American schoolchildren read with surface understanding but have great difficulty when asked to think more deeply or to elaborate on what they have read, a national study reported yesterday.

In an extensive assessment of reading ability, 80 percent of third graders, more than half of seventh graders and 36 percent of 11th graders performed inadequately or minimally when asked to read and respond critically.

"Students had great difficulty expressing even one substantive thought," the report stated.

The study, sponsored by the Education Department, urged that reading instruction at all levels be restructured to help students more effectively evaluate what they have read.

The National Assessment of Educational Progress, in measuring the reading performance of 36,000 public and private school students in the three grades, also concluded that teachers inadvertently keep poor readers from improving by avoiding methods they use with better students.

The study also showed significantly lower reading performance among minority and disadvantaged students. Blacks, Hispanics and poor students in grade 11 scored only slightly higher than the national average for seventh graders. The study is part of a regular assessment of academic achievement.

While it does not indicate whether youngsters are better readers than students in the past, it does compare the performance of subgroups. It found that females perform better than males, that students from advantaged urban communities read better than those in disadvantaged urban communities and that youngsters from the Northeast and Central states perform better than those in the Southeast and West.

The study confirmed earlier findings that students do not learn to analyze what they read or communicate their ideas effectively. Fewer than one percent of third graders, 8 percent of seventh graders and 23 percent of 11th graders performed at the highest level in this exercise.

The report focused heavily on teaching practices as reported by students, concluding that "better readers are taught more effective reading strategies, while poorer readers are likely to have their old, less-effective strategies reinforced."

Teachers are more likely to use basic "decoding" exercises with poor readers and spend less time working on comprehension and critical thinking, the report said.

"When we stand in front of a group of poor readers, we automatically think down," said Archie E. LaPointe, executive director of the assessment, conducted by the Educational Testing Service. "We seldom ask them to think about a reading selection or give their opinions or support those opinions. Maybe if we did that more often, they'd become better readers."

The study supported his contention, finding that schools fail to close the gap between good and poor readers as they progress in school. "Our suggestions are geared to teachers, to treat the poor readers the way they treat the good readers, and to have the same expectations," LaPointe said.

Alan Farstrup, director of research and development at the International Reading Association, expressed frustration that the study's conclusions were based only on student reports of teaching methods.

"We're not entirely comfortable with the conclusions being drawn," he said. "It may be that teachers have made the judgment that children having trouble reading need more decoding emphasis." He added, however, that the "data argues strongly for more even treatment" of students, regardless of reading ability.

Poor readers showed little development in reading strategy. At each grade level, they relied on sounding out words and asking for help. But good readers developed more sophisticated strategies in higher grades, relying more on context and rereading to decipher a difficult passage.