Is it plagiarism (or bad practice) to cite reviews instead of source material directly? [duplicate]

Question

My first source of information when preparing a manuscript are reviews. If up-to-date and comprehensive, I feel I can create most of my introduction by rephrasing and citing them. Nevertheless, reviews are based on citing other papers and, with the deluge of literature published these days, finding the source material for a single claim sometimes becomes a never-ending and time consuming play with a Matryoshka doll.

For example, in this review by Misteli (2007), the following claim is made and supported citing three papers.

The recent development of methods to probe the physical association of genome regions in a unbiased and genome-wide scale should lead to rapid progress in our still-rudimentary understanding of the functional significance of chromatin loops (Simonis et al., 2006; Wurtele and Chartrand, 2006;Zhao et al., 2006).

I consider relevant to a point in my article that there are new roads being opened to study “the functional significance of chromatin loops”, To do so, I have three options:

1. Rephrase the sentence and cite Misteli, 2007, acknowledging that I extracted the citation from the review and depriving the authors of the source material of a citation;
2. Rephrase the sentence and cite Simonis et al., 2006; Wurtele and Chartrand, 2006;Zhao et al., 2006 without reading the three source papers, which is kind of deceiving but at least acknowledges the authors of the research and not the author of the review; or
3. Read the original articles, rephrase and cite Simonis et al., 2006; Wurtele and Chartrand, 2006;Zhao et al., 2006, which is the most “honest” option but incredibly time consuming.

Sometimes, if I decide to do 3, I may find myself spending a lot of time reading, only to write something that was already said in the review.

Which option you would abide for? Am I somehow plagiarising the reviewers or primary authors if using any of my strategies? In terms of ethics and good scientific conduct, which is the best option? How deep and when should I look for source material?

I am not asking about which version of an article to cite, like here. I assume that versions cited in reviews are the most up to date. I am asking when it is correct to cite the review or the primary source found in such review.

Note: My strategy up to this days is a mix of 1 and 3: search and cite the source material only if the paper is relevant for the core of my article and cite the review when filling the generals.

Answer 1

Yes and No.

A review is not (or at least should not be) just a list of papers in the field, it is a research paper in its own right. A review should provide some form of new structure or idea that come from reviewing the scientific output the field in question. Thus a review can be useful when either pointing at the results emerging from the review or when you wish to provide an example of where a throrough description of a field is provided. So citing a review should primaruily be for general purposes.

If you need to cite a specific detail provided by one or several references in the review you should turn to the cited articles and check them. There is, first of all, no guarantee that the citations are correct and citing something that is not correct propagates errors in publications. In addition, the cited work may not be all the relevant work supporting a point from your perspective so your sources may become biased. After all your point for writing is most likely not the same as for the one writing the review.

A review therefore has a function on its own for providing overview and synthesis which can be cited but if you need to cite details you must go to the original work and also not limit yourself to the reference list of the review.

EDIT: Just to answer the question in full: plagiarism is only if you copy the text with out quotes. If you copy the references you are not necessarily plagiarizing but you will then give the false impression that you have actually read, in part of fully, the articles concerned and that is in the realm of unethical behaviour. In the best case, it is unwise for the resons given above.

Answer 2

The answer is somewhat field-specific. A mathematician's perspective here.

When citing a result which is uncontroversial and undisputed and unique, you can generally assume that the citation is accurate, and cite the contents of the reference as stated in the review article. So, for instance, if you are reading "Review of geometry by A. Euclid who cites the statement In a right angled triangle, the square of the hypotenuse is equal to the sum of the squares of the other two sides." which he attributes to A theorem about triangles by J. Pythagoras, then you can go ahead and paraphrase the above formulation and cite Pythagoras (not mentioning Euclid).

There is a thin line between results which are "classical" and those which are "modern". If it so happened that Pythagoras proved his theorem long ago that it is taught in graduate-level courses (but you still want to cite it because it is not universally known) then it is OK to give the more up-to-date reference by Euclid. Probably best avoided if Pythagoras is still alive and kicking (and would be glad to get another citation). Very advisable if the original is difficult to get your hands on, and/or written in a foreign language.

In both cases, you should (of course!) look up the original reference and see if Euclid is not making stuff up. If you don't do this, you are putting your reputation in the hands of whoever wrote the survey, so you should ask yourself how much do you trust him. If we are talking about a well-established result, and Euclid is well known to be a serious mathematician, then more likely then not everything will be fine. In all other cases - exercise caution.

Whichever way you go, it is often also a good idea to mention that the reader who wishes to learn more about the topic might want to look up Euclid's survey paper.

Answer 3

It seems to me that you owe a debt to all four sets of authors. How about

"The recent development of methods to probe the physical association of genome regions in a unbiased and genome-wide scale should lead to rapid progress in our still-rudimentary understanding of the functional significance of chromatin loops (Misteli, 2007 based on Simonis et al., 2006; Wurtele and Chartrand, 2006;Zhao et al., 2006)."

Answer 4

It all boils down to practicality. How central is that fact to your paper, and how controversial may it be?

If it is very important, and not a well known or established fact, you better make sure it is right, and it is worth spending some time on it. If it is a side detail, don't get lost and move on, and in the few cases you may trust a wrong review, the damage won't be extensive.

The ultimate goal of a paper is to convince your reader of your research, do what it is needed to convince them of your goal.

It is also a good idea to get a feeling of the knowledge of the authors on the specific field. You can usually trust that they know their field, but be a bit more sceptical of reviews of the type [X] for [people not in X], in my experience, they may get a few side details wrong, or not be fully up to date.

Answer 5

As long as you cite the sources which are responsible for you holding the belief that you have, you aren't engaging in plagiarism.

Science suffers from a replication crisis because original claims get a lot of citations while work like reviews that aggregate results from multiple sources and replications of a study get less citations.

However in many fields citing reviews instead of original discoveries goes against the norms and reviewers might not like it.

Answer 6

Suppose the authors of a review article assert that the stuff they're reviewing poses new questions or suggests new hypotheses that impinge on your own work. Then, it could make sense to cite the review authors as the originators of the new questions or new hypotheses. Effectively, this is an common criterion for citing anything (whether a review or an original research paper): cite the originators of whatever you're referring to, whether you're referring to data or to ideas.

Your particular example is rather analogous to the difference between hearsay and eye-witness accounts. If you're anticipating rapid progress just because Misteli says so, then that's akin to hearsay. If you're anticipating rapid progress because you too have read Sominis, Wurele, Chartrand and Zhao, then that's akin to an eye-witness account (with you as the eye-witness). In other walks of life, eye witness accounts carry more weight than hearsay, but where it gets tricky for you is that the reason for you being an "eye-witness" is that Misteli led you to the scene. Your call!

Answer 7

Something not covered in the other answers is that plagiarism is the re-use of ideas/work (not necessarily quotes!) without attribution. This means, therefore, that it depends on what exactly you are citing, and where no attribution is given the assumption in the you are claiming it as your own work.

For instance, if it's a conclusion or summary that the review has added that you include as part of your paper/argument, then cite the review. If it's the content of the other papers that you include, cite the individual papers.

Answer 8

For my perspective as a computer science researcher.

You mention 3 cases in your question:

1. you can depend on the review and follow the author's opinion. In this case, you will not get the whole picture.
2. That is plagiarism.
3. You can check the 3 source abstract and read the relative papers to your study then rephrase based on your opinion.