Why standardized tests are not an appropriate way to measure growth and some ideas for tests that may be more appropriate
Clearly there needs to be a way to assess the success of teachers. The purpose of the tests and the diversity of the students in any teacher's classroom are just two reasons why standardized tests are not the right assessment tool. This doesn't mean that there are not any available tests that could measure student progress in a meaningful way.
Assessment of teaching and learning is not the mission of standardized tests. A "good" standardized test produces a clear "bell curve," the statistical spread of scores that makes it possible to divide the results in stanine scales (with a mean of 5 and a standard deviation of 2). If too many students answer a question correctly it may be that teaching is interfering with the measurement, so that question is "bad" and will probably be thrown out. This leads to tests involving talking pineapples, races for riding lawn mowers, and other content that is not taught in many schools. Maintaining the standardized bell curve means that something else must be measured through the tests other than what might interfere with the test by actually being taught.
Not every student in a given class is an appropriate candidate for one particular grade level of standardized test. Students who begin the year performing several grade levels below or above the grade level measured with a particular test will not produce meaningful results. Special education teachers can have the experience of making the breakthrough in teaching a child to read and seeing their test scores plunge from scores indicating random guessing to reading with incomplete comprehension and thus getting all the answers wrong. Teachers with classes filled with gifted students may have an entire class that scored between the 95-99 percentile. How can they produce "gains" in standardized testing scores?
It seems clear that trying to measure student growth through standardized tests is doomed to fail. Standardized tests make the assumption that half of our students must be below average, and the same percentage each year are doomed to failure. Standardized tests are just not the right tools to measure educational progress.
There may be other more open ended tests that could perhaps give more meaningful information. There are reading tests, some even available online, that measure a students reading ability and comprehension on an open ended scale. It seems more than possible that similar math assessments must exist. A student that made more than one grade level progress in the course of one year could be acknowledged as having made "gains." For social studies and science much of what should be measured is understanding of the facts in a particular curriculum. This is a case where the adage that "tests drive the curriculum" is not such a bad thing. If an agreement could be reached as to what topics and information must be covered at a particular grade level, then tests could be devised to measure students mastery of those topics. These must not be standardized tests on secret topics. Instead they should be tests that measure whether students have learned a specific topic. It must be acceptable for more than half of the students to succeed.
Thank you for the opportunity to write about this important topic.
Yours truly,
Elizabeth Dejean, MLS
P.S. 360
Bronx, NY
Clearly there needs to be a way to assess the success of teachers. The purpose of the tests and the diversity of the students in any teacher's classroom are just two reasons why standardized tests are not the right assessment tool. This doesn't mean that there are not any available tests that could measure student progress in a meaningful way.
Assessment of teaching and learning is not the mission of standardized tests. A "good" standardized test produces a clear "bell curve," the statistical spread of scores that makes it possible to divide the results in stanine scales (with a mean of 5 and a standard deviation of 2). If too many students answer a question correctly it may be that teaching is interfering with the measurement, so that question is "bad" and will probably be thrown out. This leads to tests involving talking pineapples, races for riding lawn mowers, and other content that is not taught in many schools. Maintaining the standardized bell curve means that something else must be measured through the tests other than what might interfere with the test by actually being taught.
Not every student in a given class is an appropriate candidate for one particular grade level of standardized test. Students who begin the year performing several grade levels below or above the grade level measured with a particular test will not produce meaningful results. Special education teachers can have the experience of making the breakthrough in teaching a child to read and seeing their test scores plunge from scores indicating random guessing to reading with incomplete comprehension and thus getting all the answers wrong. Teachers with classes filled with gifted students may have an entire class that scored between the 95-99 percentile. How can they produce "gains" in standardized testing scores?
It seems clear that trying to measure student growth through standardized tests is doomed to fail. Standardized tests make the assumption that half of our students must be below average, and the same percentage each year are doomed to failure. Standardized tests are just not the right tools to measure educational progress.
There may be other more open ended tests that could perhaps give more meaningful information. There are reading tests, some even available online, that measure a students reading ability and comprehension on an open ended scale. It seems more than possible that similar math assessments must exist. A student that made more than one grade level progress in the course of one year could be acknowledged as having made "gains." For social studies and science much of what should be measured is understanding of the facts in a particular curriculum. This is a case where the adage that "tests drive the curriculum" is not such a bad thing. If an agreement could be reached as to what topics and information must be covered at a particular grade level, then tests could be devised to measure students mastery of those topics. These must not be standardized tests on secret topics. Instead they should be tests that measure whether students have learned a specific topic. It must be acceptable for more than half of the students to succeed.
Thank you for the opportunity to write about this important topic.
Yours truly,
Elizabeth Dejean, MLS
P.S. 360
Bronx, NY