When I was younger, I used to think the Seralini study had something worth saying. The fact that he didn’t actually release the raw data at the beginning seemed unimportant, and I saw reason to at least consider what he found.

After reading more into the subject, reading about the actual regulations and standards regarding the type of study he did, and learning more about how he went about releasing his paper, I stopped defending him. After he republished in a pay-to-publish journal, without replicating the experiment using a more appropriate species or other methodological improvements, I began criticizing his paper and statements openly in any discussion where he was brought up. Here’s why:

Reason Number 1: He made journalists who saw the study early agree to not consult or speak to any other scientists regarding the study. He literally denied them the right of the journalists to verify his interpretation before writing about it. That isn’t ethical or common, and it feels sneaky to me.

Reason Number 2: Seralini did not use enough rats for the type of study he was conducting, based both on its aims, length, and choise in rats. 

The standard number per group for a 90 day trial with SD (Sprague dawley) rats is 10 per sex per dose. The Hammond papers had double that: 20. The standard number per group for a 2 year trial with SD rats is 50-65 depending on the governing body (OECD says 65 when using the rats Seralini chose). 65 rats, per sex, per treatment group is truly the standard for SD rats in 2 year studies. So since the OECD is the relevant body, the number should have been 65 per treatment for a study that length with those rats.

Instead of claimed “2.5 times as many rats as the Monsanto group as he should have had”, Seralini had half as many (10) as Hammond, and less than 1/6 of what was called for.

This is not an objective, reliable, or even valid experiment, it was in many ways a statistical fishing expedition similar to the “chocolate weight loss” paper. Calling Seralini out is not the result of a double standard, it’s the only reasonable response given the whole picture. If you have a group of rats with an expectation of 50% mortality and 80% cancer at the 2 year mark and you have 10 rats in a group, you really only have a group of 5, since the other 5 are expected to be dead before the trial is over.

Reason Number 3: It would be a surprise to anyone who read the paper that it was not a cancer study, because discussion of the tumors takes up all the oxygen and then he took the unprecedented step of including pictures of the tumors from only the treatment group and leaving the control out. 

To say that it was not a cancer study is not the point. The issue is the paper and what it conveys, and what Seralini did with the paper afterward. If you read the original paper, it’s all about the tumors. The only images of rats with tumours in the study came from the GM test groups, even though tumours, death, and cancer were not significantly different between groups… why only show pictures of the tumours on GM rats when both groups had tumours?

Reason Number 4: To say that the tumors were an unexpected surprise is horseshit. Everyone knows that SD rats are tumor prone. That’s one of the reasons they are used. Check out Suzuki, et al 1979. In a 2 year trial using SD rats, 81 of 100 rats developed tumor on pre-GMO rat food.

Reason Number 5: There were no dose responses. In fact, straight glyphosate seemed protective in male rats, even though the paper only emphasized the higher mortality in female rats. This already shows a bias in terms of what they decided to highlight, one could just as easily have projected the unsound argument that the more male rats drank, the stronger and healthier they became. Take a look at a more objective analysis of Seralini’s data.

Here are the Seralini survival curves:

Reason Number 6: He republished his study again recently in a pay-to-publish journal, without peer-review, and without replicating the experiment. Any scientist worth anything knows that to fight doubts about the reliability of the data, replication is in order. So, why would he then decide to simply republish a paper that had been retracted, instead of redoing the study?

Reason Number 7: It is simply wrong to claim that he is the only one who has looked into the carcinogenicity of glyphosate, which I should remind you is not even synonymous with GMOs (there are even new GM crops that can ward off pests without needing any poison or extra proteins). I’m not a big fan of glyphosate, but I will admit that an analysis more than a dozen (14) rodent studies into its toxicology and carcinogenicity didn’t turn up significant effects.

An analysis of 12 multi-generational studies and 12 long-term toxicology studies of GM animal feeding studies also found no significant effects.

If Seralini really believes in his results, I again ask why he didn’t replicate his experiment? How can anyone stand behind a record of 1:14 with certainty?