We all know that 98.6° F is human body temperature … only it isn’t. A new study reconfirms something extensively covered during the COVID pandemic: Normal human body temperature falls between 97.3° and 98.2° F — with 97.9° F as today’s average.

And 5 grams per week is the amount of plastic every person consumes … only it isn’t. Like outdated body-temperature assertions, this 5-g value (widely reported in many science and news circles) is flawed. The difference is that data manipulation and memes didn’t give us the 98.6° F value … but they did help propel the 5-g-of-plastic assertion. It has shaken my faith in the scientific community.

Now, the world widely accepts the average person consumes 5 g of plastic per week — the weight of a credit card. Thanks to one now-quite-famous picture of a credit card between two chopsticks, a credit card’s worth of plastic is now widely quoted unit for measuring microplastic exposure. In a 2023 article, National Geographic authors parrot the refrain: Researchers estimate we unwittingly consume a credit card’s weight in plastic each week. Likewise, California Attorney General Rob Bonta (currently driving a lawsuit against the plastics industry) has stated, “Every week, we consume the equivalent of a credit card’s worth of plastic.” In November 2022, U.S. Senators Jeff Merkley, Elizabeth Warren, Ron Wyden, and Bernie Sanders sent a letter to Secretary of State Antony Blinken

, advising him on engagement with the Intergovernmental Negotiating Committee negotiating the U.N. Plastics Treaty: “It has been estimated that each of us is consuming a credit card’s worth of those microplastics every week through the food chain and our water supplies.” Even the website of the U.N. exclaims, “it has been estimated that humans ingest a credit card worth of plastic per week!” A June 2022 U.S. National Statement to the U.N. Ocean Conference states the same: “Human beings ingest on average one credit card worth of plastic per week.” Today, 5 grams of plastic ingested per week is a widely accepted fact.

Shown here suspended in a liter of water are 523 plastic microparticles weighing 0.65 g. This shows the mass and particle concentration necessary to ingest 5 g per week. Such a high concentration is easily seen both in water and upon drying. The particles are cut from 1.5 mm plastic monofilament.

The picture of a credit card between two chopsticks comes from press releases about a 2019 World Wildlife Federation report, No Plastic in Nature: Assessing Plastic Ingestion from Nature to People. It introduced a credit card’s worth as a unit of measure for plastic pollution. The report then quotes a WWF-funded analysis done at the University of Newcastle in Australia. “A new study finds on average people could be ingesting approximately five grams of plastic every week, which is the equivalent weight of a credit card,” is the opening sentence of the press release heralding the report’s release, right next to the chopsticks picture. Reuters and others covered the release, and it appeared in many online outlets.

The credit card picture may have led some to conclude microplastics were hiding in food rather than drink. The report clearly states 90% of plastics come from drink, and it isn’t subtle about it … and a half-page graphic on page 8 of the report illustrates how water is more significant than all other sources. So, the chopsticks picture is a stretch. It would have been more accurate to show the credit card in a two-gallon water cooler bottle. I emailed the WWF asking about the origin of the chopstick picture but got no response.

The study quoted in the 2019 WWF report appeared in 2020 peer-reviewed literature. The 2020 Newcastle paper (titled Estimation of the mass of microplastics ingested – A pivotal first step towards human health risk assessment) describes the work responsible for the chopsticks and credit-card image.

In fact, the sole purpose of the Newcastle study was to determine the weight of plastic being consumed. The estimation of the mass ingested was not tangential; to be clear, it was the sole purpose of the study. But no measurements were made for this study. Instead, the latter relied on and reported measurements made by others. Plastic consumption would be easy to measure if the plastic was in big chunks. Weighing isn’t easy because plastic is in small particles and weighing small things is difficult. Individual particles weigh almost nothing — too little to directly weigh. Things too small to weigh can be seen. Counting small things is comparatively easy. Researchers count the particles in the food we eat, air we breathe, and feces we emit. Getting from some number of particles to the particles’ mass requires a model, applying some assumptions, and doing math.

For example, knowing the number of dimes in a pile accurately gives the mass of the pile. That’s because all dimes are the same size and composition. Knowing the number of coins in a pile of mixed coins on the other hand doesn’t allow accurate calculation of a weight. Pennies aren’t the same size and composition as nickels or dimes. They’re certainly not the same as Eisenhower dollars, the largest coin circulated in U.S. in the last 50 years. The dime, the smallest, weighs one-tenth the weight of the Eisenhower dollar. 100 coins could weigh as little as 226 g (if they were all dimes) or more than 2.26 kg (if they were all Ikes).

Microplastics are even more challenging to weigh. Composition varies, just like coins. Unlike coins, microplastics come in many shapes. Tire wear particles might be approximately spherical. Fibers are long and thin. Degradation of packaging can yield thin and flat particles. Particles range from 5 mm to tens of nanometers, covering 4 to 5 orders of magnitude in size.

The relationship between diameter and weight makes microplastics weight estimation even more daunting. An Eisenhower dollar is a little more than double the diameter of a dime but weighs almost 10 times more. A widely used approximation of weight employs the cube of the diameter with all other things being equal. Using this approach, scaling an Eisenhower dollar down by 38-fold to  ~1 mm gives a coin weighing only 410 µg — 55,000 times less than the 22.68 g the dollar weights. A pile of 100 would weigh only 41 mg. Any weight estimations of a 100-coin pile become increasingly imprecise with a size range much smaller than the observed range in microplastic particles. Unfortunately, conversions of microparticle counts to mass values requires assumptions … and the bad assumptions upon which common conversions rely give bad estimates.

The Newcastle researchers described three different models for converting particle counts to plastic mass. Three models and three different values estimated for the global average rate of microplastic ingestion or GARMI. The three calculated GARMIs are 5, 0.1, and 0.3 g per week. Statistical error was propagated but not used in reporting the results. Instead, application of the three different models to the 102,527 particles per year average ingestion was presented as a range. The researchers reported: “Preliminary estimates indicate that humans could potentially be ingesting between 0.1-5g of microplastics per week.”

This statement sounds like it’s expressing a statistical range, but it isn’t. A single fixed count is converted to mass in three different ways, giving three different answers. A more averaged method was reported, concluding, “a combination approach of the scenario resulted in a GARMI of 0.7 g per week.” Three different models, one giving 0.1 g per week, one 0.3 g per week, one 5 g per week and the combination approach giving 0.7 g per week of plastic ingestion became, “on average people could be ingesting approximately 5 g of plastic every week.” Thus, the 5 grams stuck.

Drinking water accounts for more than 90% of the assumed consumption in the assumptions leading to the 5-g estimate. That’s about 250 particles per day in a liter of water. Work before and after the study show the source of microplastic consumption is the particles introduced during treatment and bottling — particles abraded off equipment. Ocean plastic is not the major source. Yet, the WWF report shows pictures of ocean plastic and plastic trash.

Researchers at Wageningen University released a study questioning the Newcastle findings. Their study published in March 2021 accounted for the size ranges and particle shapes when pulling data from other studies. The researchers obtain a median plastics consumption through microparticles of only about 4.1 µg per week for adults — a million times less than 5 g. They address the 5-g per week assertion concluding it “does not represent the intake of an average person.” Even the 0.1-g per week estimate is 25,000 times too high. The number of particles consumed is not the source of the discrepancy. Indeed, the particle counts are about the same. The difference is only in the assumptions used to convert count to mass.

Martin Pletz published an article critical of the Newcastle study in 2022. He concludes errors in the assumptions led to overestimating GARMI by several orders of magnitude. Pletz details where the faults lie — largely in the way data sets were combined without correcting for particle size. Pletz concludes, “it is obvious that both the maximum and minimum [microplastic consumption] calculated by [the Newcastle researchers] considerably overestimate the [microplastic] ingestion.” Pletz references the Wageningen study concluding “5 g per week is by a factor of about 10⁶ higher than [the Wageningen study’s] median value, which means that a human eats a credit card worth of [microplastics] not every week but every 23 thousand years.” Recent studies continue to support that 5 g per week is five to six orders of magnitude too high. This is not slightly off — it’s grossly off.

The sequence of public discourse may make it appear as though science is working. Work is published, others read the work, find fault, publish results, and the record is corrected. But that didn’t happen here. As others have detailed, mass media is partly at fault, but so is scientific literature.

Researchers from University of Vienna Medical School published a paper in March 2022. Their paper, “To Waste or Not to Waste: Questioning Potential Health Risks of Micro- and Nanoplastics with a Focus on Their Ingestion and Potential Carcinogenicity” cited the Newcastle study in background. They state: “translated into more imaginable numbers, on average we ingest five grams of [microplastics] per week per person (roughly corresponding to the mass of a credit card) depending on the region in which we live, our lifestyle, and diet.” The quote isn’t in the meat of the study. It isn’t their finding. It is just there to provide background context for why examining the health risks of microplastics is worth studying. They submitted the paper in October 2021. The Wageningen University study was out at the time, but not referenced. The Pletz study was not yet available.

The result was that the 5 g per week value appeared in another peer-reviewed paper. The University of Vienna issued a press release. “Five g of plastic particles on average enters the human gastrointestinal tract per person per week. This is roughly equivalent to the weight of a credit card” opens that release. Science Daily distributed it. It was the third time the 5 grams per week appeared in that news cycle, this time referencing the Vienna work and not the Newcastle study it quoted. The New York Post, Newsweek, and others also quoted the Vienna press release. It was in the background of the article, but it became the story. In fact, it is still being referenced. The Economist in March 2023: “About five grams of plastic particles enter all our gastrointestinal tracts every week. That’s about the weight of a credit card.” Attribution is the Vienna press release. Picked up in another news cycle, 5 grams per week was again widely reported.

I reached out to the author of the Newscastle study, informing him I was writing this story. Thava Palanisami and I exchanged several emails. I asked whether the world’s fixation on 5 grams per week based on his work was consistent with current thinking. I asked about the Pletz work and whether he wanted to offer a rebuttal. When I asked about pulling quotes from our exchange, he provided this quote instead: “The authors have confirmed they are still behind this significant work which has been helping to reduce the adverse impacts of plastics, we will stand by it until someone proves that there is ZERO plastic exposure to humans.”

The goal given in the paper is to “derive a global average rate of microplastic ingestion to assist in the development of human health risk assessments and effective management and policy options.” That’s a worthwhile goal. It is important to know how much we’re ingesting. It matters whether it is 5 g or 5 µg per week. The quote provided does not address the accuracy of 5 g.

Palanisami is still active in researching microplastics and publishes frequently. His recent publications continue to reference the 2021 5 grams per week paper and do not reference either the work of Pletz or Koelmans that question its conclusions. This feels like science is failing. Work of others, work identifying and correcting errors, and work critical of the initial study is being ignored.

I had a far more fruitful discussion with Bart Koelmans, part of the Wageningen University team. He continues investigating microplastics and publishes frequently. Bart uses the chopstick metaphor in his talks, stating that weekly consumption is a grain of salt between two chopsticks. He stands behind his 2021 work indicating micrograms per week is more representative of average consumption. He pushed back hard on my assertion that science had failed.

“When we saw what they did,” he said, “and we thought that’s obviously not correct. That should be done in another way,” said Koelmans when referring to the Newcastle study.

He went on to explain his group completed its study and published it — improving the method and producing a more accurate appraisal. His work is prominent in the WHO evaluations of microplastics, though it hasn’t stopped 5 g per week being quoted in other places. Koelmans concluded the value of the credit card visual played a role. “The credit card is a strong meme,” he began. “People start being scared a bit. It’s not something that you forget so easy.” Stories correcting the record are hard to get written, he laments, and, when they are, don’t attract attention.

My interest in this story has taken many turns. I purchased a couple of microscopes and have examined many water, beverage, and environmental samples. Microplastic particles are truly everywhere. I’m shocked by particles in every lake or river I sample. Shocked by the number of tire particles. Shocked by the number of particles in water bottles. I know the literature measurements showing thousands of particles in a water bottle are correct. I counted. With my own eyes, I’ve also confirmed 5 grams per week is much, much too high to describe the particles per liter I’m observing. The particles are there, but they are small. Yes, 5 grams is simply crazy.

I was skeptical back in 2019 when I first read 5 g per week. I knew it couldn’t be right. Models need to be tested to see if the results match the inputs. Checking the model results against the inputs shows the model giving 5 g per week is impossible. The weight of plastic consumed comes from counting particles. The model gives 5 g based on 253 particles weighing 0.64 g present per liter of consumed water. The average plastic particle, if spherical, would be 1.5 mm in diameter. Particles of this size are easily seen and easily collected for weighing. The 253 particles could be smaller, but then they would weigh less, unable to add up to the 5 g per week the model output. The model giving 5 g just can’t be correct.

The 5 grams of plastic per week is troubling in several ways. It illustrates the power of meme-promoted science. A compelling, if misleading, graphic can cement fact into the public consciousness. It shows few people take time to ask whether a value they hear feels correct. It shows even nonsensical values can become accepted. It shows when facts are widely reported, they don’t get corrected, and that clicks and eyeballs are more important than truth. Erroneous data, once on the internet, still influence public perception and legislation. It shows that researchers are hesitant to back away from results even when those results are questioned. It shows there is a blurry line between advocacy and factual reporting. It shows the testing and revision — key parts of science — can and are failing.

Information moves quickly today. Wrong information, correct information, it doesn’t matter. In contrast, corrections move at more plodding pace. 97.9° F as human body temperature won’t be quickly adopted. Academic publishers could better flag work that has been superseded, linking to more up-to-date studies. Outdated NGO reports, in an ideal world, would be updated. Sadly, motivation is lacking. Corrections don’t get clicks. But hastening the correction of facts is a worthy goal. Science demands we try.