"It has wings," a boy said excitedly, looking through a pile of creek muck in Mary Carter's fourth-grade class at Glacier View Elementary School.
"It's probably a centipede," Principal Bernie Sorenson suggested, taking a look. A girl eagerly showed off her slug, and another offered up a egg from who knows what kind of creature.
In Glacier Valley, that's a reading and math lesson as much as a science class.
For the past four years, the school has made a determined effort to improve its reading, writing and math skills and test scores by incorporating those subjects in all the classes.
The school counselor tutors reading and took classes in teaching writing so he could help students with journals. The phys ed teacher has posted sports-related words on the gym wall.
And in Carter's class a project on fungus, bacteria and invertebrates from Jordan Creek will ask students how biological producers, consumers and decomposers are interdependent.
"When you think about it, it makes for a great conversation," Sorenson said. "There are no right or wrong answers. It's what did you observe."
As students pored over the muck, they were expected to draw the habitat, list their observations, and record and graph the temperatures on the inside and outside of the containers.
"So the kids are given graph paper and they're going to have to figure out how to make a graph," Carter said of the project's math component.
Since the school began its own reform movement four years ago, the fourth-grade scores on the California Achievement Test generally have risen.
The language arts score, for example, has gone up from an average of the 38th percentile in 1998, compared to a national sample of students, to the 60th percentile in 2000. The math score jumped from the 63rd percentile in 1999 to the 77th in 2000.
A 77 means the Glacier Valley average score was equal to or higher than 77 percent of the students in the national sample.
Sorenson said the results of the third-grade state benchmark tests in reading, writing and math will be more useful in diagnosing students than the CATs. The CATs are too broad, and they're not very authentic, she said. The benchmark tests, in contrast, ask for writing samples and mathematical problem-solving.
The recent results from the benchmark test, taken last March for the first time, showed that 82 percent of the Glacier Valley third-graders were judged advanced or proficient in reading. Forty-three percent had achieved those levels in writing, and 79 percent had in math. So writing will be a priority, Sorenson said in a firm voice.
Glacier Valley's efforts have come because its teachers were willing to look at test results, look for patterns, and look at themselves.
Other schools in the district will be doing that sort of thing, following the mailing to parents Wednesday of students' individual test results from last school year's benchmark tests in the third, sixth and eighth grades.
Students are judged as advanced, proficient, below proficient and not proficient. (See chart Page C4.)
This fall, parents and students got the results of the first high school graduation exam, which sophomores took last school year. They either passed or failed each section of the test in reading, writing and math. Some students took sections of the test again in October and are awaiting those results.
The test results tell schools and parents how the students performed in smaller categories of skills. Sixth-grade reading, for example, includes standards such as inferring meaning and defining story elements.
But it will take some digging for parents to know the standards in detail. The state has published a performance standards guide, which is also on the Education Department's Web site, http://www.eed.state.ak.us/. Look under the standards section.
Juneau School District administrators say the state tests, added to an upcoming districtwide computer database of information about students' grades and test results, will help teachers diagnose individual student strengths and weaknesses. Principals and school district administrators will know what schoolwide or districtwide improvements are needed.
"It will make it very easy for schools to hone in," said Charla Wright, the school district's director of writing and reading assessments.
The database will let teachers know how a child has done on a variety of tests over time. It also lets administrators readily break down test results by characteristics such as sex, ethnicity, family income, English proficiency and participation in special education.
"Before in the district we haven't been able to manipulate our data," Assistant Superintendent Peggy Cowan said.
Interpreting test results is an uncertain science. Test scores do not give causes, Jim Cox, an educational consultant from Anaheim, Calif., told state educators this fall. It takes frank self-reflection by teachers and administrators to find the causes.
The problem is that test scores aren't determined solely by the classroom program. That's why principals don't like it when the public compares schools by test results.
Scores also are related to family income, English proficiency, transiency in the population, the tests' physical environment, teachers' and students' attitudes toward testing, test-taking skills, and how well the curriculum aligns with the test content, Cox said.
In analyzing test results, it's very important for schools to look for patterns over several years, both for individual students and schools, said Carol Barnhardt, an education professor at the University of Alaska Fairbanks. Test scores should be used in the context of what teachers and parents know about the students, she said.
"In order for there to be any meaningful change, the scores have to be reviewed with significant input from the people who know the students best," Barnhardt said.
The Juneau School District wants each of its schools to have 95 percent of their students meeting the standards at their grade levels by 2005.
The schools, in their accountability plans, called for a wide variety of student assessments, staff development and curriculum development focused on the standards, and interventions for students who don't meet them.
"Not all tests tell you all that it is you want to know," said Harborview Elementary Principal Bob Dye. That's why it's important to look at multiple assessments, he said.
Diana Gifford, principal at Mendenhall River Community School, will ask teachers to meet in multi-grade groups to look at the third graders' test results, because they are the entire school's responsibility.
The statewide benchmark test results have been broken down by sex, English proficiency and family income. Eventually, school districts and individual schools will do the same and add other categories.
Girls statewide did notably better than boys in reading and writing. Most students with limited English did poorly. A higher percentage of students from prosperous families were proficient or advanced than were students from poor families.
Dye has analyzed the results of the 60 third-graders at Harborview who took the benchmark test. In the reading test, for example, all six of the not-proficient students and 10 of the 12 below-proficient students had learning difficulties the school already had identified. Those include learning disabilities, severe emotional or behavioral problems, or a lack of proficiency in academic English.
"When you look at our results and break it down like that, I feel pretty good about our results in reading," Dye said.
The writing test, on the other hand, showed that only eight of the 21 students who were judged below proficient had learning difficulties.
Juneau schools already have programs in place for students with special needs. The pressure to improve those students' test scores could lead to calls for more resources.
"We have a pretty high percentage of English as a second language students, yet we only have a half-time ESL teacher," Dye said. The school added a quarter of a teacher's time, as well. "Test data will show there's more need for resources for these kids than we currently have."
Glacier Valley, for example, benefited from winning a competitive $266,000 foundation grant three years ago.
It paid for staff time to talk about teaching practices, to create integrated lessons, and to take classes in how to teach reading, writing and math. The school also has bought "piles and piles and piles" of books, Sorenson said.
The state Department of Education has hired 40 consultants to help school districts train teachers how to teach to the standards. The agency is giving school districts about $9.5 million in grants to beef up school programs, train teachers and provide students extra help.
Juneau received $387,500, and has been putting it into reduced class sizes in the lower grades and more math and English sections in the high school, Cowen said.
"We're being asked as a public system to do something we've never done before educate all kids to a high standard," said state Education Deputy Commissioner Bruce Johnson. "And that's going to take more resources."
There are 60 to 90 special education students per teacher, Juneau-Douglas High School English teacher Bill Chalmers said. If that ratio were cut down to 10 or 15, "every special ed kid I know could compete for the college acceptances.
"But we don't have that," he said. "I doubt that we're really willing to pay for that."