variety of sources, on relevance and currency |
| 3 | Sources suited to the scope and discipline of the question, on relevance, currency and authority |
| 4 | As at 3, weighing which criteria matter most for the topic, including audience and bias or point of view |
Its written communication rubric (n.d.-b) asks at the top for "high-quality, credible, relevant sources" suited to the discipline and genre of the writing. Your own module's rubric is the one that counts: Check your paper reads its source rules, such as how many and how recent, into a checklist, free.
What makes a good academic source?
A good academic source passes four tests for the job you give it: review, closeness to the evidence, date and authority. Each has its own guide here.
Peer reviewed, where research is needed
A claim about what research found needs a source that experts checked before publication. Check the journal's stated review process and the article's received and accepted dates, as how to tell if an article is peer reviewed explains. Editorials and letters in the same journal usually are not reviewed.
Primary, not second-hand
North Dakota State University Libraries (2025) define an original research article as one written by the researchers who did the work, and contrast it with a journalist reporting on someone else's. When a sentence rests on a finding, cite the study; use reviews for the state of a field. Each kind, primary to tertiary, is in types of academic sources.
Recent enough for the field
No style sets an age limit; briefs and fields do, and a field's founding work can be decades old. The rules universities actually state are compared in how recent should my sources be.
Authority for this claim
Authority depends on the question. The Association of College and Research Libraries (2016) calls it constructed and contextual: a source reflects its creator's expertise and is judged against the context in which it is used. A national statistics office has authority on unemployment figures and none on a drug trial.
How do the CRAAP test and the SIFT method work?
The CRAAP test is a checklist for judging a source in front of you; the SIFT method is a set of moves for leaving that source to learn what it is and what better coverage exists. Our view: use SIFT first on anything from the open web, then CRAAP on what survives.
The CRAAP test
The Meriam Library at California State University, Chico (2010) gives five criteria, currency, relevance, authority, accuracy and purpose, each with questions. For an essay's list, three of its questions do most of the work: has the information been reviewed or refereed, is it supported by evidence, and does it exist to inform, or to sell, entertain or persuade?
The SIFT method
Caulfield (2019) sets out four moves: stop; investigate the source; find better coverage; and trace claims, quotes and media back to their original context. The last two raise a list's quality. Finding better coverage means setting aside the source that reached you and looking for the best one on the claim. Tracing means checking whether a cited paper really said what the page claims; his own example is a medical claim built on a research finding.
Most CRAAP questions are answered from the source itself, so a polished page can pass them. SIFT starts by looking the source up elsewhere, which is how a weak one is caught; teaching students to verify sources gives the research behind that habit.
Which weak sources should I replace, and with what?
Replace any source that cannot carry the claim you hang on it with the work it rests on. Each row below comes from a named library guide.
| Weak source | Why it is weak | Stronger replacement | Library guide | EdCitation step |
|---|---|---|---|---|
| Wikipedia article | Unclear authorship, constant editing, no original scholarship | The scholarly works it leads to | University of Michigan Library (2026) | Cite a source from each work's DOI |
| News story about a study | A journalist's account of someone else's research | The original research article, traced through the journal or researchers named | North Dakota State University Libraries (2025) | Find sources, searched by the claim |
| Textbook or encyclopedia used as evidence | Tertiary: a summary of other work | The primary and secondary sources it cites | The Master's University (2026) | Find sources, Kind of study |
| A company's or campaign's web page | May be selling, or arguing a position | Government data or published research | John M. Pfau Library (2026) | Find sources, In a journal |
| A chatbot's answer | Not a source of evidence; can cite works that do not exist | The credible source that holds the information | Dudley Knox Library (2026) | Verify references |
| An abstract read without the article | No substitute for reading the source | The full text | Lycoming College (2026) | Find sources, Free to read |
What the guides add
The University of Michigan Library (2026) calls Wikipedia a starting point toward "more reliable and rigorous scholarly texts". The Master's University (2026) finds textbooks fine for starting research but not acceptable as documentation; where a course sets a textbook, the brief wins. The John M. Pfau Library (2026) at California State University, San Bernardino, calls .gov sites generally reliable, though often influenced by the party in power. The Dudley Knox Library (2026) at the Naval Postgraduate School says to verify AI output in credible sources and cite those.
A retracted paper is the one weak source never kept for its finding. Swapping one out, sentence and all, is covered in replace a weak or retracted source with a better one.
What does a better source list look like? A before and after
A better list carries the same argument on sources that survive a marker's check. Our example, written for this guide, is a first-year essay's draft list on whether nudges make people choose healthier food. Every source on the right is real and was looked up in EdCitation on 2 October 2026.
| Before (our draft) | What was wrong | After |
|---|---|---|
| Wikipedia's article on nudge theory, for the definition | A starting point, not evidence | Thaler and Sunstein (2008), the book that named the idea |
| A newspaper story on a cafeteria experiment, for "nudges work" | A report of research | Bauer et al. (2021), a field study run in a cafeteria |
| A food marketing company's blog, for how large the effect is | Written to sell | Mertens et al. (2021), a meta-analysis, with its 2022 correction |
| Wansink et al. (2012), on attractive names for vegetables in schools | Retracted in 2018 | Broers et al. (2017), a systematic review and meta-analysis of nudges for fruit and vegetable choice |
| A chatbot's summary saying the evidence is settled | Not a source | Maier et al. (2022), a published reanalysis that disputes it |
What changed in the essay's claim
The draft said nudges reliably make people eat better; the stronger sources say something narrower. Mertens et al. (2021) pooled 447 effect sizes from 212 publications and found a small-to-medium overall effect, d = 0.43 as corrected, with food the most responsive domain (d = 0.65). Maier et al. (2022) reanalysed the same data, adjusting for publication bias, and concluded that no evidence remains that nudges work overall; for food, one of their two analyses pointed against an effect and the other was undecided. Bauer et al. (2021) tested three nudges in one cafeteria and found small, uneven effects, from healthier choices to less healthy ones. Broers et al. (2017) report a modest effect for fruit and vegetables, d = 0.30. Citing them all, and weighing where they disagree, is the critical handling the Nottingham scheme rewards.
Two things only the records showed
Verify references counted Wansink et al. (2012) as retracted; its record lists a correction and then a retraction notice, both in Preventive Medicine in 2018. Cite a source showed the Mertens meta-analysis with a published correction on record. Reading the correction explains it: the sample included four observations from a 2012 paper by Shu and colleagues that has since been retracted, as well as erroneous values and coding errors, and the overall effect moved from d = 0.45 to 0.43 ("Correction for Mertens et al.," 2022). A strong source can carry a weak one inside it.
One caution: Broers et al. (2017) is nine years old. Under a brief that asks for the last five years, it needs your tutor's agreement or a newer review beside it.
How do I find better sources with EdCitation?
Find sources is the best place to start upgrading a list: it filters about 300 million published works by kind of source, years and times cited before you read a title, and every result is a looked-up record, never a written one, which no chatbot can promise. It is free, with no account. Our two searches below ran on 2 October 2026, with every filter applied.
Search 1: in a journal, recent and most cited
nudging healthy food choices, unfiltered, returned 17,868 results. With In a journal ticked, 5 yrs under Year published and the sort set to Most cited, it returned 6,539: all 100 on the first page were journal articles (88) or reviews (12), dated 2021 to 2025. First came "A nudge in the right direction: the role of food choice architecture in changing populations' diets", Hannah Ensaff, Proceedings of The Nutrition Society, 2021, cited by 128. A journal is a sign of review, not proof.
Search 2: the reviews a field relies on
The same words, with Systematic review and Meta-analysis ticked under Kind of study and 100+ under Times cited, returned 63 results. Mertens et al. was there, cited by 749, and Broers et al., cited by 214. So was "23 Ways to Nudge", a conference paper: Kind of study goes by what a paper says about itself, so read before you keep.
Citing and checking the new list
Cite a source built every entry on the right from its DOI or ISBN, in APA 7, MLA 9, Chicago 18 author-date, Harvard, IEEE and Vancouver, and screened each article for retraction. From ISBN 9780300122237:
Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press.
We then pasted the five new entries and the Wansink paper into Verify references. Five came back verified, the book with "Title, author and year match a library catalogue record." The Wansink paper was counted as retracted: "The DOI resolves to this record and the title matches. This paper has since been retracted."
What is free and what is paid
Find sources, Cite a source, Verify references and Check your paper are free. References from a file, part of Pro at $8 a month, reads a whole paper, checks every reference, matches each citation to the list and recommends published sources for a reference that cannot stand. Max, at $24 a month, adds Theoretics QA and the Library. EdCitation never writes any part of your work: it finds, cites and checks sources, and every judgement in this guide stays yours. Plans are on the pricing page.
Quick questions
What is the CRAAP test?
A checklist from the Meriam Library at California State University, Chico: currency, relevance, authority, accuracy and purpose, each with questions to ask of a source.
What makes a good academic source?
One that suits the claim it supports: peer reviewed for research findings, the original study where a finding is cited, recent enough for the field, and written by someone with authority on that question.
Is Wikipedia a good source for an essay?
As a starting point, yes; as a cited source, usually not. Follow its references to the scholarly works, read them and cite those.
How do I find better sources for my essay?
Search for the claim, then filter. EdCitation's free Find sources keeps results to journals, recent years, kinds of study and well-cited work.
Can EdCitation tell me whether a source is good enough?
It tells you whether a source exists as cited and whether it has been retracted, through Verify references. Whether it suits your argument is your judgement.
References
- Association of American Colleges and Universities. (n.d.-a). Information literacy VALUE rubric. https://www.aacu.org/initiatives/value-initiative/value-rubrics/value-rubrics-information-literacy
- Association of American Colleges and Universities. (n.d.-b). Written communication VALUE rubric. https://www.aacu.org/initiatives/value-initiative/value-rubrics/value-rubrics-written-communication
- Association of College and Research Libraries. (2016). Framework for information literacy for higher education. American Library Association. https://www.ala.org/acrl/standards/ilframework
- Bauer, J. M., Bietz, S., Rauber, J., & Reisch, L. A. (2021). Nudging healthier food choices in a cafeteria setting: A sequential multi-intervention field study. Appetite, 160, Article 105106. https://doi.org/10.1016/j.appet.2021.105106
- Broers, V. J. V., De Breucker, C., Van den Broucke, S., & Luminet, O. (2017). A systematic review and meta-analysis of the effectiveness of nudging to increase fruit and vegetable choice. European Journal of Public Health, 27(5), 912–920. https://doi.org/10.1093/eurpub/ckx085
- Caulfield, M. (2019, June 19). SIFT (the four moves). Hapgood. https://hapgood.us/2019/06/19/sift-the-four-moves/
- Correction for Mertens et al., The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains. (2022). Proceedings of the National Academy of Sciences, 119(19), Article e2204059119. https://doi.org/10.1073/pnas.2204059119
- Dudley Knox Library, Naval Postgraduate School. (2026, September 22). Generative AI. Citation guide. https://libguides.nps.edu/citation/generativeAI
- John M. Pfau Library, California State University, San Bernardino. (2026, July 30). Bias, agenda, & domain. Evaluating websites. https://libguides.csusb.edu/c.php?g=234654&p=1557752
- Lycoming College. (2026, September 16). Citing abstracts. Citation Station Online. https://lycoming.libguides.com/citation/abstracts
- Maier, M., Bartoš, F., Stanley, T. D., Shanks, D. R., Harris, A. J. L., & Wagenmakers, E.-J. (2022). No evidence for nudging after adjusting for publication bias. Proceedings of the National Academy of Sciences, 119(31), Article e2200300119. https://doi.org/10.1073/pnas.2200300119
- The Master's University. (2026, July 28). Identify primary, secondary, and tertiary sources. Identifying scholarly sources. https://masters.libguides.com/identifyscholarlysources
- Meriam Library, California State University, Chico. (2010). Evaluating information: Applying the CRAAP test [PDF]. https://library.csuchico.edu/sites/default/files/craap-test.pdf
- Mertens, S., Herberz, M., Hahnel, U. J. J., & Brosch, T. (2021). The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains. Proceedings of the National Academy of Sciences, 119(1), Article e2107346118. https://doi.org/10.1073/pnas.2107346118
- North Dakota State University Libraries. (2025, October 9). Finding and identifying original research articles in the sciences. https://ndsu.libguides.com/findingoriginalresearch
- Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press. https://openlibrary.org/books/OL10319463M
- University of Edinburgh School of Social and Political Science. (n.d.). Marking criteria. https://www.sps.ed.ac.uk/students/undergraduate/your-studies/assessment-regulations/marking-descriptors
- University of Michigan Library. (2026, April 29). Citing Wikipedia. Resources for using & editing Wikipedia. https://guides.lib.umich.edu/c.php?g=283187&p=6278154
- University of Nottingham. (n.d.). Marking criteria examples [PDF]. Quality Manual. https://www.nottingham.ac.uk/qualitymanual/2-documents/markingcriteriaexamples.pdf
- Wansink, B., Just, D. R., Payne, C. R., & Klinger, M. Z. (2012). Attractive names sustain increased vegetable intake in schools. Preventive Medicine, 55(4), 330–332. https://doi.org/10.1016/j.ypmed.2012.07.012 (Retraction published 2018, Preventive Medicine, 110, 116)