The latest Nation's Report Card shows 9- and 13-year-olds on average are doing significantly better in reading and math than their counterparts of about 40 years ago, but the achievement of 17-year-olds on average is flat.
The findings also show a narrowing of some racial achievement gaps.
The report on long-term trends for the National Assessment of Educational Progress -- known as NAEP or the Nation's Report Card -- was released Thursday. It compared results on the 2012 long-term trend exams given in public and nonpublic schools with those of students who were tested in the early 1970s.
"When you think about it, these results very clearly put to rest any notion that our schools are getting worse," Kati Haycock, president of the Education Trust, an organization that promotes academic achievement, said in the webinar announcing the results.
"If we have a crisis in American education," she said, it is that it isn't moving fast enough to educate minorities.
"Students of color [on average] are at best performing at the level of white students a generation ago," she said.
She noted that some of the biggest gains took place years ago and a "worrisome slowdown" has occurred in recent years.
Alan Lesgold, dean of the School of Education at the University of Pittsburgh, offered two theories -- which would need further testing -- on why 17-year-olds didn't do better.
Using standardized tests less complex than NAEP to drive instruction, he said, may result in an "underteaching of core ideas that are needed to drive later advancement in school.
"As a result, I suggest, this kind of teaching to the test produces more score improvement in lower grades but without that improved performance being sufficient to drive further progress."
Another possibility, he said, would be that lack of engagement in schools grows with age.
"I think by the time kids are 17, they are sick of taking those standardized tests," said Rae Ann Hirsh, director of the undergraduate early childhood education program at Carlow University.
Ms. Hirsh said the tests cover only a "very tiny piece of a much larger puzzle" and do not address much of what is needed in education today, such as creativity and critical thinking.
"We're honing in on very specific skills, but we're really forgetting the range of skills that children are going to need to compete in a global society," she said.
The NAEP report also sheds some light on students' academic choices.
Of 13-year-olds, 34 percent reported taking algebra, about twice as many as the 16 percent in 1986.
Of 17-year-olds, 23 percent reported taking calculus, compared to 6 percent in 1978.
While reading for fun typically led to higher reading results, the percentage of 9-year-olds who said they read for fun almost every day -- 53 percent -- was the same as when the question was first asked in 1984. For 13- and 17-year-old students, the percentages declined, from 35 percent to 27 percent for 13-year-olds and from 31 percent to 19 percent for 17-year-olds.
The long-term NAEP tests in reading and math were given in the 2011-12 school year to more than 26,000 students in public and nonpublic schools.
While there weren't enough private school students to separate their results, the report showed that Catholic school students had higher average scores than public school students at all three age levels in both math and reading.
The biggest gaps were at age 17, 23 points higher in reading and 20 points higher in math on a 500-point scale.
The long-term trend tests typically are given to the three age groups every four years and have only national results. Reading was first tested in 1971, followed by the first math test in 1973.
The long-term trend tests are different from the main NAEP tests, which are given in various grade levels and subjects and have both national and state results.
Over the years, demographics have changed, with a larger proportion of Hispanic students and smaller proportion of white students.
For 13-year-olds, for example, the shift was from 80 percent to 56 percent white; from 13 percent to 15 percent black; from 6 percent to 21 percent Hispanic; and from 1 percent to 6 percent Asian/Pacific islander.
In a news conference, Peggy Carr, associate commissioner in the assessment division of the National Center for Education Statistics, noted the group of 17-year-olds is different in other ways as well because more students are staying in school instead of dropping out.
In addition, in 2004, NAEP began allowing more special education students with accommodations as well as English language learners to take the test.
At all three age levels, students tended to be in lower grade levels in 2010 than decades ago. For example, most 13-year-olds tested in math were in eighth grade -- both now and years ago -- but a greater portion of them were found in seventh grade or below, 39 percent in 2012 compared to 28 percent in 1978.
On the 500-point scale, average national reading scores rose from 208 in 1971 to 221 in 2012 for age 9, and from 255 to 263 for age 13, both of which were statistically significant.
The change in reading for 17-year-olds was from 285 in 1971 to 287 in 2012, which was not statistically significant.
The change in average math scores for 9-year-olds was 219 to 244; for 13-year-olds, it was 266 to 285; and for 17-year-olds, it was 304 to 306.
However, the lowest-performing 17-year-olds made gains of 7 points in reading and 12 points in math over the decades.
The findings show some increases within some racial groups and the narrowing of some racial achievement gaps.
Black and Hispanic students made larger gains than white students in reading and math at all three ages over the decades.
One of the biggest gains for black students was for 9-year-olds, who scored on average 36 points higher than in the early 1970s in both reading and math. Their current scale scores are 206 in reading and 226 in math.
The report can be found at www.nationsreportcard.gov.
Education writer Eleanor Chute: email@example.com or 412-263-1955.