You are here

Basic concepts

12 February, 2015 - 11:00
Available under Creative Commons-ShareAlike 4.0 International License. Download for free at http://cnx.org/contents/ce6c5eb6-84d3-4265-9554-84059b75221e@2.1

Standardized tests are created by a team usually test experts from a commercial testing company who consult classroom teachers and university faculty and are administered in standardized ways. Students not only respond to the same questions they also receive the same directions and have the same time limits. Explicit scoring criteria are used. Standardized tests are designed to be taken by many students within a state, province, or nation, and sometimes across nations. Teachers help administer some standardized tests and test manuals are provided that contain explicit details about the administration and scoring. For example, teachers may have to remove all the posters and charts from the classroom walls, read directions out loud to students using a script, and respond to student questions in a specific manner.

Criterion referenced standardized tests measure student performance against a specific standard or criterion. For example, newly hired firefighters in the Commonwealth of Massachusetts in the United States have to meet physical fitness standards by successfully completing a standardized physical fitness test that includes stair climbing, using a ladder, advancing a hose, and simulating a rescue through a doorway (Human Resources Division, nod.). Criterion referenced tests currently used in US schools are often tied to state content standards and provide information about what students can and cannot do. For example, one of the content standards for fourth grade reading in Kentucky is “Students will identify and describe the characteristics of fiction, nonfiction, poetry or plays” (Combined Curriculum Document Reading 4.1, 2006) and so a report on an individual student would indicate if the child can accomplish this skill. The report may state that number or percentage of items that were successfully completed (e.g. 15 out of 20, i.e. 75 per cent) or include descriptions such as basic, proficient, or advanced which are based on decisions made about the percent of mastery necessary to be classified into these categories.

Norm referenced standardized tests report students' performance relative to others. For example, if a student scores on the seventy-second percentile in reading it means she outperforms 72 percent of the students who were included in the test's norm group. A norm group is a representative sample of students who completed the standardized test while it was being developed. For state tests the norm group is drawn from the state whereas for national tests the sample is drawn from the nation. Information about the norm groups is provided in a technical test manual that is not typically supplied to teachers but should be available from the person in charge of testing in the school district.

Reports from criterion and norm referenced tests provide different information. Imagine a nationalized mathematics test designed to basic test skills in second grade. If this test is norm referenced, and Alisha receives a report indicating that she scored in the eighty-fifth percentile this indicates that she scored better than 85 per cent of the students in the norm group who took the test previously. If this test is criterion referenced Alisha's report may state that she mastered 65 per cent of the problems designed for her grade level. The relative percentage reported from the norm-referenced test provides information about Alisha's performance compared to other students whereas the criterion referenced test attempts to describe what Alisha or any student can or cannot do with respect to whatever the test is designed to measure. When planning instruction classroom teachers need to know what students can and cannot do so criterion referenced tests are typically more useful (Popham, 2004). The current standard-based accountability and NCLB rely predominantly on criterion based tests to assess attainment of content-based standards. Consequently the use of standardized norm referenced tests in schools has diminished and is largely limited to diagnosis and placement of children with specific cognitive disabilities or exceptional abilities (Haertel & Herman, 2005).

Some recent standardized tests can incorporate both criterion-referenced and norm referenced elements in to the same test (Linn & Miller, 2005). That is, the test results not only provide information on mastery of a content standard but also the percentage of students who attained that level of mastery.

Standardized tests can be high stakes i.e. performance on the test has important consequences. These consequences can be for students, e.g. passing a high school graduation test is required in order to obtain a diploma or passing PRAXIS II is a prerequisite to gain a teacher license. These consequences can be for schools, e.g. under NCLB an increasing percentage of students in every school must reach proficiency in math and reading each year. Consequences for schools who fail to achieve these gains include reduced funding and restructuring of the school building. Under NCLB, the consequences are designed to be for the schools not individual students (Popham, 2005) and their test results may not accurately reflect what they know because students may not try hard when the tests have low stakes for them (Wise & DeMars, 2005).