Formative and summative assessments with norm-referenced and criterion-referenced approaches

Formative and Summative Assessments

Criteria and norm-referenced assessment strategies

Assessments are often divided into two major categories: summative and formative. Summative assessments test information in a cumulative fashion. Examples of these types of assessments include standardized state or district tests, or chapter or semester reviews. The weakness or limitation of summative assessments is that they are usually so infrequent, and happen so late in the learning process, they provide little relevant information for students, and provide (arguable) value only in evaluating the success of teachers or schools in supporting student learning. In contrast, formative assessments are given at frequent intervals as "part of the instructional process" itself and are "used to make instructional adjustments and interventions during the learning process" (Garrison & Ehringhaus, 2010: 1). For example, after a series of poor quiz results that indicate the majority of the students in a math class do not understand negative numbers, the teacher might review this concept before moving forward in algebra.

Both of these methods can deploy norm-referenced or criteria-referenced grading. A norm-referenced test means that all students taking the test are 'graded' against one another. Norm-referenced measures (NRM) are "most appropriate when one wishes to make comparisons across large numbers of students" and make "important decisions regarding student placement and advancement" (Assessment purposes, 2010, Assessment measures). Modern IQ tests for children, for example, "relate the mental development of a child to the child's chronological age… the intelligence quotient was equal to 100 times the mental age divided by the chronological age" (De la Jara 2010). In other words, certain tasks are determined to be 'normal' for a child of a certain age, and the child's ability or lack thereof to perform at, below, or above this level determines the child's IQ score. Another good example of a norm-referenced test is that of the GRE (graduate record exam), a standardized test for prospective graduate students which is computer-interactive. Depending on the answers to certain gateway questions, the test will 'adjust' its level of difficulty, providing easier or harder questions, and based upon the subject's number answered correctly and the difficulty of the questions, give the student a final score.

Interestingly, the first IQ tests were criterion-referenced -- in other words, students could simply pass or fail. The first intelligence tests were created solely "to be able to identify children who might need special education due to their retarded mental development" and to exclude those who were not prepared, intellectually, for very rigid French elementary school system (De la Jara 2010). However, criterion-referenced measures (CRM) are today deemed "most appropriate for quickly assessing what concepts and skills students have learned from a segment of instruction. Criterion-referenced assessments measure how well a student performs against an objective or criterion rather than another student. Criterion-referenced classrooms are mastery-oriented, informing all students of the expected standard and teaching them to succeed on related outcome measures. The bell curve in this case is skewed heavily to the right, as all students are expected to succeed" (Assessment purposes, 2010, Assessment measures).

2. An etude comparable to those by C. Rose

3. One movement from a piece comparable in difficulty to the Concerto in a Major, K. 622, by Mozart; Concerto in F Minor, Op. 73, by Weber; sonatas by Bernstein, Saint-Sans, or Hindemith; or any of the 16 Grand Solos by Bonade.

A prospective student is judged relative to the individuals with whom he or she competes against on these exercises: during a very strong year for candidates, some highly talented students might be rejected. During a weaker year, as the school needs students to pay tuition, the most talented candidates will be selected, even though in a previous year they might not make 'the cut.' The candidates are graded against one another during a particular audition session.

Conclusion

In the real world, many standardized tests deploy both criterion and norm-based methods of grading. A formative assessment that is criterion-referenced but is also used in a norm-based manner is the High School Proficiency Assessment for Mathematics (HSPA), administered to all New Jersey students and required to obtain a high school diploma in the state. The test combines both multiple choice and open-ended questions. Students are given numerical questions, intended to determine if students can manipulate numbers and also conceptual, abstract concepts such as rational or irrational numbers. They must also be able to spatially manipulate objects in their heads as well as answer questions about basic geometry. As well as 'plugging and jugging' formulas, according to the NJPEP (NJ Professional Education Port) website, students must understand probability, statistics, and basic algebra, and are tested on these concepts in multiple-choice and in an open-end fashion.

The HSPA is a criterion-referenced test. Students either obtain a passing score or a non-passing score. It includes information that all students must know to be judged worthy of possessing a high school diploma in the state of New Jersey. However, because specific data about student's exact scores and passing rates are recorded, these formative assessments and passage rates are often used to evaluate the effectiveness of a high school. If a school, relative to other institutions, has a very low scores, even very low passing but minimally acceptable scores for students on the HSPA, it may be flagged as 'at risk,' and thus norm-based standards are also applied when assessing the test.