The check runs in both directions
Cited but missing is the failure people expect. Listed but never cited is the one that actually appears more often, usually because a source was dropped from a draft and never removed from the list, and it reads to a marker as padding.
Both are found by reconciling text against list line by line. It is dull, it takes twenty minutes, and it is skipped under deadline pressure by nearly everyone, which is precisely why it is worth doing.
Then every entry gets retrieved
Through the route your marker would take, which is your own library rather than a general search. Anything that does not come back is replaced rather than left in place looking respectable.
This is also where invented entries are caught. A fabricated citation survives every automated check there is, because it is well-formed; it does not survive somebody trying to open it.
Order, punctuation and the awkward types
Alphabetisation with corporate authors and with names carrying prefixes. Hanging indents. Multiple works by the same author in the same year. Where the DOI goes and what form it takes.
None of it is hard, all of it is scored, and it is worth fixing once as a template rather than repeatedly as an irritation.
The digital object identifier deserves its own note. It is the shortest route your marker has to the source itself, and where one exists and has been left out you have quietly made checking your work harder than it needed to be. That is the opposite of what a reference list is for, and it is a thirty-second fix per entry.
Three steps, then the record keeps itself.
Send the draft and the list
Both, because half the check is a relationship between them rather than a property of either.
Get the reconciliation
Every mismatch in both directions, every entry retrieved, and the formatting corrected against the system your assignment names.
Work from the template
What comes back is reusable for the rest of the program, which is where most of the value sits.
Queries on record.
Why does an uncited entry matter?
Because it suggests the list was assembled rather than used, which is exactly the impression you want to avoid on a piece of work whose credibility rests on its evidence. It also raises a question the marker cannot resolve: did you read it and forget to cite it, or add it to make the list look fuller? Neither answer helps you.
How many references should a paper have?
Whatever the argument requires, unless your rubric states a number, in which case it states it for a reason and you should meet it. Chasing a count produces the padding problem above. If a required number feels impossible, that is usually a sign the search was too narrow rather than that the literature is absent.
Should every source be peer-reviewed?
Not necessarily, and the answer depends on what each one is doing. Professional guidance, government data and registry figures are often exactly right for context or for practice recommendation, and wrong as evidence of effect. What matters is that each type is used for what it can actually support and labelled as what it is.
What if two sources disagree?
Say so, and say which you find better supported and why. Disagreement in the literature is a finding rather than an inconvenience, and a paper that notices it and adjudicates reads as considerably stronger than one that cites only the side it preferred. This is frequently the difference between a mid band and the one above.
Can you build the list from scratch?
The list is built from the sources the argument needs, which means starting from the claims rather than from a target number. That is a search job as much as a formatting one, and it is usually done alongside finding the sources rather than after. A list assembled independently of the text tends to produce exactly the mismatches described above.
Elsewhere in the hall.
Send the assignment and the deadline.
A specialist credentialed in your own field reads what you actually have and replies within two hours with a plan and a price. The reading costs nothing, and if the honest answer is that we are not the right hall for it, you will get that instead.
