Talk:Genetically modified crops/Archive 5

Archive 1

←

Archive 3

Archive 4

Archive 5

Archive 6

Archive 7

Panchin

From this

We performed a statistical reanalysis and review of experimental data presented in some of these studies and found that quite often in contradiction with the authors’ conclusions the data actually provides weak evidence of harm that cannot be differentiated from chance. In our opinion the problem of statistically unaccounted multiple comparisons has led to some of the most cited anti-genetically modified organism health claims in history

Kingofaces43 came up with this

but review of [studies reviewed by Domingo 2011 and Krinsky 2015] show the statistical methodologies were flawed and do not show evidence of harm or lack of substantial equivalence

Mention of statistical significance seems to have been added from left field, this was not a subject of either review.

The wording is so misleading it should be removed ASAP, but I will let someone else wade into those waters. You cannot use this paper to refute these reviews, and I am shocked by the attempt. If it is true that all 22 studies covered by Krimsky as well as those in Domingo have been shown to be invalid, the source added by King is not sufficient proof.

If you wish to use the paper to refute whatever studies the authors were referring to, great. I'm interested to see what they've uncovered. But to extrapolate from "some" unnamed studies to the present 'no study reviewed by Domingo or Krimsky actually showed harm, they were all flawed', without specifying what the authors were referring to, is not very encyclopedic, to put it mildly.

Is there a reason editors are hell bent on summarizing sources (shoddily, I may add) rather than giving readers an idea of their content by elaborating just a bit? petrarchan47คุก 09:22, 9 February 2016 (UTC)

I'll preface this with a reminder that WP:MEDRS is clear we do not engage in personal peer-review in terms of study inclusion criteria, etc.

Panchin is clear that they aren't just using a few studies as example, but making a statement on the state of the literature as a whole. You're cherrypicking pieces of text to change the meaning. Some studies had issues with multiple comparisons, while some simply didn't do the stats at all. There were a range of issues, so trying to say Panchin doesn't have an overall consistent message on statistical methodologies because of the word some is incorrect. Panchin states, "We reviewed published articles in which undesired and statistically significant differences between GMOs and conventional crops were reported." There is no qualifier that it was only a partial review. They just simply say this is their review of the literature. If there was some flaw in the methodology where a study should have been included, a later review will need to comment on that.

To call statistical significance coming from left field is laughable. It's the basis of all claims in this science. If a previous review neglected to check for basic stats in the papers, that's still a flaw with those reviews as well when a corrective review comes along later commenting on the exact same aspect of the literature. When a study has been found to be flawed in use of statistics, it doesn't matter who cited it previously, especially when such a citation makes no mention of the validity of the stats. Claims later found to be false are simply dealt with as such under WP:WEIGHT. The fact of the matter is that Panchin is extremely critical of claims of harm in this review.

At the end of the day, all studies in the review had major statistical flaws. Panchin also does not give any weight to the idea that there is legitimate evidence of harm in the literature as a whole in any of their introductory or conclusionary statements. That contradicts in multiple ways with Krimsky and Domingo who try to claim there is evidence of harm. If there is something actually incorrect in the content I added, I am willing to discuss that, but I would expect that to be with an understanding of what it means in the literature when studies claiming statistical significance are later found to not be significant. This is simply how we address flawed science if we make any mention of it at all. Kingofaces43 (talk) 16:33, 9 February 2016 (UTC)

Comment: After attempting to verify Kingofaces43's addition by reading the source, it seems entirely misrepresentative. (Please correct any ignorance displayed, I am a simple layperson attempting to verify what I read by following citations.)

The text, in citing Domingo and Krimsky, gives the impression that it is the studies mentioned in those reviews that are under consideration:

Some studies have claimed that genetically modified crops can cause harm,[199][200] but review of these studies show the statistical methodologies were flawed and do not show evidence of harm or lack of substantial equivalence.[201]

In fact, counting, it seems that Panchin re-examines exactly six studies. Where is the connection to any other studies? I found none. The title of the paper should more accurately be (my bolded addition): "Six Published GMO studies find no evidence of harm when corrected for multiple comparisons".

Panchin first spends some time supporting the choice of critical tool, proposing multiple comparisons errors, evaluated by applying Bonferroni correction. Looking into Bonferroni, it is described as a simple tool that can at times produce overly conservative results. Hmm.. Then I discover What’s wrong with Bonferroni adjustments, an interesting read, which says:

This paper advances the view, widely held by epidemiologists, that Bonferroni adjustments are, at best, unnecessary and, at worst, deleterious to sound statistical inference.

Now I'm left wondering about Panchin's a basic premise: Is Bonferroni valid? Am I to believe Wikipedia and accept that these six studies are fatally flawed because Panchin's analysis says so, when a little looking casts such doubt?

Finally, Panchin's "Conclusions" seems to consist in large part of an ideological-sounding plea for people to wake up and look at sheer numbers, even if a big percentage of those studies may in fact be invalid:

It has been argued that we might be underestimating the number of false-positive results in science in general due to bias, improper use of statistics, analysis of highly improbably hypothesis and other factors.[26] The suggested solution was to take preference for large studies or low-bias meta-analyses and to take into account the pre-study probabilities of a finding being true.

We argue that the totality of the evidence should be taken into account when drawing conclusion on GMO safety, instead of far-fetched evidence from single studies with a high risk of bias due to a large number of multiple comparisons. Perhaps more focus should be drawn to clear and relevant outcomes such as mortality rates, life expectancy or reproductive success.

Unfortunately, it takes just a single article claiming a mild difference between GM and non-GM products to stir the public debate and cause a long-lasting hysteria. ...

In the first quoted paragraph, Panchin cites Ioannidis' Why most published research findings are false. That and his ongoing work in this area appear to be widely accepted,and Panchin appears to be acknowledging that, and saying that even if a large percentage of, say, Nicolia's 1500+ studies, are false, we still have a lot of studies left. I'm confused by where I have ended up - how is this argument arrived at from Bonferroni correction applied to six studies?

This is what a straightforward attempt to verify the source by reading it resulted in. I don't think this new, limited review should be used, at least, without direct attribution, adequate background (at least mentioning Bonferroni), and in a very limited way. --Tsavage (talk) 18:29, 9 February 2016 (UTC)

• I went back and re-read Panchin (the whole source, not just cherry-picking passages and pasting them here with misleading bold font), to look specifically at the issues editors are raising here. I agree in part with Tsavage that Panchin specifically does a statistical analysis of six studies, as opposed to of all studies. As I read what Panchin concludes, the only way I can interpret it, as written, is that Panchin concludes that what they found in those six studies can be extrapolated to the literature as a whole. Editors may disagree about that extrapolation, but such disagreement is original research. We should word things on the page to attribute to Panchin the conclusion that the experimental evidence for problems is weak and that much of the presented data actually demonstrates the opposite, but we should not imply that Panchin actually analyzed every study. I think the wording KofA added does have a flaw in that regard, and I am going to correct it now. --Tryptofish (talk) 18:48, 9 February 2016 (UTC)

Tryptofish: As I read what Panchin concludes, the only way I can interpret it We shouldn't be interpreting it in the way you mean, if Panchin wanted to say, "Based on the re-examination of these six studies, for x-y-z reasons, we conclude that all such studies are flawed," he could have done that. Panchin did not do that. All the pieces to form your own conclusion may be provided, but it's not up to us at Wikipedia to form that conclusion and then commit it to content.

Also, while I find the general level of re-examining sources around here is completely beyond simply identifying reliable sources, looking into the central mechanism of a study, here, Bonferroni correction, is completely reasonable - if I read, "using a microscope, we found these things," and I don't know what a microscope is, I need to find out to make sense of the statement. If Bonferroni correction is a questionable method, then we have to take that into account. Is it? --Tsavage (talk) 18:59, 9 February 2016 (UTC)

No, Panchin say quite clearly what the title of their article says. We can argue about the exact amount to which they extrapolate, but in any case, I think that the revision that I just made of the sentence represents Panchin accurately. (Panchin say: "We emphasize that with over 1783 published articles on GMOs over the last 10 years it is expected that some of them should have reported undesired differences between GMOs and conventional crops even if no such differences exist in reality." That sounds like extrapolation from the 6 studies to me.) On the other hand, you have said: "Is Bonferroni valid? Am I to believe Wikipedia and accept that these six studies are fatally flawed because Panchin's analysis says so?" No, don't believe Wikipedia, believe Panchin, or at least believe that Panchin says that they conclude what they conclude. Wikipedia doesn't care whether editors disagree with Panchin. There is nothing wrong with attributing that conclusion to them. You have Panchin saying that Bonferroni correction is useful for their purposes, and that other source saying that it is often not necessary for many purposes, and you are doing OR when you propose that we should conclude that Panchin misused Bonferroni correction. You then say: "Finally, Panchin's "Conclusions" seems to consist in large part of an ideological-sounding plea for people to wake up and look at sheer numbers". It's a peer-reviewed review paper in a scientific journal. If you are painting it as ideological, it isn't the source that has a problem with ideology. --Tryptofish (talk) 19:19, 9 February 2016 (UTC)

I'm good with your update to the wording as that clears up the meaning better than what I initially had in mind. The main idea is spot on that Panchin selected studies as part of their review methodology representative of literature that can be deemed critical. They don't need to analyze every single study out there as there are always pre-screenings that go on in conducting literature reviews as to whether particular studies will even be considered. That's a bit of an art sometimes, so if there is criticism on that part, we'll need to wait for newer reviews rather than editor criticism of it. We don't know all the inclusion details (rarely is this covered to everyone's satisfaction), but Panchin does portray these findings as representative of the literature. Kingofaces43 (talk) 20:29, 9 February 2016 (UTC)

It doesn't appear that your wording is an improvement, the content still seems unverifiable. Going by the source, this:

Some studies have claimed that genetically modified crops can cause harm,[199][200] but an analysis of such studies concluded that the statistical methodologies were often flawed, and that as a group they do not show evidence of harm or lack of substantial equivalence.[201]

...trying for neutral wording, should read...

Some studies have claimed that genetically modified crops can cause harm,[199][200]; an analysis of six such studies concluded that the statistical methodologies were flawed in the six studies examined.[201]

That is really all the source supports, with some redundancy to balance the association with the "some studies/Domingo/Krimsky" (and if Bonferroni correction is significantly controversial, then that should be mentioned, see the following...). It's still not a good sentence. In fact, the clean, neutral statement would be:

An analysis of six studies that claim that genetically modified crops can cause harm concluded that the statistical methodologies were flawed.[201]

Panchin say quite clearly what the title of their article says. We can argue about the exact amount to which they extrapolate. The title, "Published GMO studies find no evidence of harm when corrected for multiple comparisons" is misleading when it refers to just six studies. And there is no explicit extrapolation, what commonality would that even be based on, that they all suggest harm? Please quote where the extrapolation is established. If you can't quote it, then it's not there, it is an original conclusion based on what is there. A paraphrased summary does not arrive at new conclusions that aren't stated in the source.

you are doing OR when you propose that we should conclude that Panchin misused Bonferroni correction. Isn't this the point where, if Bonferroni correction is indeed controversial, we say, "Panchin was published on 14 Jan 2016, less than a month ago, there's no rush, it's reviewing old data, let's wait and see if the method is criticized"? It's not a small, quibbling point.

It's a peer-reviewed review paper in a scientific journal. If you are painting it as ideological, it isn't the source that has a problem with ideology. Fair enough, let's put aside Panchin's "a long-lasting hysteria," and just call it personal opinion. The conclusion is largely based on author's opinion about things not directly addressed by the study's actual work. --Tsavage (talk) 20:39, 9 February 2016 (UTC)

I suppose I could argue that "let's wait and see if the method is criticized" should also apply to Krimsky, but I'm not going to change your mind. I could support: An analysis of six studies that claimed that genetically modified crops can cause harm concluded that the statistical methodologies were so flawed that the data actually demonstrated the opposite, and that it is unlikely that there is really any published evidence for harmful health effects. --Tryptofish (talk) 22:07, 9 February 2016 (UTC)

Tryptofish: It would be helpful to the discussion if you could point out from the Panchin text how their review of six studies indicates that most or all other suggesting-harm studies are similarly flawed. --Tsavage (talk) 22:48, 9 February 2016 (UTC)

There's a very important distinction here. As an editor, I'm neutral as to whether or not "their review of six studies indicates that most or all other suggesting-harm studies are similarly flawed". So I'm not saying that. What I'm saying is that Panchin says that. Not that they are correct or incorrect, in terms of Absolute Truth^TM. Just that that's what the source says. So here are some quotes where they say it: "We emphasize that with over 1783 published articles on GMOs over the last 10 years it is expected that some of them should have reported undesired differences between GMOs and conventional crops even if no such differences exist in reality." "We argue that the totality of the evidence should be taken into account when drawing conclusion on GMO safety, instead of far-fetched evidence from single studies with a high risk of bias due to a large number of multiple comparisons." They are not saying that they know specifically that the statistics are flawed in studies that they did not analyze. But they are saying that they think that what they found in the six studies can be applied to an overall assessment of the literature. The wording currently on the page reflects that accurately, and really does not go beyond that. --Tryptofish (talk) 23:12, 9 February 2016 (UTC)

My objection is that, in this consensus statement debate, we're picking studies to try to piece together or shore up something that should be clear in one source, using a combination of multiple sources, and working them in to Wikipedia sourcing rules. Panchin appears to be doing essentially the same thing, they believe we should pay attention to the overall amount of research, and not focus on single studies, and they have found a novel way to support that contention, one which has been accepted in a peer-reviewed journal. How we use Panchin is the issue.

If editors believe this is a useful addition, we should make it clear what Panchin is about, not just summarize it in a few words, without establishing relative weight compared to other papers.

We should acknowledge: The Bonferroni correction is conservative yet justified [6] in such cases as those discussed below because virtually any difference between GMOs and conventional crops is presented as a cause for concern, a large number of tests are carried out without a preplanned hypothesis on what these differences might be and false positive errors have important social consequences. They're apparently applying a controversial method ("criticised as deleterious to sound statistical judgment," per their cited source at [6]^[1]) that is more likely to find no significance, and proposing a justification for that approach.

I am not trying to argue the science, my concern is the use of sources in this article. I would HOPE that what I read here, in Wikipedia's words, is a straightforward, balanced, many-eyes distillation of the sources, and not something that I have to painfully parse on my own. When we're covering scientific opinions, like Panchin and Krimsky, we should present them with enough context to give general readers a reasonable idea of what they're actually about and based on. --Tsavage (talk) 01:55, 10 February 2016 (UTC)

You've been arguing the science quite a bit now. It needs to stop as it's only causing talk page bloat on things that cannot contribute to WP:CONSENSUS. You're trying to call one of the standard methods for accounting for multiple comparisons, a concept in any introductory statistics course, controversial. That's personal peer-review, which has already been mentioned as something we cannot do as editors (otherwise I could easily rip apart Domingo and Krimsky as rubbish). That's especially when it appears this introductory concept is new to you based on your comments.

(Scientist hat on for a minute) Proper accounting of multiple comparisons, power, sample size, etc. is something much more complex when you actually dig into the literature with fluency that casting doubt over a study by vastly oversimplifying multiple corrections as controversial is just plain improper. You'll usually get papers rejected in most disciplines for not accounting for multiple comparisons, and the ongoing discussion on interpreting those adjustments is on properly designing experiments in the first place so you aren't data dredging or have too little statistical power, not that the adjustments themselves are improper. This however, is not the place to discuss that.</hat>

Panchin says what it says, and you cannot engage in original research to make it seem like the paper only focused on the six studies while ignoring their overall conclusions, was only focused on the use of a Bonferonni correction, or even cherrypick Panchin's mention of overemphasizing single primary research studies in an entirely different context. This is part of a consistent string of misrepresenting sources now that needs to stop.

In the end, Tryptofish has it. Panchin is currently accurately reflected in the current text, regardless of your personal critique of the study. We summarize what they said they did and what their overall conclusions were. In this article, we aren't going to go into detailed description of every single study out there, and we aren't going to go into an intro to basic stats on this page. If someone wants that more in-depth information not suited for an article, they can read the paper itself after seeing its conclusions or take a stats course. Kingofaces43 (talk) 03:18, 10 February 2016 (UTC)

Kingofaces43: Continuing to argue for scientific consensus wording, bringing in a new load of sources on top of the first 18, seems like the root cause of Talk page bloat atm. You haven't actually addressed my comments. In any case, I've said my piece on this. --Tsavage (talk) 03:36, 10 February 2016 (UTC)

Improved wording for Panchin

From this:

Some studies have claimed that genetically modified crops can cause harm,[199][200] but an analysis of such studies concluded that the statistical methodologies were often flawed, and that as a group they do not show evidence of harm or lack of substantial equivalence.[201]

...to a more descriptive, straightforward, neutrally informative description...

Some studies have claimed that genetically modified crops can cause harm.[199][200] In Critical Reviews in Biotechnology, a 2016 review presents statistical reanalysis of the data from several of these studies, finding that "the data actually provides weak evidence of harm that cannot be differentiated from chance," and that conclusions on GMO safety should be drawn from "the totality of the evidence ... instead of far-fetched evidence from single studies."[201]

Should be as a separate paragraph, with coverage of Domingo and Krimsky (as cited in [199],]200]). --Tsavage (talk) 14:17, 10 February 2016 (UTC)

Here's an edit. Lfstevens (talk) 16:49, 10 February 2016 (UTC)

Some studies claimed that genetically modified crops can cause harm.[199][200] A 2016 review reanalyzed the data from several prominent studies including Seralini, finding that "the data actually provides weak evidence of harm that cannot be differentiated from chance" and that conclusions on GMO safety should be drawn from "the totality of the evidence ... instead of far-fetched evidence from single studies."[201]

Yes, imo, good edit for readability, and a distinct improvement with mention of Seralini, which gives general readers more high-level context, and better indicates the substance of the Panchin review, while keeping things at a summary level. (Maybe "such" to replace "prominent" for clarity, and, were all of those studies prominent?) Now, to do the same with Dominogo and Krimsky. (For the record, while I think the Panchin concluding argument makes sense as one opinion, the paper itself seems like a designer support piece to assist the pro-biotech case, which if so doesn't make it a bad thing, just an inherently biased one.) --Tsavage (talk) 17:29, 10 February 2016 (UTC)

I'm fine with Lfstevens' version. Thanks! And that wasn't so hard, was it? It's much better to propose better wording than to try to argue endlessly, as above, about who is "right" in principle. --Tryptofish (talk) 19:05, 10 February 2016 (UTC)

Speaking for myself, the discussion isn't about being "right" in principle, it is about weight of the source. --Tsavage (talk) 20:01, 10 February 2016 (UTC)

The general ideas are good, but let's remember that we don't do unneeded attribution per MOS:QUOTE and save that for things that cannot be said in Wikipedia's voice. We should try to avoid using quotes for higher quality sources like this since we usually just reflect them in Wikipedia's voice. I've seen a lot of insistence of quotes lately, but that is something left typically for much lower quality sources that are only reliable as opinions.

Picking out individual quotes even with the best of intentions can change meaning too. The actual intent of "provides weak evidence of harm that cannot be differentiated from chance" from the abstract is a little ambiguous, whereas the similarly worded second sentence of the conclusions is actually worded a bit stronger. The point constantly driven home in the article is that claims made by the studies are not supported or invalidated due to improper statistical methodology. That's stronger than just weakening the evidence, and the article isn't just about multiple comparisons as the only statistical flaws, so we need to be careful on wording. That's why the broader description in Tryptofish's most recent version lines up better with the source than Lfstevens' first quote. The current version also includes Panchin's synopsis on on evidence of harm and substantial equivalence, which isn't included in Lfstevens proposal.

I do however like the second quote on "totality of evidence . . ." That could even be changed to a summary of the first and second paragraphs of the conclusions (most likely as a followup sentence replacing the quote). I'd have to think about that a bit more though, so I'd be perfectly fine going with that quote for now. So something like, Some studies have claimed that genetically modified crops can cause harm. A 2016 review reanalyzed the data of such studies finding that the statistical methodologies were often flawed, that as a group they do not show evidence of harm or lack of substantial equivalence, and conclusions on GMO safety should be drawn from "the totality of the evidence ... instead of far-fetched evidence from single studies." What do folks think of that? Specific things like Seralini can get mention with respect to Panchin where they are focused on more since we're just doing a broad summary here. Kingofaces43 (talk) 05:16, 11 February 2016 (UTC)

I like that version even better. Just a few syntactic adjustments: Some studies have claimed that genetically modified crops can cause harm. A 2016 review that reanalyzed the data of some of these studies found that the statistical methodologies were often flawed, that as a group they do not show evidence of harm or lack of substantial equivalence, and that conclusions about GMO crop safety should be drawn from "the totality of the evidence... instead of far-fetched evidence from single studies." --Tryptofish (talk) 20:13, 11 February 2016 (UTC)

that as a group they do not show evidence of harm or lack of substantial equivalence I don't support this version. You've just circled back to giving Panchin far too much weight. They only tenuously connect their six-study reanalysis to all other studies, it's mainly left up to the reader to make the connection, so it is hard to have Wikipedia state that connection directly. And where does substantial equivalence come from, it's not mentioned in Panchin, and it generally refers to a regulatory definition, so using it otherwise seems misleading?

If you want to insist on trying to use Panchin to negate ALL negative GMO findings, then we're back to looking more closely at the strength of the study, looking for citations, reviews, critiques. Negating ALL of a diverse group of studies is an exceptional-sounding claim, so in fact it probably requires multiple sources. --Tsavage (talk) 01:23, 12 February 2016 (UTC)

Tweaks. I think the "they" refers only to the reanalyzed studies, not to the group as a whole. Lfstevens (talk) 02:22, 12 February 2016 (UTC)
Some studies have claimed that genetically modified crops can cause harm. A 2016 review that reanalyzed the data of some of these studies found that their statistical methodologies were flawed, that as a group they do not show evidence neither of harm nor lack of substantial equivalence and that conclusions about GMO crop safety should be drawn from "the totality of the evidence... instead of far-fetched evidence from single studies."

Such a difference a word or two can make - why say something less clearly, when we already had clear wording in the pre-last-Kingofaces43 Lfstevens version?

Panchin make clear what they're trying to support with their six-study review. Let's avoid the OR of nuancing degrees of...extrapolation, as Tryptofish put it, and either say what the study did as far as actual research, or say what conclusion it wants to infer from that research. The latter, that therefore ALL studies are (likely) invalid, is an exceptional claim. The former is that they looked at six studies.

The use of quotes in the previous version is consistent with covering controversial material. Use of the Bonferroni method is controversial, made clear even in the sources Panchin cites: Despite the widespread use of the Bonferroni method, there has been continuing controversy regarding its use. Hence, there are those who believe no correction should ever be made13 and those who consider correction should be mandatory.14,15[2] If you read Panchin, they go so far as to cite "important social consequences" in justifying their choice of analytical tool. Using quotes is a prudent and recommended way to avoid ambiguity with a tricky-enough source. --Tsavage (talk) 07:10, 12 February 2016 (UTC)

You've been warned multiple times to stop your personal peer-review of the paper. It needs to stop, especially when you are personally portraying the article as controversial for a basic method you've indicated you are not familiar with and severely overgeneralizing what the actual dispute is in the literature on multiple comparisons. You've been zeroing in on the words controversial and six studies in that process and cherrypicking as a result of that. As mentioned before, this cannot contribute to WP:CONSENSUS no more than if I tried the same thing in other sources. Kingofaces43 (talk) 15:49, 12 February 2016 (UTC)

Kingofaces43: Please stop framing polite discussion of content as some sort of disruptive action; please stop saying, "it needs to stop," and directly address content and the substance of comments. I AM familiar with Bonferroni correction from attempting to verify our Panchin content in sources. Panchin present reasoning for their choice of Bonferroni, with citations, I read that, and followed to sources, and even to sources within those sources.

Wikipedia relies on plainly evident sourcing, and one thing that is plainly evident in sources is that Bonferroni correction is unquestionably controversial (for example, that statement could be added to Bonferroni correction, solidly sourced). Also plainly evident in the primary source: Panchin reanalyzed six studies. The minimal hypothesis they present to connect the six studies to all the rest is an appeal to common sense extrapolation, based on using the Bonferroni method to find no significance in their sample.

No cherrypicking involved, just a straightforward reading of the source. Critical Reviews in Biotechnology published it as a critical review, but that does not make it a mainstream or widely accepted review.

And a claim that ALL such studies are invalid certainly seems exceptional, and requires multiple sources. Per policy. --Tsavage (talk) 17:39, 12 February 2016 (UTC)

I am willing to step back from having any "extrapolation" in Wikipedia's voice. Please let me suggest: Some studies have claimed that genetically modified crops can cause harm. A 2016 review that reanalyzed the data of some of these studies found that their statistical methodologies were flawed, that they demonstrated neither harm nor lack of substantial equivalence, and that conclusions about GMO crop safety should be drawn from "the totality of the evidence... instead of far-fetched evidence from single studies." --Tryptofish (talk) 01:38, 13 February 2016 (UTC)

I maintain that the Lfstevens edit is better, and versions that you, and Kingofaces43, propose, give far too much weight to Panchin. If we're not talking about their tenuous and exceptional claim to invalidate ALL research that suggests GMO effects, then we are just talking about reanalysis of six studies using a controversial method - journal publication doesn't automatically support us repeating whatever a study's authors have to say.

At some point, this type of protracted and literal negotiation over small pieces of content stops being about the content itself, and basic content policy like neutral weight and tone, and overall balance, and becomes mainly about editor preference for interpretation, which is OR and wrong.

If we want to bring in Panchin, a pretty well unrecognized study, and use it to offset Domingo and Krimsky, we should not start by giving Panchin relatively detailed coverage, while relegating, for example, an actual review study like Domingo, to "some studies" status as is proposed. --Tsavage (talk) 02:26, 13 February 2016 (UTC)

I hear you about the undesirability of protracted negotiation, and I share your desire to get past it. I do, however, think that you are misunderstanding the sources when you treat Domingo and Krimsky as having greater relative weight and Panchin as having less. Panchin is just as much a peer-reviewed scientific paper as Domingo (not Krimsky), and Panchin is more recent than Domingo. (In fact, it apparently hasn't yet come out in print, only online, which is a simple explanation why there hasn't been much response published yet. And if later reliable sources call Panchin into question, I'll agree readily to giving those greater weight.) Absent other sources evaluating Panchin (and editor opinion about statistical methods does not count), and given Panchin's revision of Domingo's conclusions, it's editorially correct to treat Panchin as better reflecting present-day knowledge. Anyway, I'm willing to go along with any number of proposed wordings. I continue to support what I proposed, I already said that I could support Lfstevens' first proposal, and I can support Lfstevens' second proposal with a bit of copyediting. So I'll go along with any of those, depending on what other editors say. --Tryptofish (talk) 19:20, 13 February 2016 (UTC)

It appears we've gone circular again. First, I said we can't give Domingo and Krimsky no weight, i.e. no direct mention. And if we want to use Panchin to negate essentially ALL studies that find potential harm, then that's an exceptional claim, and requires multiple studies to say something similar. If we don't want to make that exceptional claim, then Panchin is a review of only six studies, and that only goes so far.

The fact that their method is mainstream controversial is only more reason for caution, and it is perfectly reasonable to take that into account. We don't simply go, "It's peer-reviewed and its newer, so let's just paraphrase what we want," we do read the source, and if something front and center in the source, like their evaluation method, is controversial, we should take that into account. This is not equivalent to, say, speculative questioning of funding sources, for example, but it seems to be the same sort of consideration as the recent MEDRS issue, where, with a reliably sourced basis, country of origin can call reliability into question. --Tsavage (talk) 20:19, 13 February 2016 (UTC)

I cite Domingo and Krimsky in the sentence I propose for the lead, and I have no objection to citing them in greater detail in the lower section. Beyond that, we are likely to just disagree about how the science works. So, bottom line: I'll go along with any of those three versions that I just listed. --Tryptofish (talk) 20:28, 13 February 2016 (UTC)

disagree about how the science works If I've made fundamental errors in my reasoning above, please inform me. I don't want to waste anyone's time, least of all, my own, with badly informed discussion.

I acknowledge your support for any of the three versions, including Lfstevens. My request is simply out of interest and for future use. --Tsavage (talk) 15:39, 14 February 2016 (UTC)

Great, thanks very much. I guess there's a lot that I could potentially try to explain, but I think the point that I care the most about is the following. Panchin says that they use a particular statistical method that they chose for specific reasons. It's a peer-reviewed reliable source in a scholarly journal. You have pointed out another source, that says that this same statistical method has significant limitations in many circumstances. But that source does not discuss the Panchin study by name, and I don't think that source says that the statistical method is never valid. You seem to be asking us to conclude that the statistical method is inappropriate for the specific use that Panchin made of it. Therefore, it seems to me that you are asking that editors set aside Panchin's own conclusions and instead determine that the criticisms of the other source should be applied to Panchin's methods. I think that's a textbook example of a WP:SYNTH violation. --Tryptofish (talk) 23:55, 14 February 2016 (UTC)

Thanks. I was thinking of possible errors in referring to the science. As for Wikipedia policy, I don't see a problem:

• Panchin use a controversial method to arrive at their results.
• The controversiality is established, and discussed as such, in the general literature.

That's meta information, it requires no editor interpretation of the study itself, no SYNTH. And that suggests that we take into account that Panchin uses a controversial method. That's what I've been saying.

Bigger bottom line, this seems like an awful lot of discussion over how to use a single study that claims, based on a small sample, to negate ALL studies that may suggest unwanted GMO effects. It should be simpler. --Tsavage (talk) 00:00, 16 February 2016 (UTC)

Bottom line is that this is a lot of discussion to dedicate to your own personal WP:OR. You need to stop violating policy in this regard. You're still just glomming onto the word controversial to end up misrepresenting the actual discussion in the literature. Even applications of t-tests and assumptions of normality are "controversial" in statistics, but they are still standard practice and pretty much in the same boat as a standard multiple comparisons correction. This is not the place to get into personal interpretations of experimental design philosophy, so if you still want to have some sort of criticism of Panchin, you need sources directly discussing it in that context. Kingofaces43 (talk) 02:37, 16 February 2016 (UTC)

WP:OR states 'This policy of no original research does not apply to talk pages.' There is no policy violation unless OR content ends up in the article page.Dialectric (talk) 05:18, 16 February 2016 (UTC)

In any case, I can support any of three possible wording choices. And I think that we might as well get back to selecting what we will put on the page, as opposed to discussing meta issues. --Tryptofish (talk) 23:21, 16 February 2016 (UTC)

It's not a meta issue, it speaks directly to sources and verifiability. There's no glomming onto the word controversial to end up misrepresenting the actual discussion in the literature, and that's a pretty serious charge. When you present a study that offers a surprising and sweeping claim, an exceptional claim, editors should verify it carefully, which is what I did. Verifiability is the only policy issue here.

1. Panchin relies on a single method, Bonferroni correction, to negate the results of six studies, and argue that that should negate the results of ALL such (GMO) studies that show significant results, an exceptional claim, that editors want to use to make that point in content.

2. Multiple peer-reviewed sources indicate that the method Panchin use is controversial:

• Despite the widespread use of the Bonferroni method, there has been continuing controversy regarding its use.[3]
• Bonferroni adjustments are, at best, unnecessary and, at worst, deleterious to sound statistical inference.4,5[4]
• A number of procedures have been developed to deal with multiplicity, such as the Bonferroni correction, but there is continuing controversy regarding if and when these procedures should be used.[5]
• Use of the Bonferroni correction is a controversial issue[6]

These are not cherrypicked or quoted out of context, and there are more such sources. What we have is straightforward discovery through verifying by reading the source. It is not our place to speculate on the merits of Panchin's use of Bonferroni, only to note that it is a controversial method that all but assures non-significance.

We should not be arguing at length over something so clear and policy-based as an exceptional claim that is not supported in multiple sources, and that arrives at its results using a controversial method. --Tsavage (talk) 02:10, 17 February 2016 (UTC)

Yup, and at this point that's all personal peer-review and opinion, which we don't do even on talk pages per WP:OR and WP:MEDRS. Time to move on. If you want to change what we can and cannot do even if someone is an WP:EXPERT editor, this is not the place to do so. Kingofaces43 (talk) 04:30, 17 February 2016 (UTC)

I agree with KofA, because the fact that a statistical method is described as "controversial" is not the same thing as describing it as "invalid". Unless there is a reliable source saying that Panchin misused the method, we need to move on. And if a source emerges in the future saying that Panchin misused the method, I will be all in favor of changing what the page says. So please: let's see if we can select an improved wording. Again, there are three possible choices that I have said that I can support. --Tryptofish (talk) 23:59, 17 February 2016 (UTC)

This discussion seems to have gone quiet, but we really ought to resolve it. After re-reading the previous comments, I suggest: Some studies have claimed that genetically modified crops can cause harm. A 2016 review that reanalyzed the data of some of these studies found that their statistical methodologies were flawed and did not demonstrate harm or lack of substantial equivalence, and said that conclusions about GMO crop safety should be drawn from "the totality of the evidence... instead of far-fetched evidence from single studies." I'm thinking that we should cite the first sentence to the same six studies that were analyzed by Panchin, and the second sentence to Panchin. I reject the idea that this gives Panchin undue weight, and I think that we need to come to closure on this. --Tryptofish (talk) 23:47, 25 February 2016 (UTC)

Is this OK? --Tryptofish (talk) 00:16, 1 March 2016 (UTC)

I have not read the source yet, so I have no position at this time. But if we are to add that many sentences about Panchin, then it would be undue if we do not permit a similar number of sentences for Domingo and Krimsky each. Do you agree? --David Tornheim (talk) 03:23, 1 March 2016 (UTC)

@Tryptofish: I haven't just walked away from this, I have taken a pause. I acknowledge and appreciate that you are engaging in what appears to be an actual content discussion, which is excellent. At this point, though, your proposal still seems far from what seems reasonable, which I've tried to state clearly already. Basically:

Ff we want to include Panchin, we should do so in a proper context, framed by at least a description of Domingo, which contains what I assume considers much of the material Panchin is referring to. Assembling a paragraph summarizing the latest Domingo shouldn't be difficult. And we should not ignore the basis of the Panchin findings, which is using a controversial method. We can't claim a level of summary is required that precludes a certain amount of context, this isn't Twitter.

As it stands, content is already in the article, and your last wording, while more accurate and informative than what is there now, also gives more play to Panchin, without context, in a very brief section that doesn't develop the topic, which adds up to imo undue weight. Still, it's better than what is there now, so I don't have an objection to replacing it - imo as just one editor.

About re-arguing the sources: It has been mentioned here that we don't re-review sources, we just establish their reliability and use them. This is a simplistic view. The reliability of a source is a function of what it is specifically supporting, not the other way round, so it is necessary to examine and understand the source sufficiently to come to some conclusion for the intended use. When Panchin proposes to negate findings based on new analysis, of course we must examine and understand the nature of that analysis, and if it turns out that the methods and reasoning are questionable (based strictly on other sources0, of course that comes into play in determining reliability. We don't just say, it's in a peer-reviewed journal so it's fine, that's a starting point, we also need to understand what we're incorporating, for example, Panchin describe their method as "conservative yet justified." There's no OR in vetting a source. (At the moment, I don't have the links to a letter by Panchin to the publishing journal about one of the studies they review in this paper, and the reply from the study authors, both letters published in that journal, which illustrates the review method problem I've been mentioning - you can easily find it if you're interested.) --Tsavage (talk) 17:38, 1 March 2016 (UTC)

I'm good with Trypto's text as it's mostly just a slight tweak from my original edit, though I have preference for keeping the two clauses linked rather than separate sentences. We do need to be careful about encouraging standalone use of sources like Domingo and Krimksy. I went ahead and added it in as I can assume others would agree this is an improvement over our current status quo version. If not, we can go back to the text that's been in the article for about a month now. Kingofaces43 (talk) 18:10, 1 March 2016 (UTC)

I made a couple of changes in the text, as noted in the edit summary. Let's not overstate things. "Some" studies is accurately, six studies, let's say that. In particular, the study did not say it concluded anything about substantial equivalence, the clearest finding statement reads, "we conclude that the data presented in these articles does not provide any substantial evidence of GMO harm." --Tsavage (talk) 19:09, 1 March 2016 (UTC)

I'm fine with the wording as it is after Tsavage's revision. Because Panchin is a response to such earlier sources as Domingo, I believe the weight is balanced correctly. If there is a subsequent response to Panchin in the form of a letter to the journal, I would welcome seeing a link to it. --Tryptofish (talk) 21:56, 1 March 2016 (UTC)

FYI, substantial equivalence is mentioned in the first paragraph of the conclusion: "The presented articles suggesting possible harm of GMOs received high public attention. However, despite their claims, they actually weaken the evidence for the harm and lack of substantial equivalency of studied GMOs." Sunrise (talk) 05:20, 3 March 2016 (UTC)

"Despite their claims" is ambiguous - were the results refuted, or not, if so, it's not "despite," it's "falsely claimed." Panchin are editorializing, which is up to them in their conclusion, but we aren't bound to echo that. Extending our coverage to include substantial equivalency gives more weight to the "extrapolation" towards that "all" harm studies are false, which becomes an exceptional claim and requires multiple sources.

And we haven't given time see how the study is received. At least one of Panchin's findings was challenged by the study authors in a journal-published Response to Panchin's Letter on Statistics. From what I can gather, Panchin seem to gone further than simply apply missing correction as the study seems to indicate; the authors say, The multiple comparisons problem was considered and corrected when we analyzed the data in Table 2 and Table 3. Table 2 and 3 were considered as two separate tables, and thus were submitted to statistical analysis respectively ... Correct me if I'm wrong.

In any case, exceptional claims seems to apply here if we're going to give more weight to the finding. We have included the opinion that the preponderance of evidence should be considered rather than single studies (a common, reasonable argument, now with something to cite it to), how much further do you suggest we go based on Panchin? --Tsavage (talk) 16:12, 3 March 2016 (UTC)

I actually wasn't making any suggestions in that comment - I was just pointing out that your claim that substantial equivalence was "not mentioned" in the source is incorrect. "Despite their claims" is pretty clear to me - it refers to the claims made in the six studies they reviewed, and draws a contrast between the claims and the evidence presented for them. I don't think we're in a position to decide that this particular statement about the studies they reviewed is "editorializing," while other related statements are actual conclusions.

For the Xu letter, I'd have to read it carefully, but it's not about Panchin having done extra analysis beyond the correction. It looks like they're arguing their study should be "counted" as two separate studies, each with only half as many comparisons...Regardless, I'd point out that they agree with Panchin on GMO safety ("are safe for human beings," etc), so I don't think it would be appropriate to use it as a rebuttal or counterpoint for the specific statement about safety that we're citing. Sunrise (talk) 00:11, 4 March 2016 (UTC)

(edit conflict) I read the letter-to-the-editor in response to Panchin, and I don't think that it indicates any reason to change what is currently on the page. The authors of the letter were the authors of a study of gut microbes after animals ate GM foods. The quote above about Tables 2 and 3 doesn't dispute Panchin so much as justify why the authors think that their data in those tables is still ok. But they go on to say: "However, it's a “complex effect” as we don't know these effects are beneficial or harmful." and "In fact, we believe that GMO are safe for human beings," and also "If professor Panchin is concerned about the statistics problem of this article, we believe the Bonferroni correction strategy of this article (respective Bonferroni correction for each separate table) is a compromised and suitable one." They are not really disputing Panchin's view of the statistics, and they certainly are not saying anything that should change what we currently say about Panchin on the page. --Tryptofish (talk) 00:21, 4 March 2016 (UTC)

I wasn't suggesting any changes to the existing text either, so I'll keep my reply to that! --Tsavage (talk) 17:26, 5 March 2016 (UTC)

Domingo and Krimsky

• I have not yet reviewed Pachin, but I have reviewed both Domingo (2011) and Krimsky. The beginning of the sentence "Some studies have claimed that genetically modified crops can cause harm,[199][200]" is not an accurate NPOV description/summary of what those reviews say. I was shocked to see the footnote references added to this existing for this sentence where the material in the reviews did not support the material in the sentencewhere the material was added. A more accurate description of what Domingo talks about can be found in the Abstract:

The number of citations found in databases (PubMed and Scopus) has dramatically increased since 2006. However, new information on products such as potatoes, cucumber, peas or tomatoes, among others was not available. Corn/maize, rice, and soybeans were included in the present review. An equilibrium in the number research groups suggesting, on the basis of their studies, that a number of varieties of GM products (mainly maize and soybeans) are as safe and nutritious as the respective conventional non-GM plant, and those raising still serious concerns, was currently observed. Nevertheless, it should be noted that most of these studies have been conducted by biotechnology companies responsible of commercializing these GM plants.

Krimsky says:

I found twenty-six animal feeding studies that have shown adverse effects or animal health uncertainties.

I see a big difference between saying studies "raise still serious concerns" vs. saying GMOs "can cause harm". This is typical of the misrepresentation of Seralini's study which never claimed that GMO corn causes cancer--Seralini instead said the finding were of concern and pointed to the need for long term cancer studies. These studies typically say there is insufficient study and too much of the studies are from industry. In Krimsky:

David Schubert, professor at the Salk Institute, summarized the state of affairs of the GMO controversy as follows: ‘‘To me, the only reasonable solution is to require that all GM plant products be tested for long-term toxicity and carcinogenicity before being brought to market’’ (2002, 969). Until the twenty-six studies, or at least the best of them, are replicated and shown to be false positives, we have an obligation to treat these studies with respect and concern.

* * *

It has been well established in social science research that in some fields there is a funding effect in science from corporate sponsorship of research. That means that corporate-funded science tends to produce results that are consistent with corporate financial interests.

I frequently see statements like these even from those who support the "scientific consensus" (e.g. the AMA [7]), and yet that is not what our articles say about the science. We need to report what is in the WP:RS rather than this black/white causes harm or does not cause harm. It seems some editors prefer to write what they think the RS says and then look for RS to back it up. It should be the other way around. Because of that, I often find sentences that look like PR that is twisting material to suggest conclusions that are not at all in the RS, just as Petrchan47 pointed out above. --David Tornheim (talk) 22:03, 9 February 2016 (UTC)

I've read your comment a couple of times, and I'm confused as to which footnote you mean when you say "I was shocked to see the footnote references added to this existing sentence..." Which footnote is that? --Tryptofish (talk) 22:21, 9 February 2016 (UTC)

You are correct. I thought the sentence "Some studies have claimed that genetically modified crops can cause harm" was there and the footnotes were added to it. I am almost sure I have seen it in the article before, but the diff indicates the sentence and the footnote appeared at the same time. Hence I have corrected my statement above. Sorry for the confusion. --David Tornheim (talk) 09:45, 10 February 2016 (UTC)

Thanks! No problem. --Tryptofish (talk) 19:02, 10 February 2016 (UTC)

I agree that Domingo does not really say that there are studies showing harm, so much as describing the ways in which Domingo felt that the literature was not yet complete. And I have wondered why some editors have seemed so eager to cite Domingo as a source "against" the scientific consensus.--Tryptofish (talk) 22:21, 9 February 2016 (UTC)

How can something be declared safe if it has not yet been adequately/fully studied/tested? --David Tornheim (talk) 09:45, 10 February 2016 (UTC)

I do not think that Domingo actually declares GMOs safe, which is why I favor citing that source along with Krimsky as representing a view that to some extent disagrees with the sources that say that there is a "scientific consensus". --Tryptofish (talk) 19:02, 10 February 2016 (UTC)

What does "adequately/fully studied/tested" mean? They do safety testing, e.g., for allergenicity and have cancelled products based on the results. GM Salmon were "studied" for 17 years before gaining approval. Is that enough? How much is enough (according to an RS)? Aren't they tested adequately according to consensus science? Lfstevens (talk) 02:04, 11 February 2016 (UTC)

Krimsky, on the other hand, does clearly say very explicitly that sources such as Seralini should, in Krimksy's opinion, be treated as serious science. And I looked at that Huff Post piece about the AMA, and I'm not seeing the AMA saying anything there about corporate interest. What they do say is what all the mainstream sources say about "case-by-case" testing. The piece also cites advocacy groups that criticize the FDA for not requiring such testing. --Tryptofish (talk) 22:21, 9 February 2016 (UTC)

I mostly agree. I did not notice any mention of the corporate interest issue by the AMA, but they clearly called for pre-market testing in the Huffington post [8] and Chicago Tribune [9] articles. That is what I was referring to above. In other words, the AMA is echoing both Krismky and Domingo saying there is insufficient study/testing. And I do bring this up again here. --David Tornheim (talk) 09:45, 10 February 2016 (UTC)

(Just noting that you divided my post above into two parts for purposes of replying.) I think there is an important distinction to make. Krimsky says that he thinks the scientific evidence actually demonstrates that GMOs are not safe. Domingo does not conclude that, but says that there wasn't enough testing to establish a scientific consensus that GMOs are safe. The AMA, in contrast, agrees with a bunch of other sources that testing must include case-by-case testing that occurs pre-market, but they explicitly agree with other sources that there is a scientific consensus about existing GM foods. To portray the AMA as taking the position that there is not enough science to justify the safety of existing GM foods seriously misrepresents what that source says. They insist instead that case-by-case testing must continue, which is entirely in accordance with the WHO and many other mainstream sources. --Tryptofish (talk) 19:02, 10 February 2016 (UTC)

I did not say "the AMA as taking the position that there is not enough science to justify the safety of existing GM foods", but you can look at what the Chicago Tribute reported:

The American Medical Association called for mandatory pre-market safety testing of genetically engineered foods as part of a revised policy voted on at the AMA's meeting in Chicago Tuesday.

Currently biotech companies are simply encouraged to engage in a voluntary safety consultation with the Food and Drug Administration before releasing a product onto the market.

It seems hard to conclude that the AMA feels the current requirements of pre-market testing are sufficient. So, I'm not really clear on what exactly you think I am misrepresenting. Are you interpreting it to say that they think the current system is fine as it is and has plenty of required testing to ensure food safety? --David Tornheim (talk) 06:36, 11 February 2016 (UTC)

If we have 1700+ safety studies, isn't it reasonable/likely that 5% (85+) of those that show statistical significance at the 95% confidence level are simply random (both pro and con)? And given the above reanalysis, don't we have to reject the conclusions of some of the con studies (and possibly pro studies, too) based on their demonstrably poor methodology? (I'm asking for more sources who address these matters.) Lfstevens (talk) 02:04, 11 February 2016 (UTC)

That's a bit of a misunderstanding (though common) on how stats work, but I won't get into that here. The short of it is that studies not showing an effect don't fall into that 95% error rate. It's only those that claim statistical significance that matter for that. Panchin describes that a little bit in why they only needed to analyze significant studies if you're wondering. Kingofaces43 (talk) 03:47, 11 February 2016 (UTC)

Thanks for the explanation! Lfstevens (talk) 04:59, 11 February 2016 (UTC)

At this point, it is getting difficult to see what the objection was to how Domingo and the AMA are described on the page, but I guess it comes down to the fact that the AMA, on the one hand, agrees with the scientific consensus that existing GM foods are no more dangerous than conventional foods – but on the other hand, the AMA is in favor of legally required testing instead of voluntary testing. But there is a difference between scientific conclusions, and opinions about law. To my knowledge, nothing on the page says that the scientific consensus is that legislation about testing is sufficient. But there is no problem with the fact that the scientific consensus is that the testing that has actually occurred is sufficient to establish the safety of the crops that exist up to now. There is nothing illogical about saying, in effect, we scientists think the testing that has occurred has been fine, but we caution against abandoning that testing in the future. --Tryptofish (talk) 20:24, 11 February 2016 (UTC)

I made the objection at the top and it is just as correct as it was when I first articulated it (with the slight revision). The AMA is just one more example along these lines. The sentence I pointed out does not articulately accurately and summarize the concerns raised in the articles or by the AMA. This sentence is in the body under "controversy" not as part of any "scientific agreement" or "scientific consensus". I will try a WP:BRD that is more true to what these articles are saying. My problem is this black/white thinking of either "shows harm" or shows "no harm": Krimsky and Domingo say it is gray, which is what Seralini was saying as well, and they suggest a need for more study per the Precautionary Principle. The AMA is saying IMHO that there is no KNOWN harm from GMO, that being a GMO does not inherently make it more dangerous than ordinary food, so therefore why label it? But that there is potential for harm for reasons and mechanisms unknown, which is why it needs to be tested, which is not required, unlike in Europe where it is tested the way artificial additives are in the U.S. And Seralini is saying that even the required testing in Europe is insufficient because the studies are all 90-days feeding trials by the companies who keep the data confidential. and he thinks they should be at least 2 years. So a sentence like the above just does not show what the authors are saying. --David Tornheim (talk) 01:18, 12 February 2016 (UTC)

Maybe I'm just missing something, which is certainly possible, so I'll wait and see what you present. --Tryptofish (talk) 01:41, 13 February 2016 (UTC)

Sentence on Regulations by Country

Above Tryptofish and Kingofaces43 say that the last sentence of Tryptofish's revised 1st proposal could be added now. That sentence specifically is:

The legal and regulatory status of GM foods varies by country, with some nations banning or restricting them, and others permitting them with widely differing degrees of regulation.

Citations

I agree. Is there any objection to adding this sentence as proposed above to GM articles where the "scientific agreement" on GM safety sentence is now found? --David Tornheim (talk) 09:17, 8 March 2016 (UTC)

I'm happy that we appear to have consensus about this! Obviously, I'm in favor of the sentence, but I have a purely procedural objection to adding it now. Because of the planned RfC to evaluate the (currently three) proposed versions of the language for which there is not yet consensus, it will be necessary to compare and contrast Proposals 1 and 2. Proposal 2 says something about regulation that contradicts what the sentence here says, and I feel that it is necessary for RfC participants to be able to see the regulation language in both proposals, side-by-side. There is no hurry about this, so let's please wait until after the RfC. --Tryptofish (talk) 19:54, 8 March 2016 (UTC)

Okay. I will wait. By proposal 2, do you mean "Proposal 2 revised" or "Proposal 2"? --David Tornheim (talk) 23:46, 8 March 2016 (UTC)

Thanks. I meant revised, because it's the most recent version of it (there's another sort-of-version in the original Prop 2 section, but it was never completed). By the way, I'm simply numbering these things in the order in which they were submitted, no implied value judgment. --Tryptofish (talk) 23:54, 8 March 2016 (UTC)

Okay. Proposal 2 revised does not contradict the above. (Proposal 2 (unsourced) did, and it was wrong for doing so.) If you think it does contradict the above sentence, please explain how. @Tsavage: Do you think the above sentence in anyway contradicts Proposal 2 revised? --David Tornheim (talk) 00:11, 9 March 2016 (UTC)

My reason for saying that is at the bottom of #Second proposal revised. Although it does say "where available", it makes it sound like there is a "consensus" among national regulatory agencies that GMOs are safe, and I don't think that that kind of consensus exists. --Tryptofish (talk) 00:18, 9 March 2016 (UTC)

Right, I just noticed your concern and was about to strike the above. I will respond up there. I have somewhat similar concerns about the language. I'll explain up there. I don't think the language from the WHO ("GM foods currently available on the international market have passed safety assessments and are not likely to present risks for human health.") which is where the language for Proposal 2 revised originated contradicts your sentence above, except by omission. --David Tornheim (talk) 00:24, 9 March 2016 (UTC)

Also, for the record the material in the above sentence should have been in our articles for some time (years). The absence of that very basic material (also observed by Tsavage and others) is one of the major reasons why I have alleged our GM article are not WP:NPOV, since I first looked at them. They have improved somewhat since ArbCom, but this kind of missing information is still a problem. --David Tornheim (talk) 00:12, 9 March 2016 (UTC)

I agree with you completely that governmental positions should indeed be adequately covered on these pages. At the same time, I want to be precise about the differences between the positions taken by governments and the positions taken by scientists, which are two different things entirely. (Just look at how the current majority party in the US Congress regards evolution and climate change!) --Tryptofish (talk) 00:21, 9 March 2016 (UTC)

I agree that science and government positions should be distinguished for exactly the reasons you give above. I should point out that under WP:MEDRS the WHO is listed as a "scientific organization", but I consider it to be more like a government organization similar to the FDA but on a global level (our article describes it as a "specialized agency of the United Nations"). --David Tornheim (talk) 00:50, 9 March 2016 (UTC)

The reason why the diversity of regulatory stances is not covered is because it's very difficult to do so from a weight perspective. A number of government organizations adopt WP:FRINGE views on this topic that run counter to the scientific community consensus. Anti-gmo activists consistently try to claim there isn't consensus because X number of countries "ban" GMOs or try to allude to such things by constantly trying to insert that into conversation. Because of all that, having minimal content here on regulation is part feature, not flaw, and also partially just something nuanced that can't easily be worked in yet given the atmosphere here in this topic. Kingofaces43 (talk) 16:17, 9 March 2016 (UTC)

• Re ping from above. Perhaps "respective" is not clear enough in context. More explicit is something like: Genetically modified foods, in countries where they are approved, have been assessed for safety by the respective national regulatory agencies, and determined to be as safe to eat as conventional food. That also changes "available" to "approved"—apparently around 60 countries regulate cultivation and importation of GM food, so strictly speaking, GM foods may be available as imports in countries with no regulation. I'll revise and re-post. --Tsavage (talk) 04:54, 9 March 2016 (UTC)

I will express my concern about that language in a second above. --David Tornheim (talk) 05:28, 9 March 2016 (UTC)

Thinking it over, I agree with David that, technically, what we have here is an omission moreso than what I previously called a contradiction. @Tsavage: please make sure that you have clearly indicated whatever ends up being your final revision, before we start an RfC, thanks. --Tryptofish (talk) 18:55, 9 March 2016 (UTC)

Survey

The American Society of Plant Biologists did a survey of their team on GMO safety. Lot of people signeed up for "safe". How does this affect our ongoing blood sport? Lfstevens (talk) 06:41, 18 March 2016 (UTC)

Blood sport indeed, unfortunately. Obviously, it doesn't really say anything about how the general public views it, but I guess it adds slightly to the documentation of scientific consensus. I suppose there is no way, however, to know how many of the signers work for GM businesses, nor how many scientists declined to sign. --Tryptofish (talk) 22:56, 18 March 2016 (UTC)

Looking at it further, I see the petition was designed as a rebuttal to something published by ENSSER: [10]. I was amused to see that ENSSER, in turn, cites Wikipedia. It also presents Seralini's papers as scientific fact. But I gave the link here, in case anyone else wants to propose citing it. --Tryptofish (talk) 23:05, 18 March 2016 (UTC)

Notice of request for RfC

Please see: WP:ARCA#Amendment request: Genetically modified organisms. Thanks. --Tryptofish (talk) 23:37, 25 March 2016 (UTC)

Citation format for Government Orgs

Is it okay if I do this here? @Tryptofish: I would like to also change the refs in our various proposals to standardize them--not just what is in the article already. By doing the kluge proposed by Boghog, they will reflect MLA format. --David Tornheim (talk) 10:06, 29 March 2016 (UTC)

My personal attitude about such things is that I really don't care, although I know from experience that other editors sometimes care very much. Looking at the discussion you had with Boghog, I see that there has been a subsequent comment by a third editor, made after the posts you made here and at the "Foods" page about this, in which that third editor objects to switching the "publisher" and "author" fields – so I don't feel comfortable that there is consensus to do what you propose, even though it's not a big deal to me personally. Please let me suggest this: please feel free to make the change in Proposal 4 anytime you feel like it, but wait for further editor input before changing any of the other proposals or changing the page. Does that seem reasonable? --Tryptofish (talk) 19:17, 29 March 2016 (UTC)

As a convenience, I'm putting this discussion link here: Help talk:Citation Style 1/Archive 86#Question: Which Template to use.. --Tryptofish (talk) 19:27, 1 April 2016 (UTC)

Suggested change to Controversy Section

Firstly I would like to congratulate those responsible for the article. I haven't used a talk page before, so please be patient if I get the protocol wrong. I am trying my best.

I would like to suggest that the Controversy section makes mention of the fact that many beleive that genetically modified crops have positive benefits. Scientists widely hold this point of view and it is expressed in summary consensus documents from very prestigious and authoratitive sources, and therefore could be called in Wikipedia a "consensus" or "of general scientific agreement". For instance:

The USDA in the web page "Biotechnology Frequently Asked Questions" A whole section is dedicated to "What are the benefits of Agricultural Biotechnology?" The section begins "The application of biotechnology in agriculture has resulted in benefits to farmers, producers, and consumers. Biotechnology has helped to make both insect pest control and weed management safer and easier while safeguarding crops against disease." and then goes on the cite benefits in more detail

More recently in a USDA report summary from the Economic Research Service "Genetically Engineered Crops in the United States" Second page "Farmers generally use less insecticide when they plant Bt corn and Bt cotton. Corn insecticide use by both GE seed adopters and nonadopters has decreased—only 9 percent of all U.S. corn farmers used insecticides in 2010. Insecticide use on corn farms declined from 0.21 pound per planted acre in 1995 to 0.02 pound in 2010. This is consistent with the steady decline in European corn borer populations over the last decade that has been shown to be a direct result of Bt adoption." The report also mentioned ahat the adoption of HT crops has enabled farmers to use glyphosate instead of more toxic and persistent herbicides.

In the (British) Royal Society's report "Reaping the benefits" Page 23 Regarding BT corn "Because damage caused by insect feeding allows entry of mycotoxin-producing fungi, a secondary benefit is that Bt maize also has lower levels of fungal mycotoxins in the grain than non-Bt maize, thus enhancing its safety as food or feed" "Control of insect pests with insecticides poses a greater risk of damage to non-target organisms than control with transgenic Bt protein." Page 30-31 Regarding herbicide resistance crops "Control of weeds in conventional cropping systems is achieved by tillage combined with herbicide application. However, the use of herbicide-resistant plants provides good weed control with little or no tillage and so a secondary benefit from the use of these crops has been the spread of reduced tillage systems in which soil erosion is reduced"

I beleive that mentioning benefit is necessary for a neutral poitn of view, as is, the Controversy section only mentions allegations of harm, and allegations of no harm. Without benefit, no reasonable person would consider using them.

With thanks RAMRashan (talk) 10:21, 4 April 2016 (UTC)

RAMRashan, thank you for discussing your ideas on the talk page. I agree that the article should reflect what reliable sources say in terms of benefits of GM crops and food. In my view, the article already does reflect the positive view in sections including 'Economics' and 'Yield'. As such, these views are already given due weight and their inclusion in the controversies section would be redundant.Dialectric (talk) 14:53, 4 April 2016 (UTC)

Dialectric, thank you for your reply. Sorry I didn't put what I wrote in context. I was giving my reasons for a small edit made on 09:45, 3 April 2016 (by myself) that was quickly reverted. This edit makes very brief mention of benefits, nothing big that would be redundant, but something small that would move make better balance. As it is now, the Controversy section mentions concerns on "long-term impact on human health", "contamination of the non-genetically modified food supply", ""effects of GMOs on the environment and nature", "the rigor of the regulatory process", "consolidation of control of the food supply in companies that make and sell GMOs", and "concerns over the use of herbicides with glyphosate", but doesn't even mention the word benefit. This seems an unbalance within this section to me. The suggested edit inserts eight words, and would change the text to the following:

"There is a general scientific agreement that food from genetically modified crops is not inherently riskier to human health than conventional food, and can provide benefits to consumers and farmers, but that they should be tested on a case-by-case basis" How does that sound? RAMRashan (talk)

The discussion at the top of this page, 'General agreement sentence, continued', has several editors discussing the consensus statement you mention. The wording of this consensus statement has been a subject of debate for months on wikipedia, and any change will likely have to go through some discussion. You could join that conversation and add to the upcoming rfc if you are interested in changing the statement.Dialectric (talk) 18:59, 4 April 2016 (UTC)

Dialectric, thanks, I have done what you suggest, kind regards RAMRashan (talk)

• It is at #Fifth proposal: Controversy Section Doesn't mention GMO benefits cited by prestigous scientific organizations, above. RAMRashan, welcome to the discussion, and thanks. --Tryptofish (talk) 21:17, 4 April 2016 (UTC)

Tryptofish, most sincere thanks for welcoming me to the discussion! RAMRashan (talk)

A reliable source directly addressing existing scientific consensus in GM food safety

This statement squarely addresses the issue of scientific consensus in GM food safety. It's from an article that discusses the general GM food acceptance issue in the US, and includes interviews with scientists at the University of California. It's quite current, published in Oct 2014, in UofC Berkeley's alumni magazine:

Unlike global warming, which an overwhelming consensus of scientists accept as real and inevitable, there is less agreement about the value of GMOS. Many scientists on both sides of the controversy say there is no conclusive evidence that eating genetically modified foods is directly harmful to human health. But that’s where the consensus stops.[11]

Of the sources presented so far, this is the only one I've seen that directly and specifically addresses the question of scientific consensus that we are discussing, and does so using opinions solicited from scientists working in the field. Overall, the article provides an overview of GM food science opinion from a scholarly perspective, and as such appears to be a useful source for expanding this area of coverage. Comments? --Tsavage (talk) 01:31, 10 April 2016 (UTC)

If anyone wants to cite it in any of the proposals about to be presented in the RfC, please go right ahead. But what you quote is a summary written by a non-scientist [12], in an alumni magazine. That is not in any way a scholarly or expert publication, nor even would it have been vetted to reflect the consensus of the faculty at that university. It quotes a couple of faculty members, most of whom are not disagreeing with the safety of GM foods to eat, although several of them do indeed express environmental and economic concerns. Two faculty members are quoted as opposing GMOs. Primarily, this is one non-tenured faculty member who self-describes as having expertise in "sociology and political ecology" [13] and is described in the source as "anti-GMO". A second, tenured, faculty member [14] is quoted as being very concerned about GMO farming practices being non-sustainable in terms of excessive monocropping, loss of genetic diversity, and insecticide use, but she is not quoted at all about the scientific consensus on the safety of eating these foods. --Tryptofish (talk) 18:59, 10 April 2016 (UTC)

Something else occurred to me. In the quote above, about "where the consensus stops", I actually do not think that editors here disagree that there is indeed ongoing controversy beyond that stop point. Where editors are disagreeing is about what our pages should say about the effects of GMO foods on human health, and even the source here isn't really saying that there is a lack of scientific consensus about that specific point. The "value of GMOS" is about more than food safety, and of course there is well-sourced disagreement about that overall value. There are plenty of scientists who say that the food is safe to eat but that it hurts society when companies sue farmers for saving seed. It would be misleading if we were to present this source as really upending the sourcing for a scientific consensus specifically about food safety. --Tryptofish (talk) 23:43, 11 April 2016 (UTC)

Indeed, the main issue continues to be the use of the term scientific consensus regarding GM food consumption findings. The UofC Berkeley California mag article, excerpted above, and Nicolia (main source for consensus proposal) both consider a range of GM crop safety concerns, not just consumption (the 1783 papers reviewed by Nicolia are broken out as GE general (diff between US and EU regs) 9.3% - 166/1783; GE environmental 47.5% – 847/1783; and GE food & feed 43.2% – 770/1783). The cited Nicolia summary statement refers to all of these concerns collectively (where the mag excerpt applies the same to the subtopic of consumption):

We have reviewed the scientific literature on GE crop safety for the last 10 years that catches the scientific consensus matured since GE plants became widely cultivated worldwide, and we can conclude that the scientific research conducted so far has not detected any significant hazard directly connected with the use of GM crops. Nicolia, 2013

There appears to be general agreement that no direct hazard has been definitively identified so far, and the California article supports that in a well-written, easily verifiable way (especially keeping WP readers in mind). Going beyond that point is where sourcing gets more tenuous and the disputes arise. Ultimately it comes down to whether we can paraphrase "no significant hazard found" as "as safe to eat as conventional food," without a reliable, independent source that clearly states the latter. --Tsavage (talk) 23:27, 12 April 2016 (UTC)

There are multiple reliable, independent sources that do indeed use wording like no greater risk to human health than conventional food. As just one example, Yang and Chen, cited in proposal 1, say: "Overall, a broad scientific consensus holds that currently marketed GM food poses no greater risk than conventional food... many independent international science organizations agree that GMOs are just as safe as other foods", and there are more sources like that. --Tryptofish (talk) 23:49, 12 April 2016 (UTC)

At least it's hopefully becoming clearer that what is being argued is summary wording, and not some editors trying to imply that GM food is unsafe against the evidence.

It is easy to make the jump from "no evidence of harm so far" to "as safe as," however, we need a solid source to do that in Wikipedia's voice. Nicolia doesn't make that leap (and in fact establishes substantial equivalence as the safety science of commercialization, which is strictly case by case; the WHO citation supports this). If Yang and Cheng do find consensus, then that should be proposed as the main source for a consensus statement.

Critical to a general reader's understanding of GM food safety is an explanation of the opposing US and EU stances, and of the common ground of substantial equivalence, which IS the current consensus-accepted science of GM food safety. These are ignored in the type of consensus statement being proposed. --Tsavage (talk) 13:49, 13 April 2016 (UTC)

You will find some of the latter in Proposal 4. --Tryptofish (talk) 16:09, 13 April 2016 (UTC)

BTW, I looked at Yang and Chen. The paper is paywalled; the abstract concludes: Mandatory GMO labeling is not supported by science. Compulsory GMO labels may not only hinder the development of agricultural biotechnology, but may also exacerbate the misconception that GMOs endanger people's health. (c) 2015 Society of Chemical Industry This gives the impression that the paper is in favor of the proliferation of GM food. Of course, this consideration does not make it any less evidence-based than any other publication, however, it does suggest bias, which speaks to independence of the source, and its choice of wording in discretionary areas. Considering that scientific consensus is a subjective assessment, not a scientific finding, we would probably be better served by a more independent source.

Once again, this is the particular problem with scientific consensus, it is a politicized term that we should avoid in tricky topics, in favor of more neutral language with which to state the facts. I believe that's the intention of some or all of the alternative proposals. --Tsavage (talk) 21:33, 13 April 2016 (UTC)

That is circular logic that cannot be applied in terms of reliability or weight. To allege bias and lack of independence because a source reaches a particular set of conclusions (especially in line with mainstream science) would be equivalent to saying the same of a paper saying certain policies should or shouldn't be enacted with respect to climate change because they earlier stated they agreed with the scientific consensus that climate change exists. I don't think anyone can reasonably entertain that line of thought with this source. Kingofaces43 (talk) 23:06, 13 April 2016 (UTC)

Tsavage, the quotes from the main text were provided by KofA, who has access to the full paper. The authors of the paper are both university professors, unaffiliated with industry, so they are not saying what they say out of what Wikipedia would call a COI. It's their opinion, based on their scholarship, and they cite the Domingo paper, so they even took that into consideration. Although "scientific consensus" is indeed a term that gets politicized, it's not a political term when scientists use it in the context of peer-reviewed scholarship. And as I said earlier, it's just one example of such sources, not the only one. Anyway, this is a debate that will have to be settled via the RfC. --Tryptofish (talk) 01:42, 14 April 2016 (UTC)

Notice of Discussion of Rules for RfC on GMO food safety

A discussion is taking place here about a proposed RfC on GMO food safety based on the above five proposals. The Wordsmith and Laser_brain have graciously volunteered to oversee the RfC. In addition to discussing the rules, The Wordsmith has created a proposed RfC here. This is not notice that the RfC has begun. I am hereby also pinging those who have either created and/or commented on these proposals or on food safety above: @Tryptofish:, @Kingofaces43:, @Sunrise:, @The Four Deuces:, @Semitransgenic:, @Dialectric:, @Aircorn:, @Tsavage:, @Petrarchan47:, @RAMRashan:, @Lfstevens:. I will also put notice on your talk pages if you are not already at the discussion. --David Tornheim (talk) 07:47, 2 May 2016 (UTC)

Wikipedia:Requests for comment/Genetically modified organisms

This is a notice that Wikipedia:Requests for comment/Genetically modified organisms is open for public comment. AIRcorn (talk) 04:32, 17 June 2016 (UTC)