Summer’s gone, almost. It’s time to get serious about fixing our schools.

I had fun in July and August jousting with readers about provocative topics such as teachers making unannounced home visits and getting D.C. schools leaders to take test tampering seriously. But many of the comments expressed a yearning for more. What can we do to make our schools work better without putting too much strain on staffs and students?

Let’s find out. Give me your formula for a good school. If you have evidence from schools that already exist, so much the better. Hopefully your ideas will work in schools with many low-income students. E-mail them to me at or post them as comments on my blog at I will publish the most interesting ones. How should we handle instruction, curricula, discipline, parent involvement, hiring and firing? How much power should the principal have? What assessments should be used?

Here are two suggestions from me. They are very different but appear to be working well with low-income public school children.

Model A: New York Performance Standards Consortium. These 28 small public high schools have student populations that, on average, are 64 percent low-income, 44 percent Hispanic, 26 percent black, 20 percent white and 10 percent Asian. Their most unique feature is a waiver from four of the five New York state tests (they do take the English exam). Each student is judged instead by a series of tasks: an analytic literary essay, a social studies research paper, an original science experiment and some application of higher mathematics. Assessments are done by external evaluators.

Most consortium schools are open to all New York City students. Some focus on specific populations, such as recent immigrants and students who have dropped out. One research study found that approximately 90 percent of consortium graduates attended college, with 77 percent of those attending four-year colleges and 19 percent attending two-year schools. A 2011 report by the National Student Clearinghouse found that of those who entered four-year colleges within a year of high school graduation, 92 percent enrolled for a second year, compared with a national average of 75 percent, and 74 percent enrolled for a second year in a two-year college, compared with a national average of 54 percent. Consortium students had lower dropout rates, higher college-going rates and higher daily attendance than the average for regular New York City high schools.

Model B: KIPP schools. All but one of the 109 schools in 20 states and the District that make up this network are charters. They include middle, elementary and high schools, all small like the consortium schools. Eighty-five percent of the students are low-income; 60 percent are black, 35 percent Hispanic, 2 percent white and 2 percent Asian. KIPP schools are open to all students who apply, with admission by random lottery if there are more applicants than slots. The schools are distinguished by longer school days and years and principals’ power to hire and fire staff at will. The students take annual state tests.

The first 1,000 KIPP middle schools to complete the four-year middle school program rose on average from the 32nd to the 60th national percentile in reading and from the 44th to the 82nd percentile in math. By eighth grade, more than 90 percent of KIPP students outperform their local districts on state tests. Students completing KIPP kindergarten outperform a national sample of similar students by 63 percent in reading and 47 percent in math. Eighty-nine percent of KIPP students attend college. Based on the first two years of KIPP middle school graduates, 33 percent graduate from college six years after graduating from high school — slightly higher than the national average and three times the average for similar students.

Tell me as much as you like about models you favor. Let’s see if we can distill the essence of a good school so everyone can have one.