'Similar' Schools Often Aren't All that Similar
Ohanian Comment: So the state comes up with an algorithm to measure and compare schools. But no one is responsible for verifying that the computer's calculations produce groupings of schools that are actually fair comparisons for each other. What will happen with the proposed federal algorithm for mental health screening?
The corporate-politicos will shrug and say, "The computer did it."
According to the state of California, James Lick High School and the Preuss School are similar.
Most students at both schools are children of working-class Latino parents who did not graduate from four-year colleges. So, the state says, the two schools -- and 98 others around California with similar demographics -- should be measured against each other.
But Lick and Preuss could hardly be more different -- in what they offer academically, and how their students perform.
The stark contrast between the two schools illustrates the problem with the state's computer-generated groupings of 100 ``similar'' schools that result in rankings of 1-10 released Tuesday. School officials brag about their rank, families use them to decide where to buy a house and they can even factor into whether voters are willing to tax themselves to fund teachers' salaries and school improvements.
So it raises the question: Is the state being fair when it compares Lick to Preuss? And is there a better way?
Lick, in San Jose's East Side Union High School District, is looking better these days, though some windows are broken and skateboarders have scuffed paint on the concrete planters. Many students come from immigrant families and are striving to master grade-level skills. The school's Academic Performance Index score is 572 out of 1,000 -- up 49 points over last year, but still earning a low state ranking.
Preuss is on the University of California-San Diego campus in upscale La Jolla, housed in a pristine $14 million facility. Students in grades six through 12 enroll only after admissions officials invite them to attend. Its API score is 845, down three from the year before -- but high enough for the highest state ranking.
The ``similarities'' between those two schools, and others, seem to baffle even the experts.
``We hear from school district leadership about this all the time. Many of them are puzzled,'' said Steve Rees, president of School Wise Press, a San Francisco firm that helps school districts issue accountability reports. ``They'll point to schools that they know, and they don't know how the algorithm could have resulted in that match.''
The state computes the rankings using a complicated formula that weighs 14 characteristics, including parental education level, level of turnover in the school population, ethnicity and teacher credentials. Funding levels, and the number of special programs, are not factored in.
Not all of the similar schools pairings are head-scratchers -- many make intuitive sense. On the Peninsula, for instance, Palo Alto and nearby Gunn high schools are considered similar to each other. In San Jose, Quimby Oak Intermediate is similar to John Muir Middle School.
But no one is responsible for verifying that the computer's calculations produce groupings of schools that are fair comparisons for each other, said Bill Padia, director of the policy and evaluation division at the state Department of Education.
``Every school that gets that report ought to be looking at that list of 100 similar schools to make sure it makes human sense,'' he said. ``Similarity is in the eye of the beholder.''
At the Preuss School, every student takes college prep classes, and parents commit to 15 volunteer hours per year. One in three take Advanced Placement tests, and 91 percent of last year's seniors went on to four-year colleges.
At Lick, one in 10 takes an AP test, and a fifth of the student body drops out. The district doesn't track how many go on to four-year colleges.
``I don't see how we could be similar to a school like that,'' said Joel Herrera, one of James Lick's co-principals. ``It doesn't seem like a fair comparison.''
Other strange pairings abound. Andrew Hill High School last year was not considered similar to any other school in Santa Clara County. But San Jose's Gunderson High School was similar to Middle College High School in San Bernardino, which admits motivated sophomores with high test scores and low grades and has them take college courses in the mornings.
Dale Russell, director of standards and assessment for Santa Clara County, is not a fan of similar schools rankings. ``There's been a lot of discussion about how the match is done, and what constitutes similar,'' he said. ``Anybody that gets a lower ranking claims that their school's really not similar to the others, and the ones that get a high ranking, are fine.''
While the rankings don't influence the money schools get from the state, they can influence the support schools get from the community.
Evergreen resident Paul Robichaux, who wants Silver Creek High School to split from East Side Union, cites its low similar-schools rank as one reason. In the East Bay, an opponent of a San Ramon parcel tax two years ago said taxpayers deserved better performance before giving the schools more money; though every school in the district scored a perfect 10 in the statewide rankings, their similar schools ranks were not as good. The parcel tax failed that November.
The rankings were created to satisfy a mandate in the state's Public School Accountability Act of 1999 that schools be compared with others with similar characteristics, said Edward Haertel, professor of education at Stanford University. Haertel led the committee of experts that crafted the formula used to compute the ranking, and though he thinks the results have some value, he also has misgivings.
``If I had my druthers,'' Haertel said, ``we wouldn't have a similar schools ranking.''
Contact Jon Fortt at firstname.lastname@example.org or (408) 278-3489.
San Jose Mercury News