Do you remember taking SATs, ACTs, or other college entrance examinations? Did you ever wonder how those tests were constructed, what the scores meant, and the degree to which your scores represented what you really knew or could really do? The SATs and other college entrance examinations are examples of standardized tests. Unlike the teacher-made tests discussed in Chapter 13, a standardized test is typically given under the same, "standardized" conditions to thousands of students who are similar to those for whom the test is designed. This allows the test publisher to establish norms to which any individual score can be compared.

Traditional standardized tests have been subjected to a great deal of criticism and controversy, and today a wide variety of assessments are used. However, standardized tests of many kinds continue to be used for a wide range of purposes at all levels of education.

• Standardized tests are typically carefully constructed to provide accurate information about students' levels of performance.
• Standardized tests are often used to select students for entry or placement in specific programs.
• Standardized tests are often used to diagnose individual students' learning problems or strengths.
• Perhaps the most common use of standardized testing is to evaluate students' progress and teachers' and schools' effectiveness. For example, today, this process is used to fulfill the NO CHILD LEFT BEHIND ACT to improve overall schooling, a trend which began in the 1970's.

The accountability movement stems in part from the public's loss of confidence in education. Legislators (among others), upset by examples of students graduating from high school unable to read or compute, have demanded that schools establish higher standards and that students achieve them. The accountability movement has its critics, however. Many argue that minimum competency testing focuses schools on minimums rather than maximums. Others are concerned that schools will teach only what is tested, emphasizing reading and mathematics at the expense of, for instance, science and social studies, and emphasizing easily measured objectives (such as punctuation) over more important hard-to-measure objectives (such as composition).

Three kinds of standardized tests are commonly used in school settings: aptitude tests, norm-referenced achievement tests, and criterion-referenced achievement tests. Several kinds of scores describe test results according to their position on the normal curve. A normal curve describes a distribution of scores in which most fall near the mean, or average, with a symmetrically smaller number of scores appearing the further we go above or below the mean. One important concept related to normal distributions is the standard deviation, a measure of the dispersion of scores. The standard deviation is, roughly speaking, the average amount that scores differ from the mean.

We use test scores to make inferences about the students we are measuring. The validity of a test is the extent to which those inferences are justified. The types of evidence that are used to evaluate the validity of a test vary according to the test's purpose. For example, if a test is being selected to help teachers and administrators determine which students are likely to have some difficulty with one or more aspects of instruction, primary interest will be in how well the test predicts future academic performance.

Whereas validity relates to the skills and knowledge measured by a test, the reliability of a test relates to the accuracy with which these skills and knowledge are measured. Test scores are supposed to result from the knowledge and skill of the students being measured. But when a test is administered, aspects related to both the test itself and the circumstances surrounding its administration could cause the results to be inaccurate. Reliability is commonly measured using a coefficient that has a theoretical range from 0 to 1. The higher the number, the more reliable the test. In general, good standardized achievement tests should have coefficients in the .90 range and higher.

Some major criticisms of traditional standardized tests relate to issues of validity and reliability. Critics argue that such tests

1. Give false information about the status of learning in the nation's schools
2. Are unfair to (or biased against) some kinds of students (e.g., students from diverse backgrounds, those with limited proficiency in English, or females, and students from low-income families).
3. Tend to corrupt the processes of teaching and learning, often reducing teaching to mere preparation for testing.
4. Focus time, energy, and attention on the simpler skills that are easily tested and away from higher-order thinking skills and creative endeavors.

One major issue in the interpretation of standardized test scores is the possibility of bias against students from low-income or diverse backgrounds. In one sense, this is a question of test validity: A test that gave an unfair advantage to one or another category of student could not be considered valid. Of greatest concern is the possibility that tests could be biased because their items assess knowledge or skills that are common to one group or culture but not another.

The use of computers to administer tests is becoming more common. In its simplest form, the same multiple-choice items (in the same order) are administered to students as they would take them if they sat for the typical, paper-and-pencil test. However, the use of a computer makes it possible to tailor the selection of items to the performance of the student.

The general U.S. Department of Education site at http://www.ed.gov provides links to the Educational Resources in Education (ERIC) system, which contains a searchable database of educational research and other resources, and to the National Assessment of Educational Progress (NAEP), which documents the educational attainment of individual states and the nation as a whole. Particularly interesting are the ERIC Digests, which are brief overviews of important educational topics, and the NAEP content standards, which represent the best thinking of national panels of experts about what each student should have learned in each educational discipline by the fourth, eighth, and twelfth grades. The Buros Institute of Mental Measurements at http://www.unl.edu/buros publishes in book form a series of test reviews by knowledgeable assessment professionals. The reviews are thorough, unbiased, and concise; they should be anyone's first stop for a quick evaluation of a published test instrument.

Today, accountability under the No Child Left Behind (NCLB) legislation has been established to ensure that all children move to academic success on state standards by the year 2014. The assessment provisions are as follows:

• Annual testing - grades 3-8 plus one high school grade, first in math and reading, with science added in 2006-2007, etc.
• Disaggregated reporting of scores - each subgroup scores gathered with a minimum number in that category, varies from state to state (i.e. special education, ethnic groups, males/females, etc.
• Adequate Yearly Progress (AYP) - defined differently in each state, but meant for all subgroups to be scoring at a higher level or with an increasing percentage of students in all subgroups who score at the proficient level (cautions in place for missing students)
• Consequences for not meeting AYP - Schools not meeting AYP for two years or must offer their children supplemental educational services and offer parents opportunities to transfer their students to more successful schools, with more sever consequences to follow if improvements still are not evidenced.

NCLB has been criticized especially for its accountability provisions because most indicate that:

• The testing is narrow and excessive
• There are unequal standards caused by the variations from state to state
• It is underfunded

Most questions regarding NCLB surround the main question of whether or not it will lead to improvements in overall schooling. Time will provide the answer to this ongoing question. In the meantime, NCLB is seemingly supported by both Republicans and Democrats as evidenced in the 2004 election processes.