Defending Performance Assessments Without Being Defensive

by Douglas B. Reeves

In the politically charged atmosphere surrounding most educational discussions these days, rarely do emotional convictions give way to facts and logic.

Certainly this is the case when it comes to debates about performance assessments in K-12 education. Both sides have perpetuated myths in this particularly muddled debate. Here are some of the most commonly argued falsehoods.

* Myth 1: Multiple choice tests are inherently objective, while performance assessments are inherently subjective.

Those who cling to this myth begin with the premise there is one true answer to every question. Only an educational assessment specialist or an economist, they joke, could believe that mathematical questions have more than a single answer.

Apparently such people are unacquainted with some of the mathematics worksheets that are common in elementary schools today in which an enterprising second-grade student can respond in a variety of ways to this problem: "Sally has four balls and Mary has seven balls. How many balls do they have all together? Show your work."

Some students will write a number sentence, such as 4 + 7 = 11. Others will draw a picture showing four balls with one girl and seven balls with the other, perhaps circling all 11 balls to demonstrate the sum. If such a simple problem can offer multiple responses, how much more likely is this to occur when students are faced with more complex issues not only in mathematics, but in language arts and social studies?

Guessing Game

* Myth 2: Multiple-choice tests are more rigorous than performance assessments.

Educators who have administered multiple-choice tests know that a certain number of the correct answers are simply the result of random selection. Contrast this random error to a performance assessment in which a student is required to demonstrate mastery of a task, such as an algebraic proof, an elaboration of the causes of the American Civil War, or the dissection of a frog. No opportunity exists for a student to guess in any of these instances. Indeed, virtually any performance assessment is more rigorous than its multiple-choice counterpart.

* Myth 3: Performance assessments yield inconsistent results.

As is true of many criticisms of performance assessments, there is an element of truth in this myth. Indeed, performance assessments can yield inconsistent results, but that is true of every test of any type that has a reliability coefficient of less than 1.00. Multiple-choice tests also yield inconsistent results that can be studied and measured.

Most of the literature surrounding performance assessments indicates that poor levels of consistency, frequently measured by agreement among raters, is a direct function of the amount of training the raters have received. When teachers rely upon a two-hour workshop and a cursory review of instructions, they are likely to provide inconsistent ratings to student work. Extensive training, combined with clear and specific evaluation rules, provide for much higher degrees of consistency.

This requirement for clear evaluation rules not only leads to statistical soundness, but also leads to a greater fairness for the student who must take the assessments. To ask any teacher or any school district to state in simple, clear language exactly what students must do to succeed does not seem to be an excessive requirement.

A Quick Remedy

* Myth 4: Performance-based assessments are inherently virtuous.

The exaggerations by the defenders of performance-based assessments are just as damaging as the myths of those who attack alternative assessments. Successful education reform depends not upon the mere implementation of something labeled performance-based assessments, but rather upon excellent, rigorous, valid, and reliable performance assessments. Unfortunately, many school districts and states have fallen into the trap of taking a traditional multiple-choice test, adding a quick essay or short-answer response section to it, and labeling the whole affair a "performance assessment."

Real performance assessments must be linked clearly to an objective standard. These assessments also must be supported with significant amounts of professional development time and energy, so the results of the assessments are similar, regardless of which teacher is doing the rating. If students are to be held to higher standards, and this surely is the core of most educational reform efforts, then the assessment of those standards must be the best we can devise.

In the continuing controversy over educational policy, we would all do well to separate myth from fact and disagreement from disrespect. Although my bias is clearly in favor of greater use of performance assessments, neither exaggerated claims of effectiveness nor slanderous attacks on its advocates serves the debate well.

Douglas Reeves is author of Making Standards Work: How to Implement Standards-Based Assessments in the Classroom, School, and District.