Numbers game: Making sense of this year's rankings

Statisticians warn against making too much of small differences, but it's a competitive world. A hundredth of a second made the difference between Olympic gold and silver in the men's 4-by-100-meter relay race in Athens.

So here, with a few words of caution, is the Globe's annual ranking of school districts by their MCAS scores.

The Globe ranked 373 public school districts in Massachusetts on their average rank on all MCAS tests from spring 2004. The ranking includes all school systems: regular, charter, regional, and vocational.

For each of the 10 tests, districts were ranked on the percentage of students reaching proficiency or better. The Globe combined the percentage of advanced and proficient scores into a single percentage for each test to calculate an average rank for all tests, because the state Education Department is pushing to help all schools reach proficiency. (The two lower levels are needs improvement and warning/failing.)

In previous years, the Globe has ranked districts only if they gave all 10 tests. But that left regional districts off the list. For example, the state considers Concord and Carlisle (with elementary and middle-school grades) and Concord-Carlisle (with the high school only) to be separate school districts. This year, towns in regional school districts may appear more than once in the list. Concord-Carlisle is ranked on its high school scores, but so are the separate Concord and Carlisle districts, based on their scores in elementary and middle school.

The rankings include only districts with more than 10 students in a grade, so some small schools are excluded. The only town school districts excluded are Gosnold, Great Barrington, Hancock, Lancaster, Rowe, Rutland, and Savoy. Many charter schools are too small to be ranked.

A few points to keep in mind:

A school that ranks 10th may have only slightly better scores than the 30th school.

Schools draw students from communities that vary widely in wealth and other factors. Research has shown that test scores largely reflect the income and education of parents, so a district with higher scores doesn't necessarily have ''better schools."

The share of students in special education or English as a second language classes varies greatly from district to district.

The Globe has changed its ranking method, partly because the state phased out its average scaled scores, instead reporting percentages at each level.

A school's ranking does not give details of test scores. One school might have many students who barely reach the proficient range, while in another most students could have scored in the advanced range.

Bill Dedman can be reached at dedman@globe.com.