Communication, Latest, Research

387. Citing references well

Some recent experiences reviewing other people’s work made me think there would be value in a post about what to do, and what not to do, when citing references in academic writing.

There are several reasons why you need to cite references in academic papers, books and theses.

1. To show that the point you are making is supported by evidence

When you make a point, particularly if it is a bit surprising, critical readers will ask themselves, “What is your evidence for that?” The evidence could be somebody else’s research, new data or analysis that you are presenting in your current document, or it could be that there is no existing evidence that supports that particular point – you are just making a judgement call or speculating. If there is no evidence cited or provided, the reader should assume that it is judgement or speculation. Judgements and speculations are allowed, but it should be clear that that’s what they are, to help the reader make their own judgements about how much confidence they can have in what you have written.

2. To be fully transparent about the evidence you are relying on

Full transparency allows readers to check for themselves: what the original paper really says, whether it is from a reputable source and whether the research it reports was done well. This is part of the general need to be as honest as possible in any research.

When I’m reading or reviewing a paper and I doubt that a citation is an accurate reflection of the original, I do sometimes go and look at the original. If it’s not an accurate citation, that affects my view of the paper. I find it particularly easy to spot an inaccurate citation of one of my own papers, and that does happen sometimes.

Being able to judge the reputability of a source helps readers make judgements about how confident they can be in a result that is being cited. Non-peer-reviewed sources can be OK, but I am more cautious in checking them. I have no confidence in results published in predatory journals, and I’m very cautious about papers published in borderline journals like those published by MDPI. In my own writing, I don’t cite predatory journals at all, and I avoid citing MDPI journals.

When I was doing my PhD on the economics of weed control, I found that many papers cited Stern et al. (1959) as the original source of the idea of an “economic threshold”: a weed or pest density above which it is worth taking action to control the weed or pest. I initially went along with that and included a citation and a sentence that repeated the usual story. But then I thought I’d better check the original paper. I found that they did talk about a pest threshold concept but not the same concept as was talked about in the many subsequent papers that cited them. It was an important lesson for me and ever since then, I’ve tried hard never to cite a reference without looking at the original and checking what it really says.

3. To acknowledge the source of the information so that you are not plagiarising their work.

Plagiarism is a big deal in academia and is to be avoided at all times.

Within those reasons for citing other research, I’ve indicated some things that we should or should not do. Here are some more.

Try to cite original sources that provide solid evidence, rather than just a judgement or speculation. If judgements and speculations are all that is available to cite, say that’s what they are.

Cite the original paper, rather than citing someone who cites the original. An exception could be if it’s a review paper or a meta-analysis that has looked at lots of papers and drawn broad conclusions.

If you cite a paper in support of a point that could be context-specific, say which context the research was done in. For example, research on a particular crop, or on a particular soil type, or in a particular country, may or may not be transferable to a different crop, soil type or country, so spell out the context to avoid confusion. If the research was done in New South Wales, don’t indicate that its conclusion applies to all of Australia. If it was done for wheat crops, don’t indicate that its conclusion applies to all crops or all of agriculture.

When citing a reference to support a point, make sure that it really does support your point. I recently was reviewing a document in which three citations had been given to support a particular point. The point was along the lines of A is bigger than B, but I had my doubts that that was really true. When I went to check the original papers that had been cited, I found that one actually said that A is not bigger than B, a second one said that A is less than B, and the third one was for a different context and effectively said that C is bigger than D. You just can’t be that fast and loose with your referencing practices and have any credibility at all. I presume that the author did look at the original papers but was not careful enough with their reading or note-taking or something. It may have just been carelessness, but from an academic point of view, what they did was lying.

Be clear about exactly what point you are citing a reference to support. For example, consider this sentence:

“In 2022 corn farmers in the US on average used less nitrogen fertilizer than they had in 2021, which was partly a result of record high fertilizer prices (Johnston, 2023).”

Is the Johnston reference in the right place? If the point that you are citing Johnston for is that farmers used less fertilizer, then it is in the wrong place. You would need to move it next to the relevant point:

“In 2022 corn farmers in the US on average used less nitrogen fertilizer than they had in 2021 (Johnston, 2023), which was partly a result of record high fertilizer prices.”

In that case, it would be good to find a reference to support the second point too.

If the point that you are citing Johnston for is that the reduction in fertilizer use was due to high prices, then its location in the sentence is OK, although as a reader I might still wonder whether you put it there incorrectly and really meant to support the first part of the sentence. You could remove that ambiguity by splitting the two parts into separate sentences:

“In 2022 corn farmers in the US on average used less nitrogen fertilizer than they had in 2021. This was partly a result of record-high fertilizer prices (Johnston, 2023).”

I sometimes see people citing a paper that is in a similar area to the point that is being made but doesn’t actually support the point. In the above example, there might be a study that shows that farmers have responded to fertilizer prices historically, but doesn’t actually say that their reduction of fertilizer use in 2023 was due to high prices. The existing wording above is not suitable for that reference, but you could find a modified wording that was honest about what the paper did and still indicated that price probably played a role:

“In 2022 corn farmers in the US on average used less nitrogen fertilizer than they had in 2021. Past research has shown that farmers respond to changes in fertiliser prices (e.g., Johnston, 2023), so the record high fertilizer prices in 2022 (Smith, 2023) probably contributed to the low fertiliser use.”

In general, be specific about what the paper said and why it is relevant. An exception to that could be where you are listing a set of papers that did something similar. “Various papers have studied the impact of fertiliser prices on farmers’ demand for fertilisers (e.g., Smith, 2022; Jones et al., 2021; Williams, 2023).” In this situation where you are providing examples, include “e.g.,” at the start. If you don’t include “e.g.,” then you seem to be saying that each of these references said that “various papers have studied the impact of fertiliser prices on farmers’ demand for fertilisers”, which is quite different from providing examples of papers that have done so.

The references you cite should not all be old and not all new. All old means you aren’t up to date with the latest research in the area. All new means you are not recognising the seminal papers that came up with important knowledge (give credit where credit is due), and it means that you risk repeating work that was already done, or you risk falling into traps that were recognised long ago.

Most citations are not direct quotes from the original source, but some are. When making a direct quote, put it in quotation marks, and after the usual indicator in the text, include the page number(s) where the quote comes from; for example: “blah blah blah” (Pannell et al., 2021, p.34).


Stern, V.M., Van De Bosch, R.F., Hagan, R. and Kenneth, S. (1959). “The integrated control concept”, Hilgardia 92(2), 81-101.