In a new type of advanced government class at Seattle’s Garfield High, the students rarely sit quietly taking notes while their teacher stands and lectures.

Instead, they debate each other. They write legislation. They run for president in mock elections and pretend they’re lawyers arguing cases before the U.S. Supreme Court.

They sometimes even stand up and holler, as Sanai Anang did recently, playing a member of a Virginia-based group that lobbies for strict immigration controls.

In a simulated public hearing, Anang, who loves to ham it up, jumped to his feet without being recognized and declared, in a mangled Southern accent, “Ee-lee-gals come over and take our jobs. They don’t bee-long here.”

His classmates and teacher Jerry Neufeld-Kaiser cracked up.

They are all part of a teaching experiment that began six years ago in the Bellevue School District when a handful of frustrated government teachers teamed up with University of Washington researchers and turned the usual Advanced Placement curriculum inside out.

Instead of lectures sprinkled with discussions and occasional projects, they put role plays and simulations at the center of the curriculum — the entree, rather than a side dish or dessert.

Their goal was to solve two problems with the A.P. program, the largest set of college-level courses offered in high schools across the nation.

First, they wanted to address the criticism that A.P. classes cover so many topics so quickly that students spend too much time memorizing facts and too little time analyzing their meaning and significance.

The team also wanted to test whether a steady diet of hands-on exercises would help address the rising failure rate on A.P. tests among some minority groups.

The team members started with A.P. U.S. government and politics — one of the most popular A.P. offerings — dumping most of the lectures that usually are the core of the course, and replacing them with five in-depth projects.

They then tackled A.P. environmental science and are now working on A.P. physics.

The transition hasn’t been easy for students used to being told, at the start of each assignment, exactly what they’re supposed to learn.

Students and teachers alike complain the projects can be time-consuming to complete — and to plan.

And, done poorly, they can be a waste of time.

But the results so far are promising, showing that the project-based classes can provide depth and enough breadth for students to pass the spring A.P. exams.

Students in the experiment, now under way in about five dozen classrooms in Washington, Northern California and Iowa, have done as well and often better on the A.P. exams compared with classmates in the experiment’s control schools that use a lecture-heavy approach.

They’ve often scored higher on a separate test that researchers designed to probe how well students truly understand what they’ve learned — although those results have been mixed.

The researchers are not examining results by race because they believe achievement gaps are grounded in differences in class rather than ethnicity. They have found their approach can yield results for students from low-income homes as well as those from middle- and upper-class neighborhoods.

Last year, for example, 88 percent of students in two of the experiment’s high-poverty schools passed the A.P. U.S. government test in the spring — much higher than the 24 percent for comparable schools nationally.

Program booming

The A.P. program began in the 1950s as a way for elite high-school students to earn college credit.

In the past two decades, participation in the program has exploded, with more than 2?million students taking one or more A.P. exams last year.

But as A.P.’s popularity has grown, so have questions about its quality.

The spring A.P. exams are supposed to cover what students would learn in typical introductory college courses, but many teachers complain there’s so much material that their classes turn into extended cram sessions.

Critics also question whether the A.P. boom, driven by a push to open the program to all interested students, sets up those without strong preparation to fail.

The program is run by the nonprofit College Board, which is addressing those same concerns itself, steadily streamlining the exams to allow students more time for in-depth study.

The board is watching the teaching experiment carefully, interested in its promising results. In 2012 the board invited project leaders to its A.P. conference to present their ideas to A.P. teachers from across the nation.

It’s important that students gain an in-depth understanding of a subject, said Auditi Chakravarty, an A.P. program vice president. “And that requires more than the passive sit-and-get kind of learning.”

Enlisting an old idea

The A.P. experiment that started in Bellevue grew out of conversations between the U.W. researchers and a former Bellevue superintendent, Mike Riley, who’d led a big expansion in A.P. participation in his district.

They thought they could improve the classes by using an idea that dates back to the 1890s, when education reformer John Dewey promoted “l(fā)earning by doing.”

At its best, project-based learning can help students grasp the importance of their lessons and retain more of what they learn. At its worst, it can be entertaining but little else.

The research into its effectiveness is mixed, in part because the project approach can mean so many different things.

The U.W.-Bellevue team members dubbed their approach rigorous project-based learning, to distinguish it from unfocused efforts that have given the term a bad reputation.

They didn’t throw out traditional instruction altogether. Students still take tests and do homework. They still take the regular A.P. test at the end of the class.

The team spent a year planning the first project-based class in U.S. government, extending it to a yearlong course and searching for projects they could adapt so they didn’t have to create them from scratch.

Then it recruited the experiment’s first set of students, promising a still-tough but more engaging — even fun — A.P. experience.

Bumpy ride at first

The first year turned out to be tougher than many anticipated.

Some students complained they didn’t know what they were supposed to be learning, and they struggled to work productively in teams.

Many worried they wouldn’t be ready when it came time to take the A.P. test in the spring — and so did some of their teachers.

“To be quite frank, I didn’t think I was giving them what it took,” said Newport High teacher Tim Shultz.

Some teachers still lament that the course now takes more time, which means students can no longer take a common companion course — A.P. comparative government — in the same school year.

The experiment has been costly, too.

To date, the team has raised about $6?million to support its work, with half from the George Lucas Educational Foundation and the rest from other sources, including the National Science Foundation and the Bill & Melinda Gates Foundation. (The Gates Foundation is the main funder of The Seattle Times Education Lab project.)

Still, Shultz and other teachers embrace the new approach.

They love seeing students stop counting how many points an assignment is worth and instead lose themselves in planning a political campaign or lobbying for a bill.

“What I was doing in the past was teaching to the test,” said Shultz. “I’d say, ‘Know these 50 cases and you’ll be fine on the test.’ ”

Now, he said, he teaches students how to use legal precedents to help them make strong arguments before a mock Supreme Court, only sometimes adding, “Oh by the way, it’s also on the test.”

Teachers also say the approach helps many students who don’t come from privileged backgrounds and may not regularly talk politics over the dinner table. Until they play a legislator or a judge or a candidate, said Garfield’s Neufeld-Kaiser, they may have no concept of what those people do.

But after they write a bill and lobby their classmates to vote for it, he said, they get it.

“It’s so much more accessible because they’ve lived it.”

For the second year, teachers dropped some projects and revised others, and they figured out ways to grade students on their individual contributions — one way to avoid one student taking over and doing the bulk of the work.

Some teachers warned students they might feel uncomfortable with the new approach but that they should trust the process, and that students before them had passed the test and they could, too.

“Engagement first”

Each project follows a common set of principles based on research into how people learn best.

One is to immerse students in a challenge, then follow with lectures and reading to help students figure out how to meet it — an approach the researchers call “engagement first.”

They also design each course around a master question, which students circle back to after each project, ideally gaining a new level of understanding each time.

John Bransford, a well-regarded learning expert and a member of the U.W.-Bellevue team, said the hope is to help students gain expertise much like musicians improve with repeated guided practice.

At Bellevue’s Sammamish High earlier this year, a project on the federal budget illustrated what this concept looks like in practice.

Teacher Katie Piper first showed a documentary that presented a troubling picture of the country’s growing debt and taught students a little about entitlements and economic theory.

Then she sent them off to come up with a proposal that would significantly reduce the debt, with elements that would appeal to both Democrats and Republicans.

Within two days, students were deep in discussions about Social Security benefits, tax loopholes, Medicare and the Bush tax cuts — concepts some barely understood when they started.

The point was not to make the students financial experts but to give them more insight into how government works — in this case, all that goes into passing budgets.

On the last day, three groups presented their proposals to the class and an invited expert, the city manager of nearby Newcastle.

Seve Sandomirsky, 17, worked hard to sell his team’s plan, which was heavy on liberal solutions such as closing tax loopholes for corporations, and light on anything that might appeal to more conservative lawmakers.

Dressed for the presentation in a blue dress shirt and tie, he hoped to win over everyone with a sense of fairness, saying companies have long avoided taxes that they should have been paying for years.

“I had to dig deep,” he said later, “trying to sell this as a bipartisan deal.”

Students buying in

Sandomirsky expressed enthusiasm for the project approach, even while acknowledging it can be a lot of work.

“The greater understanding,” he said, “is so much more enriching than having a lecture and regurgitating information.”

Some of Neufeld-Kaiser’s students said the same.

“Instead of reading about what people are doing, you get to step into their shoes,” said Israel Brown.

Rather than test prep, “this is more like real-life prep,” added a classmate, Rahel Solomon.

Neufeld-Kaiser and a few other teachers like the approach so much they use it in non-A.P. classes as well.

At Sammamish High, the faculty, inspired by what they saw in the project-based A.P. classes, are redesigning most of their core courses in a similar way.

Some teachers emphasize the approach is not best for all students — that some learn just fine through lectures.

Newport teacher Virginia Evans, while a fan, also wonders if it helps or hurts students when they go to college.

“The reality is, colleges are like my non-project classes,” she said. “They lecture at you, and you write papers.”

The researchers aren’t declaring complete victory yet.

That’s partly because the early results, while promising, could reflect the so-called “early adopters” effect — that any program with enthusiastic teachers will at first show strong results that peter out when used more widely.

The research team will continue its study for a few more years, hoping to amass enough evidence to convince many more schools that rigorous project-based learning can enhance advanced classes.

Team members don’t back the notion that all high-school students should take college-level classes, but for those who do, they want the courses to be the right kind of tough.

They are still debating exactly what that is, but they know what it isn’t.

“Have you really learned something if you’ve memorized a bunch of definitions?” asked UW professor Walter Parker, one of the experiment’s lead researchers.

“It’s probably some kind of learning. But it doesn’t make the grade as deep learning — meaningful learning.”

Linda Shaw: 206-464-2359 or lshaw@seattletimes.com. On Twitter @LShawST