Knowing What Students Know: The Science and Design of Educational Assessment

Rating:
Summary: An argument for fudging the data
Review: The thesis of this book is that (1) tests of educational assessment are increasingly being looked at by politicians (2) this has implications for the employment of educators and (3) being very clever, we teachers should be able to come up with new wasy of measuring student performance that won't actually test what the students have learned.

Okay, perhaps that's a bit overly cynical. But not by much.

Politicians are indeed looking closely at student performance, and for the first time in a long time, thinking about grading how well the public school are performing. And to be fair to the teachers, the intent often has as much to do with individual political aims as it does with education. So teachers repsond in kind, ganing the system, teaching to the exams and so forth. The only parties being left out of this game are the students themselves.

Assessment is needed, both to judge how well students are doing, and to judge how well the schools themselves are doing. The finny thing is that there is a measure that's been ignored in all of this. It's called grading.

Countless studies have been done over the past 50 years to determine what the best indicator of college performance is. People have looked at SAT scores, socioeconomic status, personality and dozens of other measure, and the one measure that consistently explains most of the varience is undergraduate grades. That's it. Even given the grade inflation of the past few decades, grades are still a pretty reasonable indicator.

Rating:
Summary: decent
Review: This book is not terrible and provides a good overview and introduction to assessment. The previous reviewer of this book could not have been more wrong about grades. Actually, research shows grades are a horrible predictor of success when controlling for other factors (income, socio-economic status, etc.). Research also shows there is *no* link between a grade and what someone has learned. This is because grades are relative and measure an individual's performance at one time, although I will admit that using multiple methods and more numerous testing will increase the validity of grades. If tied to the learning goals of a course, grades can be effective. But as a policy making tool or measure of what someone has learned, grades really don't say much, if anything at all.