Wikipedia talk:Identifying primary and secondary sources for biology articles

Peer reviewed

Hi DrCrissy: I'd suggest adding something about whether a journal / other source is peer-reviewed or not, and what that means. MeegsC (talk) 12:15, 27 April 2015 (UTC)

Discipline specific definitions

I welcome this essay, because I think that clarifying the use of the primary/secondary/tertiary terminology as applied to scientific disciplines is important.

I think the essay still leans too far towards concepts developed for disciplines like history, anthropology or sociology. In these disciplines, a piece of writing can be the "raw material", the "primary data". When is this the case for biology? Consider this sentence from the essay: A scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. Sure, and a single paper written by a historian discussing the authenticity of a historical document is a primary source for the authenticity of the document, but this isn't what is usually meant by a "primary source" in history: the primary source is the document; an evaluation of its authenticity or significance by a historian is a "secondary source".

What's important about sources is not whether they are secondary or whatever but whether they reliably represent a significant consensus within the topic area.

In biology, it's often the case in my experience that so-called tertiary sources are less reliable than others. General works and encyclopedias are necessarily written by people who don't have detailed expertise in all the areas covered. Hence they often just repeat, with various forms of paraphrase, what was written by others. Time and time again it can be shown that along with accurate information such sources repeat errors, which can only be detected and corrected by going back to the original source.

My own quick view is that the hierarchy in biology goes something like this (but of course in reality it's a continuum, not a set of discrete levels):

1. "Primary" in science = the "raw data" – specimens, measurements, genome sequences, chemical analyses, etc.
2. "Secondary single" in science = single reports and evaluations of such "raw data"; both experiments and evaluations (e.g. drawing species boundaries based on genomic differences) require confirmation
3. "Secondary review" in science = repeated reports and evaluations, showing the development of a consensus; monographs by subject experts are typically at this level
4. "Tertiary" in science = textbooks, encyclopedias, etc., collating information from other sources; the broader such sources are the more they should be treated sceptically; they are often far from the most reliable

As per WP:MEDRS, it's sources at my level 3 that we should most rely on; sources at my level 2 are fine for more objective information (e.g. basic descriptions of organisms, in terms of colour, size, etc.) but not ideal for more subjective information (e.g. the existence of a new genus or species). Peter coxhead (talk) 14:19, 27 April 2015 (UTC)

Hi Peter. Thanks very much for your input. My great motivation for writing this essay was seeing valuable edits being lost because they were tagged as being from a source considered to be primary in its entirety (usually an original paper) rather than considering the actual source in context. In your 4 levels, where would you put review-like statements made in the Introduction section of an original article?DrChrissy ^(talk) 14:32, 27 April 2015 (UTC)

@DrChrissy: I think this is answered in the exchange below between Jytdog and you. A good review section is at my level 3; a poor one is not. I understand why WP:MEDRS (and hence Jytdog, who's very active there) wants to downgrade review sections of papers; it makes sense for medical claims. But it's often far too strong for biology.

As a real example, I wrote some 25 articles on the Maurandyinae (e.g. Lophospermum). The definitive source, as always, is the most recent monograph, which for this tribe is Elisens, Wayne J. (1985), "Monograph of the Maurandyinae (Scrophulariaceae-Antirrhineae)", Systematic Botany Monographs, 5: 1–97, doi:10.2307/25027602, JSTOR 25027602. This is a mixture of reviews of other people's work, reports of Elisens' work, and Elisens' judgements. It doesn't make sense to try to separate these components. It doesn't make sense to classify scientific journal articles as "primary" and then discount them. Peter coxhead (talk) 19:31, 27 April 2015 (UTC)

Peter, I could not agree with you more.DrChrissy ^(talk) 19:49, 27 April 2015 (UTC)

this is very good actually! (misread the first time, sorry) only two quibbles. The "review" sections of research papers are written by the authors to tell a story that supports their hypothesis/results/conclusions. those are not good reviews, as they are pointy, and they should be avoided. and the statement in the essay suggesting content like ""The present study shows that substance X had an analgesic effect on cats, thereby supporting the review of Reference1"" is WP:SYN. Jytdog (talk) 15:11, 27 April 2015 (UTC)

Hi. Thanks for the input and the compliment. A good review section (I am meaning the Introduction") of a research article will present a balanced view of all sides of a question/issue etc. Obviously there will be degrees of quality and the success that a neutral point of view is presented. The "Discussion" is more likely to present material biased to supporting the researchers findings, but that does not make them untrue. It is up to us as editors to ensure the content we place into Wikipedia is neutral and balanced. I am not sure who you think is doing the WP:SYN. Is it the author of the research article? This is not a problem because the WP editor simply makes the statement "substance X had an analgesic effect on cats" and cites both the research article plus the review, or just the review article. (I prefer the former as this will be more recent).DrChrissy ^(talk) 16:55, 27 April 2015 (UTC)

overlaps somewhat with an existing essay: [Wikipedia:Identifying reliable sources (natural sciences)]] - not sure if you looked at that. Jytdog (talk) 15:14, 27 April 2015 (UTC)

Yes I did look at that, and in fact borrowed parts of it for this essay. It is also in the "See also" section. Please see my comment in the "Purpose" thread below regarding conciseness and particularly the emphasis on sources having both primary and secondary content in this essay.DrChrissy ^(talk) 16:58, 27 April 2015 (UTC)

Just a couple of points from my own experiences in vertebrate taxonomy and paleontology, to me Primary Sources are as you pointed out scientific journal papers, I would perhaps add a caveat about them being peer reviewed and not on the list of predator journals. For much information a combination of Primary Research Articles and Review Articles will give a more balanced view for biology. For my own publications I write both, my original descriptions of new genera and species are of course all Primary Publications, however my work which reviews for example all the fossil turtles of the Plio-Pleistocene is a review artical but it was still published in a journal, albeit as a monograph. My point here is that many scientists actively publish in the peer reviewed scientific journals no matter what the style of the publication, be it original work, review or commentary. Apart from one book chapter and a few smaller articles I only publish in peer reviewed scientific journals, so you should be wary of assuming that all scientific journal publications are what you call Primary Publications. cheers Faendalimas ^talk 01:22, 28 May 2015 (UTC)

Hi User:Faendalimas thanks for your input. I found your list of predatory journals extremely interesting. I have not heard of it before. Have you thought about suggesting it at one of the suitable sources discussion pages?

As for the discussion about primary material, I think we are in agreement. The term primary or secondary should be applied to the material (the paraphrased/quoted text) to be added to WP, not the entire journal. For example, in your papers describing a new species, material relating to your description of the new species would be primary. However, if in that same paper you have an Introduction and you make a statement that "All sub-species of the lion extant in England have a fifth leg" this is review statement made by an expert (you) and therefore it is secondary material. You could even have secondary and primary material in the same sentence, e.g. "All previously described sub-species of the lion extant in England have a fifth leg; we show here that a newly discovered subspecies also has a fifth leg". My point here comes from my experience of having sources questioned on the basis that because the material I added to WP was from a science journal, it had to be primary material and under those circumstances it was unacceptable. I hope this helps.DrChrissy ^(talk) 11:21, 28 May 2015 (UTC)

Purpose

It is not clear to me why additional guidance specific to biology is needed on top of WP:NOR. Rlendog (talk) 15:31, 27 April 2015 (UTC)

Hi - thanks for the input. One of the reasons I wrote the essay was to provide a concise summary. WP:NOR discusses several issues not relevant to the present essay and so a Biology editor has to wade theough this to get to the relevant information. I also wanted to expand on the way that a single source may offer both primary and secondary content. This is mentioned in a couple of places, but I think that message is lost elsewhere.DrChrissy ^(talk) 16:39, 27 April 2015 (UTC)

My answer would be that most of WP:PSTS which is a major part of WP:NOR is not directly relevant to scientific topics and muddles up key issues. For example, it says: An account of a traffic accident written by a witness is a primary source of information about the accident; similarly, a scientific paper documenting a new experiment conducted by the author is a primary source on the outcome of that experiment. To equate statements made by a witness based on their memory with data reported in a scientific paper is unsound, and misses the important issues in evaluating reports of scientific research. Peter coxhead (talk) 19:42, 27 April 2015 (UTC)

Thanks PeterDrChrissy ^(talk) 19:52, 27 April 2015 (UTC)

Then, just to play devil's advocate here, Peter, shouldn't that argument be made on the TP for OR? I'm not sure the distinction you are making is supported by community consensus. While I agree that primary reports for an event are not exactly the same creature as primary reports on research, our community consensus and policies on reliable sources continue to regard both types of primary reports as problematic for supporting any kind of uncertain claim, and both varieties are seen as being particularly susceptible to bias. Even though in principle I like the nuance that DC is trying to make here with regard to primary sources (for example, with regard to the fact that some section of such a source may contain review of previous research and thus an argument can be made that they are secondary in reference to any claims supported in this way), the fact remains that any content found in a primary source is likely to be influenced by the objectives of its author(s). This essay is an interesting exercise, but I think caution ought to be taken that it does not directly contradict our actual content guidelines, or it is unlikely to gain traction, especially as its nominal purpose is to speak to the complexities of sourcing content on biological topics and I don't yet see the argument for how this subject varies in any significant way from other empirical topics. Snow ^{let's rap} 00:58, 5 May 2015 (UTC)

Hi Snow, thanks very much for your input. I take your point about bias from the author, but this happens even in Cochrane reports - the author/s start with a hypothesis (it should be a null hypothesis), objectify the papers, but then discuss these as being "weak" or "strong". I was interested in your comment that the essay might be contradicting our content guidelines - please could you expand on that. Finally to your last sentence. I am a biologist. I have been dealing with biological content for many years so I feel familiar with this. I am happy to see this essay reach other empirical topics (and be renamed as a consequence), but I would not like to see the content expanded that much - having a brief, very focussed essay was my motivation for writing it.DrChrissy ^(talk) 10:34, 5 May 2015 (UTC)

Hi DrChrissy. Sorry, I missed this post in the mix when replying to Peter earlier. To address your question about where the current reading of the essay runs against the grain of present policy, I actually don't feel that it does; the concerns I raised were more about some of the points that were being discussed here with regard to primary-vs.-secondary sourcing (see my response to Peter bellow for a more detailed discussion of my perspectives in that regard. In truth, having reviewed the essay in more detail, I really begin to see it's value, especially in that your "peer-review papers can still make secondary claims" argument is really quite useful and makes an important distinction I'm not sure I've seen anyone else make on the project (though I expect someone has, at some point, when arguing for some source or another). That distinction is nuanced, so an essay addressing it could really help some editors trying to make a secondary sourcing argument for a claim found in a source that is peer-review, but when the claim itself is a repetition of facts established in other studies or broader consensus in a given field. But the point is also so applicable to other areas, that I wonder if it shouldn't be discussed on other policy talk pages, in order that it be reflected in policy discussion itself, in addition to you essay on the biological context in particular. Snow ^{let's rap} 02:08, 6 May 2015 (UTC)

@Snow Rise: In the first part of your comments above, you use the term "primary reports", and I would agree that these need to be treated with caution. But there's no simple "type of publication" rule to tell you what is a "primary report"; it requires editorial judgement. To give concrete examples, the system of classification used in all plant articles at present is the APG III system. This is based on a single journal article: Angiosperm Phylogeny Group (2009), "An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III", Botanical Journal of the Linnean Society, 161 (2): 105–121, doi:10.1111/j.1095-8339.2009.00996.x. It would be completely absurd to refuse to accept this as a reliable source for angiosperm classification on the grounds that it's in some way "primary". Similarly in writing Tulip, editors have relied very heavily on a single journal article for the currently accepted species and the classification: Christenhusz, Maarten J.M.; Govaerts, Rafaël; David, John C.; Hall, Tony; Borland, Katherine; Roberts, Penelope S.; Tuomisto, Anne; Buerki, Sven; Chase, Mark W.; Fay, Michael F. (2013), "Tiptoe through the tulips – cultural history, molecular phylogenetics and classification of Tulipa (Liliaceae)", Botanical Journal of the Linnean Society, 172 (3): 280–328, doi:10.1111/boj.12061. Again, it would be absurd not to rely on this paper, since it's the most definitive account currently available. I can't speak for all biology editors, but my experience is that in WP:PLANTS we have a clear idea of what are reliable sources for botanical articles, an idea which is much better (although not perfectly) captured by this essay than by WP:PSTS. Peter coxhead (talk) 15:09, 5 May 2015 (UTC)

Well, my inclination is that when there is general editorial consensus that a given source supports a straight-forward claim that is unambiguously accepted within a given field, it's perfectly valid for that purpose, regardless of whether or not it is primary. Many editors forget (or are utterly unaware of the distinction) that primary sources are acceptable under policy for non-controversial claims. So given a context where all editors are on the same page, its not an issue to use such a source in the way you describe above. However, that doesn't really change the fact the source remains primary. With regard to the examples given above, both are primary in that they represent the research as presented by those who conducted it, rather than secondary or tertiary sources reporting on/evaluating said research. I completely understand the "common sense" argument you are making here (and to a large degree support it), but those would still be considered primary sources and if any editor challenged your assertion that the claims made by the source were non-controversial, you would be compelled to provide a secondary source validating that claim, under the current reading of our policies. Contributing as I often do in similar areas, I am sympathetic to those who have to deal with these complications -- but again, as regards this particular essay, my advice is to make sure it cleaves as close as possible to the current reading of policy and community consensus on these matters, as it will need to if it's to have much influence. And that means acknowledging that first-party research is considered a primary source. DrChrissy makes a nuanced point in drawing attention to the fact that some sections (abstract, introduction, ect.) may in fact be secondary in that they reference established facts in a field rather than presenting the new research that is the focus of the paper. That's both true and a very clever observation I can't recall seeing elsewhere on the project. But the point is so subtle, I would expect it to lead to problems, since not every editor who raises an objection to a given claim is going to go into that article (or is even necessarily going to have access to it), in order to parse that distinction. Many will simply see it is a primary source and remove the content it supports on the general principle of WP:PSTS. That's where I see the real potential value in this essay; it would allow a shorthand method for illustrating the distinction discussed above -- which absolutely is, insofar as I'm concerned, consistent with WP:V/WP:RS. Snow ^{let's rap} 23:52, 5 May 2015 (UTC)

@Snow Rise: we agree about the potential value of this essay. However, I disagree strongly with some of your points:

• You linked WP:BURDEN as "you would be compelled to provide a secondary source validating that claim". Nowhere does WP:BURDEN say "secondary".
• With regard to the examples given above, both are primary in that they represent the research as presented by those who conducted it, rather than secondary or tertiary sources reporting on/evaluating said research. Not so. Both papers reach conclusions based on a wide variety of sources. The abstract of the APG III paper begins "A revised and updated classification for the families of flowering plants is provided. Many recent studies have yielded increasingly detailed evidence for the positions of formerly unplaced families, resulting in..." Christenhusz et al. evaluate their cladistic results in relation to previous studies, with remarks like "in support of previous molecular studies, the results suggest..." Both papers are mixtures of reports and evaluations of previous studies and original material. You say that this point (as made by DrChrissy) is so subtle, I would expect it to lead to problems, since not every editor who raises an objection to a given claim is going to go into that article. It's this kind of attitude that drives away scientifically informed editors, who actually understand the material they are adding to Wikipedia. An editor who can't be bothered to read the source doesn't deserve to be respected.

Sources need to be evaluated in relation to the claims they are used to support, not in relation to some artificial binary classification as "primary" or "secondary". If a single author publishes a paper arguing for a split of an existing species into two, then, given that such taxonomic decisions are highly subjective, we need to know that others accept this split, so further sources endorsing it are needed. If 10 authors of the calibre of Christenhusz et al. (four FLSs, from four countries) publish a classification of Tulipa based on a wide review of the literature and some additional original research, we don't need to wait until the classification is copied by others (although actually it has been). Peter coxhead (talk) 21:18, 6 May 2015 (UTC)

Peter, I'm not trying to argue that the community shouldn't adopt a more nuanced approach in these cases -- point in fact, I tried to make it clear that I see this as the most valuable element of the essay -- I'm just trying to put forward what, from my experience, I expect the practical stumbling blocks to be, and the fact is that many editors tend to classify sources in a strictly dichotomous fashion when it comes to the primary/non-primary divide. And realistically, many of those who might do so in the case of the type of sources we are discussing here would be doing so in good-faith and in a manner that is largely consistent with our policies as they currently read. You say An editor who can't be bothered to read the source doesn't deserve to be respected. But the fact of the matter is that for the average peer-review article, the vast majority of editors are unlikely to access to the source, outside perhaps the abstract. That means some well-intentioned (and even experienced) contributors -- for example, those responding to an RfC on such a matter -- may be inclined to accept the argument of another party who classifies the coverage as simply primary, if the paper itself is obviously first and foremost a work of primary research.

On a side note, the most common gripe that I've been hearing across this project in recent years -- ad nasueum, really -- is that this or that kind of valuable editor is being "driven away" by this or that "illogical standard" or element of "bureaucracy". But the Wikipedia approach is a consensus approach, and the trade off of that valuable aggregative model is that change can be a slow process at times, and sometimes we even just have to make our peace with consensus in a niche area that seems completely counter-intuitive to us, at least until we succeed in convincing our fellow editors that a different approach is called for. Those who don't have the patience and stamina to see that process through are of course in no way compelled to continue to contribute. But I think it's hyperbolic to suggest that they are in any sense being driven out because the local or broad consensus view on a particular aspect of editing does not at present agree with their outlook. As a scientifically informed contributor myself, I've never felt pressured or frustrated even remotely to the point where I felt I had to leave the project simply because I couldn't properly source (and thus had to leave out) reference to a fact that I nevertheless knew had gained substantial support amongst researchers. Even in cases where I felt that policy was being misapplied on technicality or misinterpretation, I felt no such compelling drive. Editors in every kind of area have to frequently deal with and accept that situation and those who can't resolve themselves to it were probably not going to have the capability to, regardless of the area they were contributing in. The fact of the matter is that sometimes even good ideas (like some of DC's here) need to be repeated a lot of times before they gain traction.

Getting back to the subject at hand, though, it's worth noting that nuance cuts both ways on this issue. I support the argument that review of another researcher's findings may at times support a claim requiring secondary sourcing even if the source is, nominally, a primary source first and foremost. But if the claim being supported is too similar or intertwined with the conclusions of the research which is the basis of the paper being cited, then things become complicated. Even under ideal conditions, it's probably always going to be an uphill battle make this "secondary claim within a primary source" kind of argument, because, after-all, the purpose of presenting these previous findings is to contextualize the present research, not to present an independent review of the previous findings, and that can be problematic. Especially in cases where the study roughly confirms the findings of previous research, I wouldn't expect it to be accepted as a secondary source, even if it technically cites another work. That is, if study A notes that Study B concluded X, and then A also goes on to conclude X, I wouldn't expect A to fly as a valid secondary source for X amongst editors, even though it technically made the claim in both a primary and secondary fashion. Also worth noting is that often peer-review works will reference the conclusions of other works without actually treating them as established results. When we use peer-review works as secondary sources for other research, we should be very careful to look at the weight and validity being granted to that research and whether it any sense reflects consensus in the field. In many cases papers will reference one-another, but they do so in a manner that regards the original research as merely suggestive of a possibility which informed their own approach, not accepted fact that has been established by consensus in the field. We should be certain that we reference and attribute such information accordingly. A reference to the findings of a particular piece of research in any independent paper is not necessarily a valid secondary source just because it is one stepped removed. There needs to be appropriate and thorough review, and control for bias from parties whose own work may be confirmed or debunked by the claim being put forth. Those are always going to be conflating factors in the circumstances we are discussing here and for those, and other, reasons, a case-by-case analysis may always be necessitated for such secondary-reference-from-primary-work citations.

But I can tell you one thing that certainly is not going to be accepted is the approach suggested in what you said here: "If a single author publishes a paper arguing for a split of an existing species into two, then, given that such taxonomic decisions are highly subjective, we need to know that others accept this split, so further sources endorsing it are needed. If 10 authors of the calibre of Christenhusz et al. (four FLSs, from four countries) publish a classification of Tulipa based on a wide review of the literature and some additional original research, we don't need to wait until the classification is copied by others." I'm afraid that's completely inconsistent with policy, being textbook WP:SYNTH. We simply don't get to say "Well these primary reports all suggest the same thing, and I really trust some of them, so that's just as good as a secondary source." No matter the number of primary sources converging on the same conclusion, we can't put them together to conclude scientific consensus. That is what secondary sources are for. Now the catch is that you mentioned a "wide review of the literature". To the extent that such a source is in fact a review of literature, it is perfectly acceptable as a secondary source (tertiary in some cases, even), but it is not a secondary source for its own original research, no matter how many different similar primary sources converge on the same conclusion. Snow ^{let's rap} 01:43, 7 May 2015 (UTC)

I suspect this discussion has run its course and we aren't going to reach agreement. My view remains that a case-by-case analysis is always needed in deciding whether a source is reliable for the information it is needed to support. Of course we don't decide to put together reports and draw conclusions. What we decide is what are reliable sources for information added to Wikipedia, as editors have correctly done in the case of the two examples I gave.

There can be no secondary source for its own original research. Others can replicate the research (but very, very rarely do), but this isn't a secondary source for the original research. Others can survey a range of research including the given piece of research and draw conclusions, but this is different. Your line of reasoning leads to the nonsensical view that if X reports that Y said Z then X is is a better source than Y for Z.

We do of course need 'secondary' sources to establish notability: the APG III system of classification would not be notable if it had not been so widely taken up; similarly the Christenhusz et al. classification of Tulipa. But such sources aren't the most reliable sources for the classifications, since they are just repeating, possibly with added errors, what the original publications said. The most reliable sources for the classifications are the original publications. Peter coxhead (talk) 09:00, 7 May 2015 (UTC)

Well, of course each individual source needs to be evaluated on its own merits relative to any claim it is used to support. I think you are conflating two very different issues here. I was speaking as to complications inherent in the specific scenario being posited here -- trying to create a rule of thumb that says peer-review and similar sources can be used as secondary sources when they reference other research. I'm pointing out that selling this to other editors under ideal circumstances is difficult and that there specific circumstances where it's just not likely to ever fly. I honestly don't mean to cause offense here, but it seems to me that many of your notions of what defines a source as primary or secondary are good deal away from where basic and overwhelming community consensus lays, to the point of confusion. For example you say "there can be no source for its own original research." Now the grammar's a little muddled there, so I'm not entirely sure what you are meaning to say, but if a claim is made in a peer review paper, that's a primary source, if another source reviews that original material, said source is secondary. You may find it nonsensical that "if X reports that Y said Z then X is is a better source than Y for Z" but that statement basically defines longstanding policy as regards the relative worth of primary and secondary sources. Primary sources cannot be used to support claims whose non-controversial status has been challenged, and primary sources are generally not allowed to be used in a manner which contravenes claims based on secondary sourcing, especially in cases involving scientific consensus. (See WP:V, WP:MEDRS).

As to notability, I'm not sure why you mention it here, as it's rather a separate issue from that which is being discussed, but I'm afraid you're interpretation of sourcing with regard to that policy is a bit off as well, at least as regards the scenario you propose. It doesn't matter how highly you regard Christenhusz (or any primary source) and it doesn't matter how broadly his classification is adopted in other primary sourcing, because sources establishing notability generally need to be secondary (WP:GNG). Your position that the original publications are the best for establishing notability of a topic (or for supporting claims within such topics) is not reflective of community consensus on these matters. With respect, I think perhaps you are confusing the factors that make a source valuable to researchers with regard to establishing an objective reality with those factors which make a source of value to our purposes here on Wikipedia for establishing verifiability.

With regard to this discussion having run its course, you're of course free to comment further or not, as your whim takes you, but I suspect discussion on this page has only just begin, if DrChrissy plans to continue work on the essay. This is not a content discussion, so there is no immediate issue to resolve which requires agreement between us. The only purpose or value of discussion of these issues is to possibly help DC iron out the wording of the essay, that it might be made to be consistent with (but also in some slight way expand or inform upon) existing policy, so that it may gain some wider community consensus. Seeing as Chrissy reached out for such perspectives, I'm sure they would just as soon this line of discussion be considered a live one, given that only two contributors have thus far put forward their perspectives on the matter. Snow ^{let's rap} 03:00, 8 May 2015 (UTC)

Thanks again for the input. The notion that a single source may have both primary source content and secondary source content is discussed somewhere eles on the project, but it is rather brief/sketchy and hidden in one of those long, very detailed guidelines/policies somewhere. I will dig it out if this becomes necessary. I concede this essay may cause problems, but only in that it will make editors more careful about there edits. Editors inserting material will be best to check whether their material is from a primary or secondary source. Editors reverting material will also have to check whether the material is from a primary or secondary source within the referenced material. I am hoping that this essay will help prevent editors seeing that inserted material is from, e.g. a research paper, designating this as being entirely primary material and deleting on that basis (very often an edit summary such as "primary source" is left). This is not really an additional problem - it is only what careful editors should already be doing, but this essay highlights the issue for biologists (at the moment).DrChrissy ^(talk) 10:48, 6 May 2015 (UTC)

Thanks for this excellent essay! I was just wondering myself whether it was appropriate to consider as secondary sources those sections of a research-presenting paper which do not themselves present research, particularly content from introductions or general discussions. As a nonexpert, I intuitively agree that "All previously described subspecies of lion in England have 5 legs" is a secondary source but one to be used with caution; authors don't need a special "secondary source permit" in order to make secondary observations, but a primary paper has (I imagine) less of an obligation to give a exhaustively balanced perspective on the state of a field.

It would be great to get this principle more generally validated by those in the WP community who have experience using peer-reviewed papers in a formal context. I look forward to seeing that happen! FourViolas (talk) 12:14, 1 August 2015 (UTC)

Thanks very much for the positive comments. I looked at the question you were asking, and I agree entirely. My concern is that some editors appear to make quite swathing deletions on the basis that a research paper is entirely a primary source when the Introduction and Discussion sections are effectively mini-reviews (secondary sources). The issue of presenting a balanced view in research articles is a little more difficult. Because scientists are encouraged to be extremely concise in their writing, there may be little or no mention of opposing evidence, although this will likely depend on how strong that evidence is. Let's not forget, scientists are writing for an expert audience and those experts will almost certainly be aware of any opposing hypothesis or evidence. It is different when we write for WP. We, as editors, are obliged to seek out and report opposing hypotheses/evidence and to report this in a way which makes a balanced article.DrChrissy ^(talk) 13:20, 1 August 2015 (UTC)

Great, thanks. So "background" sections are reliable secondary sources as far as what they discuss, and they're allowed to perform SYNTHESIS for us; but we mustn't assume they give a comprehensive or balanced view of the topic, and where they conflict with a real narrative review they should usually be set aside. FourViolas (talk) 22:56, 2 August 2015 (UTC)

Just a couple of things I notice as I read through this. Basically as I read it I am thinking of examples in taxonomy which agree or disagree with what your saying, how other fields may be affected I am not so sure.

First up authorship of papers, general acceptance etc. It is completely legitimate under the ICZN code to name and describe a new species based solely on re-interpretation of already published data. That is you do not have to present any new evidence just refer to it elsewhere, for this reason I earlier mentioned predatory journals as these are often taking peoples molecular phylogenies then naming all the clades. A rather despicable practice. However there are some legitimate publications also that may do a similar thing. The number of authors on a paper is also not relevant, how many authors on a paper depends on the project at hand, for my own species I have had everything from myself as sole author to up to 3 co-authors.

Under the code any correctly published name is valid from the moment it is published and technically should be followed. If we wait for it to be accepted, well I ask by who? CITES does an examination of all new names every year either rejecting them or using them for its own purposes, but they do not publish this. Or do we wait for another paper to use the name, in poorly studied taxa that may take years. In the end the decision on whether to adopt new names is going to have to be a case by case decision and it can only really be made by people who know the group in question. So anything these policies do must encourage editors not turn them away. I think the more people Wikipedia can attract that are specialists the better off Wikipedia will be. I do not deny that is a double edged sword, some specialists do fall into the "I am holier than thou" syndrome and that is a problem that must be dealt with also, but this is where consensus will have to prevail, with multiple people who understand the process of the various sciences are able to have input.

just a couple of points for you... cheers Faendalimas ^talk 19:58, 1 August 2015 (UTC)

This vs MEDRS?

I gave a quick read through the essay and like how it gives guidance on the likely primary or secondary nature of various parts of a biological paper. Two comments came to mind:

• As medicine is part applied biology, it would be good to distinguish early on in the essay how this essay differs from MEDRS. For instance, among WP:MED editors, fully secondary sources are much preferred and mining primary publications for secondary material is frowned upon.
• At least in genetics, meta-analysis methodology papers may be secondary for their source analyses, but usually also have primary statistical results to report--that is, the meta-analysis itself is primary and not always as objective as one would like.

--Mark viking (talk) 18:56, 11 May 2015 (UTC)

Hi - many thanks for the input. I take your point about WP:MEDRS and I will think about emphasising this. I like your point about meta-analysis being both primary and secondary. Do you think it would be fair to say that conclusions are secondary, but any data generated by the study are primary? I think I have been told this is excessive, but it does seem logical.DrChrissy ^(talk) 19:19, 11 May 2015 (UTC)

I agree that is a good way to describe meta-analyses: conclusions are usually secondary, but the data produced and any novel analysis methods introduced may be primary. --Mark viking (talk) 19:36, 11 May 2015 (UTC)

I think we need to be careful about that approach/wording. I would say rather that any affirmations of general trends explicitly reported in the studies the meta-analysis is based on can in many instances be considered secondary, but any conclusions forwarded as unique claims in the meta-analysis itself (be it the statistical data analysis itself or whatever) remain primary. If you follow me, anyway -- this is getting to be quite the recursive discussion! Snow ^{let's rap} 04:34, 12 May 2015 (UTC)

MEDRS prioritised systematic reviews. Which is excellent in areas where they exist and should be given highest priority. Rather than thinking in terms of primary-secondary, we should be thinking in terms of research findings and reviews, I think. Not to the point where the say that research findings should be excised from articles (as sometimes happens) but rather that we make it clear that research findings are often wrong (even if your methods and stats are perfect, you're going to expect to be wrong 1 time out of 20) and very often tentative (correlation v. causation, and all that). So we should give precedence to more recent studies, more systematic studies, and reviews. We should never present a single study to rebut many, especially when the single study is older. Things like MEDRS says. But with the understanding that the kind of systematic reviews that are often done for medical studies simply don't exist for a lot of biology. Guettarda (talk) 19:11, 14 May 2015 (UTC)

Definitions section

It's not useful to repeat the stock definitions (much of which have to do with news reporting); we already have a page for that. Rather, it should identify what these terms mean with regard to biology publications.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  22:06, 10 August 2015 (UTC)