The Journal of Agricultural and Food Chemistry sits at a busy crossroads: food chemistry, plant metabolites, nutritional biochemistry, analytical chemistry, veterinary residues, natural products, and agricultural science. That makes it a fascinating journal to inspect in the Retraction Watch database.
In the uploaded dataset, I found:
8 retraction records from the Journal of Agricultural and Food Chemistry.
That is not a huge number, so the analysis must walk carefully, not stomp around in statistical boots. But the pattern is distinctive: JAFC retractions are not dominated by paper mills, fake peer review, plagiarism, or mass publisher cleanups. They are mostly data, reliability, image, and misconduct-related cases, a more classical “scientific evidence failed” pattern.
I compared JAFC with a broader set of 462 agri-food-related non-conference records from 129 journals, defined by journal titles containing terms such as food, agriculture, agronomy, crop, horticulture, cereal, grain, soil, dairy, or aquaculture. I excluded conference/proceedings records to avoid the large batch-retraction distortions we saw earlier.
1. How many JAFC retractions are there?
In the uploaded Retraction Watch file:
| Journal | Retraction records | Publisher |
|---|---|---|
| Journal of Agricultural and Food Chemistry | 8 | American Chemical Society |
Among ACS non-conference records in the database, JAFC contributes about 1.9% of ACS retraction records. So JAFC is not one of the major ACS retraction hotspots. It sits far below ACS Omega, JACS, ACS Applied Materials & Interfaces, and Organic Letters in raw count.
Within the agri-food-related journal set, JAFC is a mid-sized cluster: tied with Food & Function, slightly below Food Chemistry, and far below the recent agri-food batch clusters such as Acta Agriculturae Scandinavica, Section B and Journal of Food Quality.
Non-conference records only. Journal titles were selected using food, agriculture, agronomy, crop, horticulture, soil, dairy and aquaculture terms.
Calculated from the uploaded Retraction Watch CSV.
The important first conclusion:
JAFC is not a high-volume retraction journal in this dataset. Its importance comes from the kind of retractions, not the number.
2. What were the JAFC retractions about?
The eight JAFC records were:
| Retraction lag | Country | Topic, shortened | Main reason pattern |
|---|---|---|---|
| 0.18 years | China | Oxidative stability of unsaturated monoacyl trehalose | Limited/no information notice |
| 0.37 years | China | Sea cucumber fucosylated chondroitin sulfate and atherosclerosis | Error in analyses/data, unreliable conclusions |
| 0.67 years | China | Cyadox pharmacokinetics and tissue depletion in pigs/broilers | Unreliable conclusions |
| 2.06 years | Taiwan | Asiatic acid, glycyrrhizic acid, oleanolic acid | Data concern, image error, unreliable conclusions |
| 2.66 years | United States | Garlic and cardioprotection | Fabrication/falsification, institutional investigation, misconduct |
| 3.46 years | Italy, United States | Red and white wine, “French paradox,” cardioprotection | Fabrication/falsification, consent issues, misconduct |
| 4.21 years | United States | Broccoli and cardioprotection | Fabrication/falsification, institutional investigation |
| 6.55 years | United States | Arabidopsis root exudate metabolic profiling | Data concerns, not reproducible, unreliable conclusions |
The biological flavor is interesting. Several JAFC retractions sit in the “functional foods / phytochemicals / cardioprotection / natural-product bioactivity” region. That is a high-risk interpretive zone because it often combines chemistry, cell biology, animal physiology, nutrition, and disease claims. A molecule starts as a food-derived compound, then quickly begins wearing a biomedical superhero cape.
That cape needs checking.
3. Main reasons: JAFC looks data-heavy, not paper-mill-heavy
JAFC’s reason profile is very different from the broader agri-food-related set.
| Reason theme | JAFC | Other agri-food-related records |
|---|---|---|
| Data/results/methods/unreliable conclusions | 87.5% | 35.2% |
| Image concerns | 37.5% | 4.0% |
| Fraud/misconduct/fabrication/falsification | 37.5% | 5.3% |
| Plagiarism/duplication/copyright | 0% | 33.5% |
| Paper mill/peer-review/AI-related | 0% | 39.2% |
JAFC retractions are data, image and misconduct-heavy, while other agri-food-related journals show more plagiarism and peer-review/paper-mill signals.
Calculated from the uploaded Retraction Watch CSV. Categories overlap and percentages do not sum to 100.
Because JAFC has only eight records, the statistical tests are exploratory, not courtroom verdicts. Still, Fisher’s exact tests show that JAFC is significantly enriched for:
| Comparison: JAFC vs other agri-food-related records | Fisher’s exact p-value |
|---|---|
| Data/results/method concerns | 0.004 |
| Image concerns | 0.004 |
| Fraud/misconduct | 0.008 |
| Paper-mill/peer-review tags are lower in JAFC | 0.026 |
| Plagiarism/duplication lower in JAFC | 0.057, borderline |
So the evidence says:
JAFC retractions are more like failed experimental claims than publication-system cleanup events.
That is the opposite of what we see in some recent agri-food-related journals, where retractions are dominated by paper-mill, peer-review, or plagiarism signatures.
4. Time to retraction: JAFC is neither especially fast nor especially slow
JAFC’s median time from publication to retraction is:
2.36 years
For the broader agri-food-related set:
1.81 years
The difference is not statistically significant by Mann-Whitney test, p = 0.46. In plain language: with only eight JAFC cases, we should not claim JAFC is meaningfully slower or faster than related journals.
But JAFC has a notable temporal pattern:
In this dataset, JAFC has no retraction records after 2018, while agri-food-related retractions surged strongly after 2020.
JAFC retractions are concentrated before 2020, while several related journals show strong 2020-2026 clusters.
Calculated from the uploaded Retraction Watch CSV. The 2020-2026 era includes partial 2026.
This is one of the sharpest findings.
JAFC’s retraction pattern is old-fashioned: a small number of individual records, mostly before 2020, centered on data reliability, images, and misconduct.
The newer agri-food retraction landscape is different: large post-2020 clusters, often linked to peer-review or paper-mill-style reasons.
5. The agri-food field has undergone a retraction regime change
Across the 462 agri-food-related records, the reason profile changed substantially over time.
| Era | Records | Median lag | Dominant pattern |
|---|---|---|---|
| ≤2009 | 21 | 0.34 years | Plagiarism/duplication-heavy |
| 2010–2014 | 48 | 0.69 years | Plagiarism/duplication, some data/fraud |
| 2015–2019 | 126 | 1.01 years | Mixed, still plagiarism-heavy |
| 2020–2026 | 267 | 1.97 years | Paper-mill/peer-review and data concerns rise |
In the 2020–2026 era, 59.2% of agri-food-related records were tagged with paper-mill/peer-review/AI-related themes. That is not a JAFC pattern at all.
This suggests a genuine shift:
Older agri-food retractions looked like individual paper problems.
Recent agri-food retractions increasingly look like publisher-audit and integrity-screening events.
The villains have changed costumes. Earlier: duplicated text, unreliable claims, individual misconduct. Later: paper-mill fingerprints, compromised peer review, batch correction.
6. How does JAFC compare with specific related journals?
Here is where the comparison becomes spicy.
| Journal | Records | Median lag | Main reason signature |
|---|---|---|---|
| Acta Agriculturae Scandinavica, Section B | 82 | 2.43 years | 97.6% paper-mill/peer-review |
| Journal of Food Quality | 42 | 1.66 years | 100% data/results, 100% paper-mill/peer-review |
| Food Science & Nutrition | 21 | 2.81 years | Data/results plus peer-review/paper-mill |
| Food and Chemical Toxicology | 14 | 0.28 years | Plagiarism/duplication-heavy |
| Aquaculture Research | 10 | 2.60 years | Mixed, peer-review-heavy |
| Genetic Resources and Crop Evolution | 10 | 1.54 years | 100% data/results, 100% peer-review |
| Agronomy Journal | 10 | 1.43 years | 70% peer-review/paper-mill |
| Food Chemistry | 9 | 0.74 years | Mixed plagiarism, fraud, data |
| JAFC | 8 | 2.36 years | Data, image, fraud/misconduct |
| Food & Function | 8 | 1.33 years | Data, fraud/misconduct, some peer-review |
| Critical Reviews in Food Science and Nutrition | 7 | 2.21 years | Plagiarism/duplication-heavy |
The most important contrast is this:
JAFC resembles journals where original experimental evidence failed.
It does not resemble journals where retractions came in large paper-mill or peer-review batches.
For example:
- Journal of Food Quality: 42 records, all in 2020–2026, with 100% paper-mill/peer-review and data/reliability tagging.
- Acta Agriculturae Scandinavica, Section B: 82 records, 80 in 2020–2026, 97.6% paper-mill/peer-review tagging.
- JAFC: 8 records, none after 2018, 0% paper-mill/peer-review tagging.
These are completely different retraction ecologies.
One is a lab notebook under forensic light.
The other is a publisher pipeline being flushed.
7. Country-specific patterns
JAFC’s eight records involve only a few country tags:
| Country | JAFC country-paper occurrences |
|---|---|
| United States | 4 |
| China | 3 |
| Taiwan | 1 |
| Italy | 1 |
One paper involved both Italy and the United States, so country occurrences sum to nine rather than eight.
The country patterns differ:
- United States-linked JAFC records include the longer-lag misconduct/data cases, especially the cardioprotection cluster involving broccoli, garlic, wine, and Arabidopsis root exudates.
- China-linked JAFC records are faster, mostly under one year, and include unreliable conclusions, data/analysis errors, and limited-notice cases.
- Taiwan appears in one data/image/unreliable-conclusion case.
- Italy appears in the red/white wine cardioprotection paper, co-listed with the United States.
In the broader agri-food-related dataset, the country map looks very different.
Top country-paper occurrences in the agri-food-related journal set. Multi-country records count once for each listed country.
Calculated from the uploaded Retraction Watch CSV.
In the broader agri-food set:
- China dominates in count, with 141 country-paper occurrences.
- India is second, with 89.
- Pakistan, Saudi Arabia, Iran, Russia, and the United States follow.
- Russia has a distinctive long median lag, 4.27 years, and a very strong plagiarism/duplication signature.
- China, Pakistan, Saudi Arabia, Ethiopia, Ghana, Azerbaijan and others show strong paper-mill/peer-review patterns in this agri-food subset.
JAFC does not mirror that broader country structure. It is older, smaller, more US/China-centered, and less tied to the recent post-2020 integrity-sweep pattern.
8. Publisher-specific comparison
The agri-food-related retraction ecosystem is distributed across many publishers:
| Publisher | Records | Journals | Median lag | Main pattern |
|---|---|---|---|---|
| Taylor & Francis | 124 | 16 | 2.23 years | Paper-mill/peer-review-heavy, some plagiarism |
| Elsevier | 100 | 36 | 0.71 years | Plagiarism-heavy, fewer paper-mill tags |
| Wiley | 68 | 15 | 1.40 years | Mixed, peer-review/paper-mill and data |
| Hindawi | 43 | 2 | 1.66 years | Nearly all data plus paper-mill/peer-review |
| Springer Nature | 18 | 7 | 1.54 years | Data plus peer-review |
| Springer | 15 | 10 | 2.22 years | Plagiarism-heavy |
| ACS | 8 | 1 | 2.36 years | Data, image, fraud/misconduct |
| RSC | 8 | 1 | 1.33 years | Data, fraud/misconduct, some peer-review |
ACS’s JAFC cluster is small but more serious in the classical sense: data reliability, images, and misconduct. It is not part of the huge Taylor & Francis/Hindawi/Wiley post-2020 agri-food correction wave.
The ACS versus RSC comparison is neat because both are society publishers:
| Journal | Publisher | Records | Median lag | Main signature |
|---|---|---|---|---|
| Journal of Agricultural and Food Chemistry | ACS | 8 | 2.36 years | Data/image/fraud |
| Food & Function | RSC | 8 | 1.33 years | Data/fraud, some peer-review |
They have the same record count, but JAFC has a higher image concern share and no peer-review/paper-mill tags. Food & Function has one peer-review/paper-mill-tagged record in this dataset.
9. Subject profile: JAFC is more chemistry-biochemistry-medical than the broader agri-food set
JAFC’s subject tags are concentrated in chemistry, biochemistry, food science, plant biology, and medicine.
| Subject tag | JAFC | Broader agri-food-related set |
|---|---|---|
| Chemistry | 87.5% | 19.0% |
| Biochemistry | 62.5% | 14.9% |
| Food science | 62.5% | 23.2% |
| Medicine | 50.0% | 10.6% |
| Plant biology/botany | 37.5% | 22.7% |
| Agriculture | 12.5% | 42.9% |
| Veterinary science | 12.5% | 1.7% |
This explains why JAFC’s reason profile is more data/image/misconduct-heavy. It is not simply an “agriculture journal” in the agronomy/proceedings/paper-mill sense. Its retracted papers often sit closer to chemical biology, bioactivity, metabolism, phytochemistry, pharmacokinetics, and nutritional biochemistry.
That is a higher-interpretation-risk zone. A food-derived compound can be chemically real, biologically active in one system, overstated in another, and clinically seductive in a title. The path from molecule to health claim is paved with attractive potholes.
10. Statistical tests and what they mean
Because JAFC has only eight retractions, hypothesis testing must be treated as exploratory.
JAFC versus other agri-food-related journals
| Hypothesis | Result |
|---|---|
| JAFC has different time to retraction | Not supported, Mann-Whitney p = 0.46 |
| JAFC is more data/results-heavy | Supported, Fisher p = 0.004 |
| JAFC is more image-concern-heavy | Supported, Fisher p = 0.004 |
| JAFC is more fraud/misconduct-heavy | Supported, Fisher p = 0.008 |
| JAFC is less paper-mill/peer-review-heavy | Supported, Fisher p = 0.026 |
| JAFC is less plagiarism/duplication-heavy | Borderline, Fisher p = 0.057 |
JAFC versus other ACS journals
Compared with other ACS records, JAFC is not statistically unusual:
| Feature | JAFC | Other ACS journals | Test result |
|---|---|---|---|
| Median lag | 2.36 years | about 1.68 years overall ACS median | Mann-Whitney p = 0.89 |
| Data/results concerns | 87.5% | 70.9% | Fisher p = 0.45 |
| Image concerns | 37.5% | 36.4% | Fisher p = 1.0 |
| Fraud/misconduct | 37.5% | 21.1% | Fisher p = 0.38 |
| Paper-mill/peer-review | 0% | 0.5% | Fisher p = 1.0 |
Translation:
JAFC is unusual compared with agri-food journals overall, but not unusual compared with ACS chemistry journals.
That is a very useful conclusion. JAFC’s retraction pattern is more “ACS chemistry/data-integrity style” than “modern agri-food paper-mill style.”
11. So what is the story of JAFC?
The story is not that JAFC has many retractions. It does not.
The story is that JAFC retractions belong to a different species from many recent agri-food retractions.
JAFC’s retraction phenotype
- Small number: 8 records
- Mostly pre-2020
- No paper-mill/peer-review signature
- No plagiarism/duplication signature
- Strong data/unreliable-conclusion signature
- Substantial image and misconduct/fabrication signal
- Strong chemistry, biochemistry, food science and medicine overlap
- Country pattern centered on the United States and China, with one Taiwan and one Italy/US record
Broader agri-food retraction phenotype
- Larger recent surge, especially 2020–2026
- Strong China and India representation
- Major journal clusters in Acta Agriculturae Scandinavica, Section B, Journal of Food Quality, Food Science & Nutrition
- Much more paper-mill/peer-review signal
- More plagiarism/duplication in some older or regional clusters
- More publisher-batch behavior
So JAFC is a small old cathedral of individual evidence failures. The recent agri-food landscape is more like a warehouse audit, with entire shelves being checked and relabeled.
Final thought: JAFC’s retractions are few, but scientifically sharp
In the uploaded database, the Journal of Agricultural and Food Chemistry does not look like a retraction hotspot. It has only eight records. But those eight records are not trivial administrative dust. They cluster around the integrity of experimental claims: unreliable conclusions, data problems, image issues, fabrication/falsification in some cases, and failed reproducibility.
That makes JAFC different from many newer agri-food-related retraction clusters, where the dominant signal is paper mills, compromised peer review, or publisher-wide cleanup.
The deeper lesson is about field identity. JAFC is not merely “agriculture plus food.” It is chemical evidence applied to biological and nutritional meaning. That is powerful territory, but it is also dangerous territory. When a phytochemical, food compound, residue, metabolite, or plant extract is linked to health, toxicity, cardioprotection, inflammation, or disease, the paper becomes more than chemistry. It becomes a narrative about what food can do in the body.
Those narratives travel well. They deserve tough evidence.
JAFC’s retractions are therefore few sparks, not a wildfire. But they illuminate the combustible edge of food chemistry: the point where molecules become health claims, and where the literature must ask, very sternly, “Does the data actually hold?” 🔬🌾
No comments:
Post a Comment