UPDATED

This November, Denver taxpayers will be voting on whether to expand a math tutoring program districtwide. But more than that, the vote will determine whether the school district’s efforts to innovate on a larger, faster scale is proven enough to be supported by taxpayers.

Last week, Denver Public Schools successfully pushed onto the voting ballot measures that would increase education spending by more than $500 million. Most of that money—$466 million—is included in a bond proposal that will pay for facility renovation and construction. But a measure that sets aside $49 million in property taxes for education services would include $15.5 million to expand a year-old daily math tutoring program—which produced large academic gains (see below) among some of the city’s lowest-performing students—to the entire district.

The tutoring is conducted daily among 4th, 6th, and 9th graders within Denver’s Summit School Network, 10 schools in a poor area of northeast Denver that were among the lowest-performing in the city. The 75 tutors in the program were recruited through a national effort; most are college students, Teach for America applicants, and retired teachers. Each tutor works part-time and is paid $21,000 per year plus benefits and a chance to earn up to $4,000 in bonuses.

The Summit School Network is run by Denver Public Schools staff members, but supported and supervised by Blueprint Schools Network, a nonprofit group that is an offshoot of Harvard University’s Education Innovation Laboratory, or Ed Labs. The group was founded with a goal of quickening the pace of innovation in low-performing schools by beginning new initiatives, rapidly researching them, and then scaling whatever works.

The approach, based on research from Harvard economist Roland G. Fryer, applies five best practices seen in high-performing charter schools: extended school day and year, strong school leadership, data-driven instruction, a “culture of high expectations,” and the increased math tutoring.

As I wrote in June, in a feature on the Denver turnaround efforts, the idea is to scrap what isn’t working and support what does, but do so at a pace much faster than the typically slow rate of change in education. If the November referendum is passed, one of those tenets (the tutoring) will have moved from a research-backed idea, to a pilot program, to a districtwide program in under two years. And one of the schools, McGlone Elementary, will pilot a literacy tutoring program this year that is based off the math program.

Academic performance data released earlier this month may help persuade voters to favor districtwide implementation.

Below are the percentages of students scoring “proficient or advanced” on the 2012 Transitional Colorado Assessment Program in math (EdNews Colorado has a very useful database). Before the 2012 school year, staff at most schools within the Summit Schools Network were replaced, the school day and school year were extended, and central office administrators were moved to locations within the community.
Among students in the Denver Summit Schools Network:

• 52 percent of 4th-graders, compared with 34 percent in 2011
• 37 percent of 6th-graders, compared with 23 percent in 2011
• 23 percent of 9th-graders, compared with 9 percent in 2011

City-wide:

• 55.7 percent of 4th-graders
• 50.1 percent of 6th-graders
• 24.8 percent of 9th-graders
• 43 percent of all students

There is more to efficacy than state test scores, but it seems clear the math tutoring program had some effect, especially in 4th grade, where participating students have almost closed the gap between themselves and the city’s average students.

When broken down on an individual school level, the numbers are more interesting.

At McGlone Elementary School, 4th-grade students scoring “advanced or proficient” in math jumped from 23 percent in 2011 to 46 percent in 2012. Third graders jumped from 20 percent to 31 percent and 5th graders actually dropped from 30 percent to 29 percent. You can also follow the same cohort of students. In 2011, 20 percent of 3rd graders scored “advanced or proficient” while 46 percent of those same students earned that score as 4th graders this year.

At Green Valley Elementary school, each of grades 3-5 scored at least 16 percent higher in the math TCAP scores this year over last, but 4th graders, with a jump from 48 to 69 percent, saw the largest increase. And the Green Valley cohort that scored 69 percent as 4th graders this year scored just 39 percent as 3rd graders last year.

The results should be taken with at least a small grain of salt. Colorado is among the states with the largest gap between the proficiency scores on its state-specific tests and National Assessment of Educational Progress scores. And because the gains are smaller among grades that didn’t receive math tutoring, there should be questions about how to improve scores among those students, especially if expanding the program to all grades is too expensive.

It also shouldn’t come as a major surprise that intensive, daily tutoring raises scores. The issues are more of scale and cost. Because the entire three-year turnaround effort at the Summit Schools costs nearly $11 million, with tutoring taking up a large chunk, there have been questions about the sustainability of the efforts.

Henry Roman, the president of the district teachers’ union,suggested to me at the beginning of the summer that the turnaround effort couldn’t last without a long-term funding commitment from outside sources. Now there is a measure asking for those funds for a more expensive program.

But in order to achieve performance gains money must be spent, and Matthew Spengler, the executive director of Blueprint Schools, said other initiatives, like reducing class size, for instance, could cost just as much.

“If you look at the potential gains to investment, this may be much more effective,” Spengler said in a phone interview Tuesday. “But we are just at the beginning of this.”

It’s unclear how closely the districtwide tutoring would resemble that of the Summit Schools; I’m told it’s undecided who would provide the tutoring if the bond passes. But under the districtwide proposal, schools that have greater needs for instructional support will receive higher levels of funding and 8th-graders will receive tutoring instead of 9th-graders, according to the district.

None of this suggests that voters will agree to pay what the district calculates as an extra $12 per month for the owner of a home valued at $225,000, because of one year’s results from some math tutoring. But if the goal of Denver’s turnaround effort is to move quickly, this is one way to do it.