But cause and effect is less clear when it comes to nutrition, and so most studies can merely show associations, rather than causes.
"Associations are a weak form of science; they help us ask questions but they rarely give us definitive answers," Katz said. For instance, he said, few people with television sets get tuberculosis, although the two have nothing to do with each other.
Small studies may help scientists design future, larger studies, like a study of whether pistachios can help prevent lung cancer, but they typically do not stand on their own.
"This study is an invitation to scientists to start asking questions. This is not a finding that's ready for prime time," said Katz.
Even if pistachios don't prevent lung cancer, they do have other benefits.
"I think we already know from multiple lines of research that nuts are nutritious foods," said Katz. "I wouldn't put pistachios quite at the top of that list," he added, saying he would put walnuts first, followed by almonds.
And people shouldn't jump to eat the latest superfood in excess -- or even at all if they don't like it.
"We don't want to hang our hat on any single food as the answer in disease prevention," said Connie Diekman, the director of university nutrition at Washington University in St. Louis. "Single studies should not give the consumer a message that says, 'I must do.'"
But that can change as more studies confirm possible benefits from a food. One example making headlines recently is coffee, which is being shown to have benefits, but by multiple studies over many years.
In the end, though, Diekman said, individual food studies should show a consumer an overall message of a balanced and varied diet rather than one of choosing superfoods.
"The food we eat does play a role in our overall health. How and what that role is and what it looks like is less clear, but the food choices you make are going to make a difference to your health," she said. "I hope what they hear in that message is think as you make food choices, as opposed to doing whatever's the easiest, the fastest."