It’s something we hear from policymakers again and again: The world squanders too much energy. And wringing out that waste should be one of the easiest ways for the United States and other countries to save money and curb pollution.

But as it turns out, much of what we know about the topic of energy-efficiency is still fairly hazy. Sure, it’s technically doable to make cars more fuel-efficient or insulate homes to prevent heat from leaking out. But which of these efforts are really the most cost-effective? And if it’s such a no-brainer, why aren’t people already taking these steps?

The fact that we still don’t have great answers to those questions is what inspired a group of economists at MIT and the University of California, Berkeley to launch a big new project, called E2e, that will try to apply more scientific rigor to the whole topic of energy efficiency.

“Almost all of the previous work on energy efficiency comes from engineering studies, which look at what’s possible under ideal conditions,” says Michael Greenstone, an economist at MIT and co-director of the E2e project. “We wanted to ask a slightly different question — what are the actual returns you could expect in the real world?”

Here’s what he means. In 2009, McKinsey & Co. released an eye-popping study demonstrating that the United States could hugely improve the efficiency of its homes, offices and factories, through strategies like sealing leaky building ducts and upgrading old appliances. By doing so, McKinsey estimated, the country could save $680 billion dollars over 10 years and do the climate equivalent of taking all the nation’s cars off the road.

Yet as economists scrutinized those numbers, they realized the picture is more complex. ”Those engineering studies can’t account for the behavioral changes you might see in response to efficiency improvements,” says MIT’s Christopher Knittel, who also co-directs the E2e project. “People could, for instance, start adjusting their thermostat if it becomes cheaper to cool the house.” (This is known as the “rebound effect.”)

Ideally, says Knittel, researchers would start conducting rigorous, randomized controlled trials to find out precisely how effective various efficiency policies are. The E2e  Web site lists some of the detailed work that has been done on this front — though there aren’t many such studies.

One recent study of Mexico, for instance, found that a government program to help people to upgrade their refrigerators with energy-saving models really did curtail electricity use. However, a similar program for air conditioners had the opposite effect — when people got sleeker A/C units, they used them more often, and energy use went up.

“The point is that policymakers aren’t going to spend an infinite amount of money trying to save energy or reduce greenhouse gases,” Greenstone says. “So the motivation is to find the places where the return is the greatest. If you could reduce a ton of carbon-dioxide for $100 or two tons for $50, you’d choose the latter.”

The researchers are also asking why, if it’s so compelling, people and businesses don’t already take steps to become more energy efficient. Is it because people aren’t aware that they can? Are there actual market barriers that could be addressed by policy? (For instance, landlords may have little incentive to invest in energy-saving appliances for their tenants.) Or is it just that the purported savings aren’t worth it in the first place?

“It’s easy to come up with conjectures for why people aren’t choosing more efficient options,” says Catherine Wolfram, an economist at the Energy Institute of Haas in Berkeley. “Maybe people don’t have the right information, maybe people are procrastinating. But right now, these are just stories. It’s an area where we need more evidence.”

Some work is being done on this front. Knittel, for instance, is conducting an experiment to see whether people will buy more fuel-efficient cars if they simply receive more detailed information about gasoline costs and mileage. Greenstone and Wolfram are carrying out a randomized controlled trial to scrutinize a U.S. government program to help weather-proof the homes of low-income people.

“Part of the reason we started this project is that efficiency is one of the few areas where there’s broad agreement across the political spectrum that these are policies we should be pursuing,” Greenstone says. “And we want to be able to show what actually works and what doesn’t.”