The roughly 200 high school teachers in the Elmbrook, Wis., school district were recently asked to describe the way they grade students -- and no two approaches were the same. Elmbrook is now trying to devise a more consistent grading system.

Revisions to the grading policy are also under consideration in Montgomery County, where public school officials worry that a child could easily receive different grades for the same work in different schools.

In an era in which student assessment has taken top priority across the country, educators are looking anew at how teachers determine grades for tests, projects and papers -- and at the factors that should be considered for semester and year-end marks.

"The whole idea of grading and reporting is to accurately reflect a student's work often enough, to provide feedback between the home and the school and the teacher," said Montgomery County Superintendent Jerry D. Weast. "And it should be as consistent and reliable as possible."

But how consistent and reliable can grading systems ever be?

Issues of race and class sometimes come into play, as do factors as basic as how a teacher feels when grading a pile of humanities essays, said Amy Wilkins, executive director of the Trust for Early Education and an assessment expert. And judging the creativity of an art project or the originality of an English literature essay can never be a process as rigid as determining whether proper nouns are capitalized, she said.

"I think one can be more or less objective, but we are obviously not dealing with chemicals in a test tube or atoms bouncing around," said Amir Hussain, religious studies professor at California State University at Northridge. "We are dealing with people and ideas, and there is always an element of subjectivity in that."

Teachers have used some type of grade -- letters, percentages, numbers -- to sum up student achievement for more than a century. Yet many educators have become critical of the traditional system of letter grades for a variety of reasons.

The critics say grading is an imprecise reflection of student performance because school districts set different standards for achieving a particular grade and because teachers weigh factors other than academic achievement in grading.

"Teachers all did different things," said Jennifer Smith, 21, a student at Northern Virginia Community College who attended public schools in Fairfax and Arlington counties. "Some teachers were more lenient than other teachers. Some would grade a little more on attendance and tests, whereas others included class participation."

In recent years, some districts have altered the way elementary students are graded, often changing the traditional A-B-C-D-F to number systems that measure specific skills.

The Elmbrook district in Wisconsin and others are seeking to change grading in upper levels, too, not simply by using techniques such as having teachers grade papers after masking the students' names, but also by aligning report cards with content standards. These standards spell out what students are supposed to know in specific subjects.

In the San Juan Unified School District near Sacramento, parents of students in grades 1 through 8 receive detailed trimester report cards that show the content that students are supposed to know and provide an assessment in each area.

Take, for example, first-grade Mathematics Standard 3: Measurement and Geometry. These benchmarks are on the report card:

* Sort objects and data by common attributes.

* Identify, describe, classify and compare geometric figures.

* Compare the length, weight or volume of objects.

* Tell time and relate time to events.

Students are given a numerical assessment for each: 4 is advanced, 3 is proficient at the grade-level standard, 2 is approaching the grade-level standard and 1 means below the grade-level standard. On a traditional report card, all of that, including assessments for four other standards, would be put meshed in a single letter grade.

Montgomery officials don't plan to go that far in changing their grading system, but Weast said they are struggling with the proper balance of "tools" to include in an overall grade. For example, one proposal would increase from 25 percent to 30 percent the weight of a final exam on the semester grade for high school computer science, English, foreign languages, mathematics, science and social studies.

The county school board will make final decisions this year after community input, Weast said.

The limited research that has been done on grading shows there is no ideal way to approach it, said Ken O'Connor, a curriculum coordinator with the Toronto District School Board who has conducted some of the most influential research. But several experts say grading systems work best when they assess only those things that are achievement-related.

Researchers are concerned, in particular, about a seeming disconnect between grades earned in the classroom and standardized test results. A few years ago, the South Carolina Department of Education compared math grades given by teachers in the state, as reported by eighth-graders in a nationwide survey, with results of the state's Palmetto Achievement Challenge Test.

Thirty-two percent of the students reported getting mostly A's, but only 7 percent earned top scores on the PACT, which is aligned with the state content standards.

"What teachers do often is factor in a lot of motivational pieces," said John Holton, a South Carolina state education coordinator for math and science. "I think we're still in the learning curve of what it means to be standards-based."

Some educators readily acknowledge that grading is uneven -- and that they like it that way.

"Part of life is that we experience different things," said Rich Crowley, coordinator of guidance services for Fairfax County public schools. "And sometimes that has to do with who our boss ends up being and what they value. It is a microcosm for preparation for life."