A recent study, and the various blog posts and news articles covering the study, seems to make a baffling and contradictory claim: It is possible that adjusting our diets from meat-heavy to produce-heavy could actually result in an increase in greenhouse gas emissions. This flies in the face of almost every assumption and previous study, but the writing is clear as day. One image caption on the study's press release states it in no uncertain terms: "Eating lettuce is more harmful to the environment than eating bacon."
The press release for this study, from Carnegie Mellon University, in Pittsburgh, Pennsylvania, titled “Vegetarian And ‘Healthy’ Diets Could Be More Harmful To The Environment” was published on December 14th. The release was created to alert people to a new study from scientists associated with the university, but was not actually written by the scientists who conducted the study. Here are some headlines from articles that covered the study (notably, through the lens of the press release), which was first published online on November 24th:
- Lettuce Produces More Greenhouse Gas Emissions Than Bacon Does (Scientific American)
- Vegetarians Cause More Greenhouse Gas Emissions Than Meat Eaters: The Worst Foods For The Environment, According To The USDA (Medical Daily)
- Vegetarian diet more harmful to environment (Economic Times)
- Lettuce ‘three times worse than bacon’ for the environment, scientists claim (The Telegraph)
- Salad Is 3 Times Worse For the Environment Than Bacon: Study (Details)
These are all wrong, and belie a fundamental problem (perhaps a set of problems) in the world of science journalism. The basic issue is the lack of communication—and often of understanding—between the scientists who do the research and the people who write the press releases, and a further problem of laziness from journalists who merely parrot the releases. The people who write releases are crafting inflammatory, often incorrect documents about scientific research in order to snare the attention of journalists, who all too often don’t bother looking into the research at all. And why bother? When you have to write six or eight posts a day, and a press person at some university writes up an attention-grabbing headline, it’s an awful lot easier to just rewrite the release than it is to actually look into whether the document accurately represents the research—let alone whether the research itself is any good.
The latest victim of this trend is Michelle Tom, who co-authored the Carnegie Mellon study. Her title was “Energy use, blue water footprint, and greenhouse gas emissions for current food consumption patterns and dietary recommendations in the US.” That report, which looks into the environmental effects of different types of food, was retitled “VEGETARIAN AND “HEALTHY” DIETS COULD BE MORE HARMFUL TO THE ENVIRONMENT” by Shilo Rea, a director of media relations at the university, where Tom is a Ph.D candidate. (Shilo “represents the College of Humanities & Social Sciences, including the areas of psychology, decision sciences, behavioral economics, education and literature,” according her bio. She did not respond to requests for comment.)
“It seemed to me like they were really trying to be controversial.”
“I think the title of this press release is definitely misleading and is not an accurate portrayal of our research,” Tom told me yesterday by phone. She went on to say that she wished she’d gone further in putting her foot down about the wording of the release. “I never met the person who wrote this,” Tom said. “I did take a look at it and made a few revisions. I should have had them change that title.”
“It seemed to me like they were really trying to be controversial,” said Kai Olson-Sawyer, a Senior Researcher and Policy Analyst with the GRACE Communications Foundation, a non-profit based in New York City that studies the connections between food, energy, water, and emissions and campaigns for sustainable options. (He focuses on water use.)
What Does the Study Say?
After drilling down through the study and talking to Tom and to Olson-Sawyer, here’s my understanding of what the study is actually trying to say: Not all vegetables and fruits have an equal impact on the environment (it looks at water use, energy use, and greenhouse gas emissions), and it’s even true that with some very careful rejiggering, you can create a possible produce-heavy diet that is worse for the environment than a meat-heavy one.
In her study, Tom sets up three possible scenarios, all related to the current caloric intake of the average American (they figured this at about 2,390 calories per day on average, with about half again that much in food waste) and the United States Department of Agriculture’s (USDA) recommendations regarding the percentages of each food group (grains, fruits, meats, vegetables, dairy) we should be eating. Those are all laid out in the USDA’s current nutrition guidelines, though Americans still have a ways to go before the reality matches the recommendations.
- Scenario 1: We continue to eat the same types of foods in the same proportions—meaning very meat-heavy diets, unlike the more produce-heavy USDA recommendations—but we reduce our calorie counts to something sensible. (Tom and her co-authors created their own ideal calorie counts based on activity levels.) That would mean instead of eating a Big Mac, an average eater might switch to a simple McDonald’s cheeseburger: basically the same ingredients, but smaller and thus lower in calories.
- Scenario 2: We keep the calorie count the same, but we change the mix of food types to match the USDA’s recommendations. This one is complex: It means instead of eating that Big Mac, which has 563 calories, you’d have to split that 563 calories into about 150 calories of vegetables, 100 calories of fruits, 100 calories of protein, 100 calories of whole grains, and 100 calories of dairy.
- Scenario 3: We reduce our calorie counts AND change the mix of food types we eat. So we’d have the same mix of fruits, vegetables, grains, protein, and dairy as in Scenario 2, but less of them, calorie-wise.
According to Tom’s findings, only Scenario 1 would result in a positive effect on the environment. This seems totally unexpected; we are told, again and again, that eating more fruits and vegetables and grains is good for the environment, that livestock emissions are a huge problem, and that changing our diet to be more like the USDA’s recommendations—meaning, less meat—would have a positive effect on the environment as well as our bodies. Tom’s study says this isn’t the case, and journalists took notice.
Lettuce vs Bacon: The Final Debate
The press release, and the coverage that followed, relied heavily on one very weird comparison: lettuce and bacon. This seems to come from one quote given by Paul Fischbeck, one of Tom’s advisers: “Eating lettuce is over three times worse in greenhouse gas emissions than eating bacon,” he said, according to the release. Tom was reluctant to say much ill about Fischbeck, for obvious reasons, but did say: “Even if that is accurate to some extent, that’s probably not the best title to our article.” Fischbeck served as a guide during the creation of the study, but did not write either the study or its press release. Regardless, Rea, the press officer, seems to have singled out his quote as the most important element of the study.
To compare [lettuce and bacon] using the information provided by this study is…insane.
Regardless of Tom’s misgivings, you can see why a media person might have grabbed on to this—the two items fairly scream “healthy” and “unhealthy.” But to compare them using the information provided by this study is…insane. The study is in fact not saying anything of the sort, and the list of caveats needed to make that statement mathematically work would run pretty long. You’d have to compare them by calorie count instead of by weight or by nutritive benefits. You’d have to assume that any human would replace a protein source with a mostly water-filled leafy vegetable, which not even the USDA is recommending. (The USDA’s dietary guidelines do, in fact, include meat.) And you’d have to assume the figures on pork processing presented in the study are accurate, which I am not sure they are (given the lack of available data on full lifecycle emissions, which I’ll get into in a bit), and that the lettuce was grown in California. (Not a bad bet, but not a given: 90 percent of the country’s leafy lettuce comes from California, and 83 percent of its Romaine lettuce. If you’re eating lettuce from anywhere else, these figures are way off, which I’ll get into in the Blue Vs. Green Vs. Grey Water section.)
“On the face of it, it’s kind of laughable,” said Olson-Sawyer.
The Real Culprit is Not Lettuce
Tom told me that the basic point of her study was to demonstrate that not all fruits and vegetables are good for the environment and that not all meats are bad—but, in fact, even that is sort of irrelevant: The variable that really threw off the figures wasn’t the calorie counts in lettuce. The true X-factor in all this is sugars, fats, and oils. These items do not appear in either the abstract of the study (written by the scientists) nor in the press release, but a short interview with Tom was all it took to reveal that the true powerhouse in the equation, the high-calorie, low-impact foods that gave the study its conclusion, are these, and not lettuce. Tom wasn’t hiding it—the press release was, intentionally or not.
“Dairy, vegetables, and fruits have higher greenhouse gas emissions per calorie than do sugars, fats, and oils,” explained Tom. “So the reduction in emissions due to our reduction in meat consumption is kind of offset by replacing sugars, fats, and oils with dairy, vegetables, and fruits.” Something like sugar is so high in calories that if you try to replace it with fruit—which, again, is not necessarily a likely outcome—you necessarily end up having to rack up pounds upon pounds of fruit to equal the calorie count of just a touch of sugar. (You’d have to eat 17 oranges to equal the calories in a cup of sugar.)
When I asked if these categories—sugars, fats, and oils—had such a powerful effect on the calorie count that they would render any discussion of lettuce irrelevant, she laughed. “Well, yes. You are kind of right about that,” she said.
Of course that makes sense, if you look at it from a caloric perspective. (Which, by the way, not everyone thinks you should; Scientific American spoke to a researcher who advocates looking at the entire nutrient makeup of a diet rather than calorie count.) Sugars, fats, and oils have spectacularly high calorie counts by weight, meaning they’re “efficient” from that very specific perspective. And they’re also represented strongly in our current diet, rather than in the USDA’s recommendations. (The researchers, by the way, used the existing recommendations from 2010, not the new ones that should come out next year.) So it computes that of the three proposed scenarios, only the one that’s super high in sugars, fats, and oils would score high in caloric efficiency. Screw lettuce: This is a study about corn syrup and olive oil.
But is the Study Even Accurate?
Getting away from the problem of the press release for a minute, there are a whole mess of potential issues with the study itself. For one thing, when looking at the impact of certain foods on water use, it only examined “blue water” (referring to withdrawals from rivers, lakes, and groundwater), completely ignoring either “green water” (rain) or “grey water” (reused water). “We felt that blue water footprint was a little bit more important given that California is drought-prone and a lot of our fruits and vegetables are grown and produced in California,” says Tom. That’s true, sort of; California produces, just for example, 99 percent of the country’s artichokes, 99 percent of its walnuts, and 95 percent of its garlic. And blue water is certainly more important in California than green water.
But that’s not true elsewhere. Olson-Sawyer notes that east of the Mississippi, green water is hugely important for agriculture. And California may be the country’s biggest agricultural state, but it’s nowhere near the majority; it exports less than 15 percent of America’s agricultural products, and employs only 27 percent of the country’s farm workers. In other words: California is important, hugely important, but there’s an awful lot of farming happening elsewhere. To ignore green water, which supplies a massive amount of water to thousands of farms, can totally throw off any water-use calculation.
“This is a major gap. It’s a giant variable.”
Another issue: We all know that livestock produces lots of emissions. (The UN estimates livestock is responsible for 14.5 percent of all anthropogenic greenhouse emissions.) But as it turns out, we know basically nothing about the total greenhouse gas emissions involved in processing and packaging meat. “Since most of this is proprietary, it’s up to the business to share that information. The public doesn’t necessarily know what the environmental impact of processing is,” said Olson-Sawyer. A 2014 study from the University of Nebraska-Lincoln backs that up: The total greenhouse gas emissions of the whole life cycle of livestock, from breeding to raising to slaughter to packaging, is, basically, a mystery. “Current methods used by the U.S. Environmental Protection Agency (EPA) associated with beef production in feedlots were found to account for only 3–20% of life cycle GHG emissions,” write the authors.
The lack of this data is infuriating to folks like Olson-Sawyer. “The data really isn’t there. We’ve looked, we’re always looking for research along these lines. This is a major gap. It’s a giant variable,” he said. When I asked Tom about this, she said, “There definitely could be some bias in our results if that’s the case.” She noted that their figures are presented in a range, but agreed that that range would be extremely narrow and not very helpful if their data only comes from a limited number of data sets which that 2014 study found to be unreliable.
This study is an interesting one, for sure; both Tom and Olson-Sawyer volunteered that, at its core, the report is a call for a more nuanced look at the way we eat and the way it affects the environment. Simply going vegetarian isn’t necessarily the answer, it says. In fact, there are (admittedly bizarre and unrealistic) ways that going vegetarian could actually be bad! But the study doesn’t suggest that vegetarianism is bad, or state that lettuce is “worse for the environment” than bacon. What it says, in a potentially flawed and messy but still valuable way, is that we have to think holistically about the way we eat, analyze it more thoughtfully and with a wider lens, if we want to have a positive impact.
And none of that is helped by the gaping flaws in science journalism. One of those issues is, well, would anyone have covered the study if it had been properly and correctly titled? A release titled “Study Finds That Certain Combinations Of Certain Fruits And Vegetables When Combined In A Way Sort Of Related To 2010 USDA Recommendations Can Possibly Maybe Have A Not Great Effect On The Environment” is nowhere near as sexy as what they came up with. Who knows, I might have ignored it too.