[♪ INTRO ].

It seems like it’s practically impossible to get a straight answer on whether a given food or supplement is good for you. First they say red wine is good for you.

Then there’s a headline announcing that it doesn’t do anything, and wait, isn’t alcohol supposed to cause cancer? If you wanted to change up your diet to accommodate all the latest news reports, you wouldn’t get very far before you’d run into all kinds of contradictions. I mean, you can’t both drink and not drink alcohol at the same time.

And the problem goes deeper than the over-exaggerated headlines on cable news. Often, the way nutrition research is done makes it really hard to translate the results in to practical advice because of imperfect comparisons. The effects of one compound might not be the same as the effects of the entire food.

What happens in animals might be different from what happens in humans. And a food that’s good for you in some ways might be bad for you in others. But it is possible to untangle all that by examining these studies more closely.

You just need to know what you’re looking for. Take what happened with raspberries not too long ago. In a 2011 study, researchers in China found that red raspberries could improve blood pressure.

Annnnd then in 2013, a team in Finland found that fresh berries didn’t affect blood pressure at all. There were plenty of differences between these two studies, though — and plenty of reasons they might have had different results. The Chinese study used raspberry extract, and was done on rats … specifically, a strain of rats prone to high blood pressure.

Whereas the Finnish study used berry puree, and the subjects were humans with symptoms that increased their risk of heart disease, like high blood pressure. The first problem with comparing them is that the other ingredients in the berries could’ve interfered with the effects of the compounds in the raspberry extract alone. Then there’s the whole “humans vs other animals” thing.

One reason scientists use animal models is that you can control everything about them. You can administer every speck of food they eat, control how much they exercise, and monitor them day and night. Try to do that with a human and see how long it takes to get punched in the nose.

But that means in this case, the humans who got the berries might have been doing or eating something else that negated the effects. Or hypertension-prone rats might be more sensitive to the blood pressure-lowering effects of berries than normal, if somewhat unhealthy, human beings. Comparing those two studies is a little bit like comparing apples and oranges … or apples and raspberries, I guess.

Different types of studies are important. It’s usually much easier and safer to test things in other animals before moving on to humans, and scientists in all fields of research rely on multiple lines of evidence to form conclusions. That’s how science works!

But when it comes to nutrition, most individual studies are intended as a piece of a larger research puzzle, not something for the home consumer to take to the grocery store. Studies on humans might be easier to translate to your life, but even those are fraught with complications. Like when a 2014 study on red wine contradicted the findings of another paper from two years earlier.

Both studies were specifically concerned with a compound in wine called resveratrol, which is an antioxidant — meaning it helps prevent potentially damaging chemical reactions in cells. In 2012, a randomized controlled trial by Hungarian researchers found that adults who received resveratrol supplements showed an improvement in a bunch of signs of cardiovascular disease. Randomized controlled trials, or RCTs for short, are the closest you can get to a controlled experiment in humans without the aforementioned nose punching and a whole passel of ethics violations.

Participants are sorted into groups and statistically randomized so the experimental and control groups are as uniform as possible. Then you give the experimental group the food or supplement or specialized diet you’re studying, which should ideally be the only thing that’s different between the two groups. And this RCT seemed to show that resveratrol made a difference.

But then, in 2014, an observational study of adults in Italy didn’t find any effect of resveratrol on signs of cardiovascular disease or mortality. Observational studies are different from RCTs. They aren’t really experiments — they involve gathering data that’s already out there, in this case by interviewing people about their eating habits and analyzing their urine.

RCTs are usually thought of as the most rigorous type of human study, but that doesn’t mean they’re always right and observational studies are always wrong. The 2014 observational study found that while participants had a lot of resveratrol in their diets, like from wine, the stuff might not have reached a high enough concentration in their bodies to have any biological effect. So the lesson there might be that supplements are more effective.

Which brings us to a fairly infamous example from the field of cancer nutrition. Beta carotene is a dietary precursor to vitamin A found in foods like carrots and sweet potatoes. A body of evidence from multiple observational studies suggests that eating fruits and vegetables rich in beta carotene lowers the risk of lung cancer.

Which was the justification for a number of clinical trials, including one in Finland in 1994 and one in the US in 1996, which gave beta carotene supplements to smokers. Except, the smokers who got beta carotene had a HIGHER rate of lung cancer. You really don’t want to give your study population cancer, so this came as a bit of a shock.

In the fallout, it was found that beta carotene in high doses could actually act as a pro-oxidant -- a compound that’s the opposite of an antioxidant, and promotes chemical reactions that lead to cellular damage. So the observational studies showed that carrots are good for most of us, but the clinical trials discovered that under specific circumstances, high doses of beta carotene have a pro-cancerous effect. These findings weren’t exactly contradicting each other … it was just another case of comparing the wrong things.

So that’s another thing to look out for with nutrition studies. When you dig into the details, findings that seem contradictory might turn out to be building on each other. The final -- and probably most important -- question to ask when you see a report announcing that something is good for you is, what does “good for you” even mean?

There’s no universally agreed upon definition for it, and it’s never what researchers specifically set out to study. Trying to figure out whether something is “good for you” is just too broad. Studies might look at rates of stomach cancer after a certain number of years, blood glucose levels, cholesterol levels, or any of the other incredibly specific markers related to cancer or diabetes or obesity or some other health issue.

Scientists who study nutrition call these outcomes, and they’re probably the most important thing to look for in any nutrition study. Going back to the wine example, red wine contains a whole bunch of plant compounds that are targets for nutrition research — resveratrol is just one of them. At the risk of stating the obvious, wine also contains alcohol.

And alcohol itself might actually be good for you in some very specific ways. A 2002 analysis of several dozen studies of the effects of wine and beer found that overall, wine drinkers were less likely to die from heart disease, or to experience other, less fatal heart problems. But the association was there for beer drinkers too.

It was statistically weaker, but clearly the effect wasn’t just coming from the special plant compounds in wine. This finding has cropped up more than once, and there’s enough evidence that some nutrition experts might recommend light or moderate alcohol consumption to help prevent cardiovascular disease. Does that mean alcohol is good for you?

Well. I mean. It’s alcohol.

Even if it helps your cardiovascular system, it also messes with your liver, interferes with fetal development, increases the risk for certain cancers, depresses your central nervous system and a whole list of other things. In some of these cases, the dose makes the poison. The evidence points toward light or moderate drinking having certain positive effects, but heavy drinking being bad in basically every way.

While alcohol is an extreme example, it’s not hard to find cases where a particular food or nutrient has a positive effect on some outcomes and a negative effect on others. These studies aren’t contradicting each other because they’re not trying to answer whether something is universally “good for you.” But when the findings show up in the news, sometimes that nuance gets lost in translation. When it comes to nutrition research, there are so many seemingly contradictory findings coming out all the time that even the scientists publishing them need to step back and look at the evidence as a whole.

That’s why the field is full of review articles and meta-analyses. The process of science is pretty much always this messy, but for some reason with nutrition it’s on public display. Maybe that’s because we want to be able to justify our craft brew hobby by believing it’s healthy, or because we’re always looking for ways to live longer and stave off diseases.

Maybe it’s because what we eat feels like the one thing we can control in a world full of unknown risks. But it helps to take every individual nutrition study with a grain of salt. Not too big a grain, though — I just read that salt’s bad for you.

Or was it good for you? Thanks for watching this episode of SciShow, which is produced by Complexly. For other awesome Complexly videos about nutrition and health research, check out Dr.

Aaron Carroll over at Healthcare Triage. [♪ OUTRO ].