The few times that Huan Kiat Koh came home with poor grades when he was growing up in Singapore, he vividly remembers his mother's response.
"She would scold me, and then tell me to sit down and work," said Mr. Koh, a sophomore majoring in materials science at Carnegie Mellon University. "In general in Singapore, parents feel grades are very important for their kids.
"I think I read that American parents, when their kids get bad test scores, tend to be more sympathetic and worry about the kids' self-esteem," he said wryly.
That "tiger mother" approach may be one reason Singapore students consistently score at the top on international science knowledge tests, and American parents' more forgiving attitude may play some role in the less impressive U.S. scores.
In the 2007 TIMSS test, which stands for Trends in International Mathematics and Science Study, students from Singapore took first or second place in all science categories. The United States ranked 11th.
Results like this have generated concern at the highest levels.
President Barack Obama has repeatedly pushed to improve science and math education in America -- most recently this month. In a speech in Arlington, Va., he complained that "the quality of our math and science education lags behind many other nations."
Carol Johnson, superintendent of the Boston Public Schools, added her own concerns after a news conference on poor science test scores in urban districts last month.
"If we are truly to equip all our students to compete in the global economy, we must give them the tools to outthink and outperform their counterparts around the world, " she said.
In the midst of this hand-wringing, a larger question remains: Why do students who live in a nation that is the acknowledged world leader in scientific research lag behind other countries on science knowledge?
The answer is complicated, say experts, and involves everything from poverty, to poor training of teachers, to attitudes about learning, to the disturbing level of scientific illiteracy among American adults.
"If I were going to write the equation for improving science education in the United States, it would look like something Albert Einstein had put on the chalkboard," said Chris Mooney, co-author of the 2009 book, "Unscientific America: How Scientific Illiteracy Is Threatening Our Future."
A major reason for poor U.S. science test scores is the huge gap between students in affluent and poor school districts, which is often reflected in racial disparities in scores.
Jack Buckley, commissioner of the National Center for Education Statistics in Washington, D.C., said American student rankings look very different when white and black students' scores are considered separately.
In one of the major international tests, the European-based Program for International Student Assessment, or PISA, American students as a whole scored 502 in 2009, slightly above the industrialized nation average of 500.
But if white students are considered separately, Mr. Buckley said, their score would have been 532, which would have ranked them sixth, while African-American students as a group had a score of only 435, putting them between Bulgaria and Romania.
Economic differences play a key role in that discrepancy, said Arthur Eisenkraft, a science education expert at the University of Massachusetts at Boston.
"Certainly we've always known there are high correlations between poverty and how kids do in school," Mr. Eisenkraft said.
Demographics also help explain the high scores in places like Shanghai, which scored at the top of the last PISA test in science, Mr. Buckley added.
"Shanghai is a magnet for some of the best and the brightest in China," he said. "Saying that Shanghai is a representative sample of China is like saying Massachusetts is a representative sample of the United States."
The downside of demographic effects in America is particularly noticeable in large urban school districts like Pittsburgh.
Of the 17 big-city school districts that participated in the National Assessment of Educational Progress science tests in 2009, only two -- Charlotte, N.C., and Jefferson County, Ky. (which includes Louisville) -- had more than a third of their students scoring in the proficient range. More disturbing, eight of the districts had more than half of their students scoring below the basic level, ranging from Los Angeles at 55 percent to Detroit at 74 percent.
"And basic means basic," said Alan J. Friedman, a science education consultant who spoke at a NAEP press conference convened by the National Assessment of Educational Progress in Boston last month. "If you're doing below basic in science at the eighth-grade level, you may be freezing yourself out of a whole lifetime of career options and advancement opportunities.
"It's not a pretty picture."
In Pittsburgh, which did not participate in the NAEP testing, school-by-school results in state science tests are just as disheartening.
About 61 percent of Pittsburgh's elementary students scored at the proficient level or above on the test, compared with 81 percent statewide, and it gets worse from there.
At the middle school level, only 32 percent scored at proficient or above, compared with 57 percent statewide. Even at the relatively new Pittsburgh Science & Technology Academy 6-12, only 38 percent of eighth-graders were proficient or advanced. At the high school level, the figures slumped to 20 percent for Pittsburgh and 40 percent statewide.
In three of the city's high schools -- Oliver, Peabody and Westinghouse -- fewer than 5 percent of the students reached the proficient range.
Stephen Pellathy, a science curriculum specialist for the district, acknowledged that the drop-off in science test scores from the elementary years to middle school and beyond is a major challenge.
"Why the big drop-off? I guess that's the million dollar question," he said. "It's not one where you can point to a single factor. The kinds of things we are trying to do to improve those scores is to have more time with our teachers in terms of professional development and try to hire and train better teachers."
The Pittsburgh schools also are using an "inquiry-based" curriculum, known as Full Option Science System, which emphasizes hands-on experiments and getting students to solve problems rather than memorize facts.
So far, though, the new initiatives have not shown strong results in science test scores.
Part of the reason for that, Mr. Pellathy said, is a broad attitudinal problem that many American students and their parents share.
In Asian countries and other parts of the world, parents believe that doing well in science is mostly a matter of hard work. But in many American households, he said, there is a more of a belief in innate ability.
As a result, "whenever students have their first failures, they say, 'Well I'm not good at science and that's just how it is.' "
That notion -- that you either have the gift or you don't -- "is particularly damaging in science because the reality of science is you're basically always failing," said Mr. Pellathy, who has a Ph.D. in physics.
"When I was a student," he recalled, "one time I got two weeks of data out of a piece of equipment I had worked on for an entire year. So if you don't have that mindset that your failures are setting you up for your subsequent success, you won't keep going."
Besides the sheer value of hard work, it's also important to teach students how to analyze problems and figure out solutions, and not just accumulate facts, said the University of Massachusetts' Mr. Eisenkraft.
He recalled visiting an elementary school one time where the principal proudly told him that every student in the school was required to memorize the order of the planets.
"I said to him, 'I'm not going to argue with that, but when you have them learn that, don't do it during science time because it has nothing to do with science.' The science question is how do we know Venus is closer to the Sun than the Earth, or how do we know the Earth goes around the Sun?"
The NAEP test in particular has questions that ask students to do that kind of analysis, which may explain some of the poor U.S. results, especially if students are taking state tests that are more about memorizing scientific concepts.
Mr. Eisenkraft said he has seen some states where the state test scores track students' NAEP scores pretty closely, but others where "98 percent of the students are passing the state test, but only 12 percent of them are proficient on NAEP."
"The only interpretation I can make is the students are being lied to -- they're being told they're doing all right at science, and how is the child supposed to know the state is allowing him to pass a test that is not the same as other students are taking around the country?"
Two consultants who are now working with science educators in the Pittsburgh and Philadelphia schools see the same problem in Pennsylvania.
The Pennsylvania state science test "is a very factoid-based test," complained Donna Cleland, director of professional development for the 21st Century Center for Research & Development in Cognition & Science Instruction, which is working to apply brain science research to middle school science teaching in both cities.
On top of that, the science test scores are not used to evaluate how schools are doing under the No Child Left Behind Act -- only reading and math scores count for that assessment.
As a result, said Ms. Cleland and F. Joseph Merlino, principal investigator for the 21st Century Project, when the Pennsylvania tests are on the horizon, teachers start cutting back on science instruction in favor of reading and math.
Another general problem: Many teachers who lead science classes don't have much training in it, and often they haven't taken many science courses in college.
That becomes particularly critical if the content is controversial, as in teaching evolution in biology classes.
Eric Plutzer, a political science professor at Penn State University, has found that 28 percent of high school biology teachers do a good job of teaching evolution and are comfortable with the subject, while 13 percent do not believe in Darwinian evolution and often won't teach it.
That leaves a broad group of nearly 60 percent of teachers who accept the concepts of evolution but are wary of dealing with critics in their communities, Mr. Plutzer said. As a result, they often tell their students they have to teach evolution because it's part of the state standards, "but everyone is free to believe what they want."
While the opposition to evolutionary biology usually has religious roots in America, there is a broader issue that affects adults in this country -- an overall lack of scientific literacy.
Jon Miller, a University of Michigan researcher who has done surveys on this issue for years, said there are two ways to look at the statistics.
On the one hand, scientific literacy has been rising, going from 10 percent in 1988 to 28 percent in 2008, which makes America second only to Sweden.
On the other hand, he said, "we live in a democracy and when only 28 percent of the people can understand climate change, for instance, that's not enough."
It's also not enough to keep American students competitive in the job market.
"People mistakenly believe that getting a diploma will guarantee them a good job for life," Mr. Miller said. "What we know about today's economy is that sometimes a piece of paper will get you an interview, but if you don't know the information, it won't get you very far."
Tomorrow: Hands-on science in the classroom
Mark Roth: firstname.lastname@example.org or 412-263-1130.