Your referencing style is set by your department, not by your subject and not by preference. Two psychology courses at different universities can require different styles, and a single university often runs several at once. Before you format a single reference, find the style guide your marker actually assesses against — everything below is a map, not a substitute for it.
How do I find out which referencing style my course uses?
Check the assignment brief first, then the module handbook, then your library's referencing guide, in that order. The brief overrides everything else: a module can depart from the departmental default for one assessment. If all three are silent, ask the module convenor rather than guessing — using a defensible style consistently and asking is a much smaller problem than using the wrong one confidently.
One trap catches people every year. "Harvard" is not a single published standard. It is a family of author-date styles, and universities publish their own variants that disagree about punctuation, italics and the ordering of elements. A Harvard guide from another institution will be close but not identical, and the differences are exactly what gets marked.
What actually differs between referencing styles?
Less than it appears. Nearly every style asks for the same facts about a source — who made it, when, what it is called, where it appeared, and how to find it again. What changes is the order those facts go in, the punctuation between them, what is italicised, and how the citation appears in your text. Once you can see a reference as a set of fields rather than a string to copy, converting between styles becomes mechanical.
The one genuinely structural difference is how a source is signalled in the body of your writing, and that splits the styles into two camps.
Author-date or numeric: the split that matters most
Author-date styles put a name and year in your sentence — (Smith, 2020) — and order the reference list alphabetically. Numeric styles put a number in your sentence — [1] or a superscript — and order the list by the sequence in which sources first appear. This decides how your reference list is sorted, so it is worth establishing before you write rather than after.
- Author-date: APA, Harvard variants, Chicago author-date.
- Author-page: MLA, which gives a name and page number but no year in text.
- Numeric: Vancouver, IEEE, and Chicago's notes-bibliography system via footnotes.
APA 7th edition
APA is author-date and dominates psychology, education, nursing and much of the social sciences. In-text citations carry the author surname and year, with a page number added for direct quotation. The reference list is alphabetical with a hanging indent.
The detail most often lost marks on is capitalisation. APA uses sentence case for the titles of articles, chapters and books — only the first word, the first word after a colon, and proper nouns are capitalised. Journal titles, by contrast, keep title case and are italicised. DOIs are given as full links in the form https://doi.org/10.xxxx.
MLA 9th edition
MLA is used across literature, languages and the humanities, and cites by author and page rather than author and year — (Smith 42), with no comma. Its reference list is headed Works Cited.
MLA 9 is built on the container system: a source sits inside a container (a journal, an anthology, a streaming service, a website), which may itself sit inside a larger one. Rather than memorising a template per source type, you fill in nine core elements in a fixed order and omit any that do not apply. This is why MLA handles unusual sources — a podcast episode, a tweet, a video game — more gracefully than styles built around per-type templates.
Harvard
Harvard is an author-date approach rather than a single authority, which is why there is no definitive Harvard manual to buy. Your institution's library guide is the authority, and it is the version you should follow to the letter.
In practice a Harvard reference looks close to APA, and the two are frequently confused. Common divergences include whether the year sits in brackets, whether article titles take single quotation marks, and whether the place of publication is required for books. If you are converting a document from APA to your university's Harvard, check those three points first — or have the reference list checked against your institution's guide.
Chicago
Chicago, now in its 18th edition, is unusual in offering two complete systems, and you must pick one and stay in it. Notes-bibliography uses superscript numbers pointing to footnotes or endnotes, with a bibliography at the end, and is standard in history and the arts. Author-date works much like APA and is more common in the sciences and social sciences.
In the notes system, the first note for a source is given in full and later notes are shortened to author, short title and page. Ibid. is now discouraged in favour of the short form, since it breaks whenever notes are reordered during editing.
Vancouver
Vancouver is the numeric style of medicine and the health sciences. Sources are numbered in the order they first appear in your text, that number is reused every time the source recurs, and the reference list follows the same numeric order rather than alphabetical order.
Vancouver references use abbreviated journal titles taken from the NLM catalogue, and list up to six authors before using et al. Because numbering follows first appearance, inserting a new source into an existing draft renumbers everything after it — which is the single strongest argument for using a reference manager on a long Vancouver document.
What about OSCOLA, AGLC, IEEE and the rest?
Several disciplines run their own styles, and where one exists it usually takes precedence over the general styles above. Law is the clearest case: OSCOLA in the UK and AGLC in Australia both use footnotes and have detailed conventions for citing cases and legislation that no general style covers adequately. IEEE is the numeric standard across engineering and computer science. CSE appears in the biological sciences, ACS in chemistry.
What does a reference actually need to contain?
Enough for a reader to find the exact source you used. In practice that means capturing these fields as you read, in whatever order your style later wants them:
- Who is responsible — author, editor, or the organisation where no person is named.
- When it was published, and for web sources when you accessed it if the style asks.
- What it is called — the title of the item itself.
- Where it appeared — journal, book, website, conference proceedings.
- How to locate it — volume, issue, page range, DOI or URL.
Recording these at the point of reading, rather than reconstructing them the night before submission, removes most of the pain from referencing. It is also what an annotated bibliography is built from. It is also the only reliable defence against accidentally citing a source you never actually read.
How do I cite a source with no author or no date?
Never invent either. Where no individual author is named, the organisation responsible becomes the author — a government department, a charity, a company. Where a source genuinely has no date, author-date styles use an abbreviation in place of the year: n.d. in APA and most Harvard variants.
If a source has no identifiable author, no date and no publisher, that is worth treating as information about its reliability rather than a formatting problem to solve. A source you cannot attribute is usually a source you should not be leaning on.
Should I use a reference manager?
For anything longer than a short essay, yes. Zotero, Mendeley and EndNote all store the underlying fields once and generate the formatted output per style, which means switching styles becomes a menu choice rather than a rewrite. On a numeric style such as Vancouver, they also renumber automatically when you insert a source mid-draft.
They are not infallible. Metadata imported from a publisher or a database is frequently wrong — a missing subtitle, an author's initials mangled, a book listed as an article. Check what the manager produces against the source itself, particularly for chapters, reports and anything not from a major journal.
What loses marks most often?
Inconsistency, more than any individual formatting error. A reference list in flawless APA loses fewer marks than one that is half APA and half Harvard, because the second suggests the references were assembled from whatever each source's website offered rather than compiled deliberately.
- In-text citations that have no matching entry in the reference list, and entries with no corresponding citation.
- Direct quotations without a page number, where the style requires one.
- Secondary citations presented as if you read the original — if you read Smith describing Jones, cite it as such.
- Mixing two styles' conventions, most often APA capitalisation inside a Harvard reference list.
- URLs pasted in place of a full reference, with no author, date or title.
A working method
Establish the style before you start, from the brief rather than from habit. Capture the five fields above for every source as you read it. Use a manager for anything substantial, and verify what it outputs. Then, before submitting, read the reference list against your in-text citations once in each direction — every citation should have an entry, and every entry should have a citation. That final check catches most of what markers actually deduct for. If a reference list has already grown past the point of untangling it by hand, referencing support is the narrowest possible thing to ask for assistance with.