Designing a State Accountability Program: Part I

Cross-posted from Education Week

In my last blog, I promised a design for a state accountability system, a design that reflects the considerations described in the earlier blogs in this series.  Here it is.

I would have the state start by deciding on a core high school curriculum, not just mathematics and English, but also the other subjects that all students should take to graduate high school.  My personal picks would be English, mathematics (including at a minimum, mathematical modeling, algebra, statistics, probability and geometry), science (physics, chemistry, biology and environmental science), American history, world history, economics, music, the arts, technology and engineering and physical fitness, but my hypothetical state might make other choices.

I’d create a high school curriculum for each of these subjects—based on the Common Core in the relevant subjects—designed to match the rigor of the curriculum in the nations with the highest performance in international comparative tests, a curriculum designed—where appropriate—to enable students to master the concepts and core ideas in each subject and, as well, to be able to apply what they are learning to real world problems.  I’d make sure that the curriculum was designed to enable students to acquire the non-cognitive as well as the cognitive skills that all students will need to be successful.  That list is now familiar, and it includes, but is not limited to, problem-solving, persistence, creativity, innovative capacity, strong analytical skills, the ability to synthesize, strong communications skills, the ability to contribute effectively in groups and the ability to lead when necessary, and so on.  Then I would have the state map that curriculum down to grade one in a way that describes the content of the progression in the curriculum from the beginning to the end of the sequence, using what we now know about children’s growth and development to map a sequence that tracks what we have learned about how young people learn.

Then I would pick no more than three key points on the trajectory from grade one to the end of high school, and for each of those points, I would develop very high quality assessments designed to capture as much as possible the full range of knowledge, skill and the other qualities we want to see in our students in each of these subjects or their precursors and across them at each of these points in a students’ trajectory through school.  If I were choosing those three points, they would be the end of fourth grade, the end of middle school, and the end of the sophomore year in high school.  The test administered at the end of the sophomore year in high school would be designed to measure what all students were expected to learn in order to graduate from high school, and its standard would be set on the basis of what it would take to be successful in the first year of an open-admissions college or a vocational program designed to meet an industry standard.  That is, the last exams would be set at an empirically-determined college- and work-ready standard.  Students could take these exams as early as the end of their sophomore year or whenever thereafter they wished and could leave high school to go to college as soon as they passed it, or they could stay in high school to prepare for selective colleges if they wished.  All high schools would be tasked with getting all of their students to this college-ready standard.

These census tests would include some multiple-choice questions, but would largely consist of performance items, many of which would require the production of such things as extended essays, working robots, art works and so on, which could not be assessed with multiple-choice methods.  The same would be true of the measures of the kinds of non-cognitive skills and abilities described above.  Some of these things could be captured using computers; others would require human judgment.  These three tests would be expensive and time-consuming to administer.

In the off years, the state would administer tests in English, mathematics and, in middle and high school, science on a sampling basis grade by grade at the end of the 2nd, 3rd, 5th, 6th, and 7th grades.  Vulnerable groups would be oversampled to make sure that small populations of such students in the schools would be accurately measured.

The year-by-year sampling assessments would be designed to capture the performance of schools, but not individual students, and would be designed so that all the assessment could be administered and graded entirely by computers, thus producing assessments that would be much less expensive to administer.  In addition, all elementary schools would be required to administer a state-determined diagnostic test of entering first graders’ math and English literacy skills.

The state would create a public web site on which would be posted the relevant results of the five sampling tests, the 1st grade diagnostic test and the three in-depth tests, for each subject and grade tested in every public school in the state and for each group within the school, but not for each student.  It would also compare the average scores to the average scores for the state and for schools serving similar student bodies.  And it would allow the user of the system to track average growth rates for the students in the school to the growth rates of students in the state as a whole and to growth rates for schools serving similar student bodies.  The system would not be designed to compare the performance of individual teachers to the performance of other teachers within the school or outside the school.

The state would take responsibility for using the data generated by this system to identify schools whose students appeared to be in danger of falling significantly behind the expected progressions through the state curriculum, and schools in which vulnerable groups of children were falling significantly behind.  Schools thus identified would be scheduled for visits by teams of experts trained and assembled for this purpose by the state.  The expert teams would be charged with identifying the problems in the school and with producing recommendations for improving school performance through actions to be carried out by the school faculty, the school district, the community and state assistance teams.

If the recommended actions did not produce the desired effects quickly, the state would require large districts to do one or more of the following:

  1. Ask high-performing school principals to take responsibility for the low-performing school, in addition to running their own school;
  2. Deploy a cadre of high-performing teachers to the low-performing school to act as master teachers in that school;
  3. Deploy a cadre of teachers from the low-performing school to a high-performing school for training and mentoring;
  4. Significantly raise the ratio of teachers to students in the school, salting the faculty with high-performing teachers;
  5. Reassign a high-performing principal to the low-performing school.

This, of course, would not work in sparsely populated rural areas.  For those areas, the state would have to maintain a register of highly competent teachers and school leaders who would be available to serve as mentors to leaders and teachers in rural schools and to be stationed in those schools for various lengths of time, as needed.  And the state would have to be able and willing to enlarge the faculties of rural schools serving hard-to-educate students with fully competent teachers on a long-term basis.  You might be wondering how the state will be able to persuade highly competent school leaders and faculty to take on low-performing schools and to serve in them for extended periods in the way this plan calls for.  For the answer to that question, you will need to read my next blog, which will also describe the last feature of my accountability plan, the feature that is probably the most important.