How do you judge the quality of a school, a district, a teacher, a student? By the test scores, of course. Yet for all the talk, what educational tests can and can’t tell you, and how scores can be misunderstood and misused, remains a mystery to most. The complexities of testing are routinely ignored, either because they are unrecognized, or because they may be—well, complicated.\ Inspired by a popular Harvard course for students without an extensive mathematics background, Measuring Up...
How do you judge the quality of a school, a district, a teacher, a student? By the test scores, of course. Yet for all the talk, what educational tests can and can’t tell you, and how scores can be misunderstood and misused, remains a mystery to most. The complexities of testing are routinely ignored, either because they are unrecognized, or because they may be—well, complicated. Inspired by a popular Harvard course for students without an extensive mathematics background, Measuring Up demystifies educational testing—from MCAS to SAT to WAIS, with all the alphabet soup in between. Bringing statistical terms down to earth, Daniel Koretz takes readers through the most fundamental issues that arise in educational testing and shows how they apply to some of the most controversial issues in education today, from high-stakes testing to special education. He walks readers through everyday examples to show what tests do well, what their limits are, how easily tests and scores can be oversold or misunderstood, and how they can be used sensibly to help discover how much kids have learned.
Prologue 1If Only It Were So Simple 5What Is a Test? 16What We Measure: Just How Good Is the Sample? 35The Evolution of American Testing 46What Test Scores Tell Us about American Kids 74What Influences Test Scores, or How Not to Pick a School 113Error and Reliability: How Much We Don't Know What We're Talking About 143Reporting Performance: Standards and Scales 179Validity 215Inflated Test Scores 235Adverse Impact and Bias 260Testing Students with Special Needs 281Sensible Uses of Tests 315Notes 335Acknowledgments 345Index 347