This document is linked to by the SAT9ISR.PDF files produced when scoring the Stanford Achievement Test, 9th Edition for hearing-impaired children. This file as well as the scoring program were developed by Kevin Cole under the direction of the Gallaudet Research Institute. For more technical details concerning the data files distributed with the test results, see the article Score Summary File Description.

Subtest

A portion of the test which covers a broad subject area, such as Reading Comprehension, Math: Procedures, Science, or Spelling.

Item Cluster

Functional areas within a subtest. There are two types of item clusters: Content clusters group items in a way which indicates how well a student has learned facts, e.g. rote memorization. Process clusters group items in a way which indicates how well a student is able to aquire new information and/or utilize information to construct new information or concepts. Individual items often fall into both a content cluster and a process cluster. The number of items in all content clusters within a subtest should equal the number of items in all process clusters within a subtest, e.g. Reading Comprehension at the Primary 3 level, has 54 items. Within the Reading Comprehension subtest, there are three content clusters with 18 items each, totalling 54 items, and four process clusters, with 14, 24, 8 and 8 items, respectively, also giving a total of 54 items.

# of Items

The total number of items in a given subtest or item cluster.

Items Right

Also known as the Raw Score. This gives the number of correct answers for a subtest, or for an item cluster within a subtest.

Items Wrong

The number of incorrect answers for a subtest, or for an item cluster within a subtest.

Items Blank

The number of items left unanswered for a subtest, or for an item cluster within a subtest.

Percent Right

The number of items answered correctly divided by the number of items in the subtest or item cluster, multiplied by 100.

Scaled Score

A "score" provided by Harcourt, Inc, which is derived from the Raw Score. It gives a method of comparing performance across different levels of the same subtest (e.g., scaled scores can be compared on the Mathematics: Problem Solving subtest from year to year, even if a student takes a different level of the test each year.) Scaled scores are valuable for measuring a student's growth in a particular subject area. Scaled scores can also be averaged to summarize the performance of all students in a class or school, on a particular subtest.

The test is designed to measure performance within a certain range. Raw scores which are at the extremes (either very high or very low) result in scaled scores, and subsequently grade equivalents and percentiles which do not have a high degree of reliability. Therefore, scores falling outside of the measureable range are NOT reported.

Notes:

• Unlike scaled scores, raw scores CANNOT be averaged because different levels of a subtest may contain a different number of items.
• Scaled scores are NOT equivalent across content areas. For example, a 560 scaled score in Mathematics: Problem Solving is NOT equivalent to a 560 scaled score in Reading Comprehension.
• Scaled scores CANNOT be used to create an average scaled score across subtests for an individual student (e.g., averaging 590 in Reading Comprehension, a 670 in Mathematics: Problem Solving, and a 680 in Mathematics: Procedures to arrive at an average scaled score of 647 would be an inappropriate use of the scores).

Grade Equivalent

A "score" provided by Harcourt, Inc (derived from the scaled score) indicating the grade at which an average hearing student would achieve the same scaled score. The score is given as "decimal" value: a grade and the month of the school year separated by a period. For example, if a student has a scaled score of 559 in Reading Comprehension, the grade equivalent is 2.3, meaning that the average hearing second-grader, in the third month of the school year would have also obtained a scaled score of 559.

Deaf/Hard of Hearing Age-Based Percentiles

A number indicating the percentage of deaf/hard-of-hearing test takers of the same age who scored lower on a given subtest.

Notes:

• Age-based percentiles are NOT grade-based percentiles.

For more definitions, see the HEM Glossary of Measurement Terms at <http://www.hemweb.com/library/glossary.htm>.

If you have any questions regarding these files, please contact me at the address below: