TEWL-3

Basic Writing

This subtest consists of 70 items that are scored 0, 1, or 2. It measures a child’s understanding about language and his or her ability to use the writing tools of language. Can be administered independently or in conjunction with the Contextual Writing subtest.

Contextual Writing

This subtest consists of 20 items that are scored 0 to 3. It measures a child’s ability to construct a story when provided with a picture prompt. The subtest measures story format, cohesion, thematic maturity, ideation, and story structure. Can also be administered independently or in conjunction with the Basic Writing subtest.

Overall Writing

Combining the index scores from the Basic Writing and Contextual Writing subtests. It is a measure of the child’s overall writing ability; students who score high on this quotient demonstrate strengths in composition, syntax, mechanics, fluency, cohesion, and the test structure of written language.

Reliability

• For age and grade, internal consistency reliability (alpha) coefficients of all scores meet or exceed .90 for all ages, with most meeting or exceeding .95.
• For alternate form reliability, only 4 of 42 coefficients do not meet or exceed .90. Of the four, all are in the mid to high 80s.
• Stability reliability of all scores approximates or exceeds .90.
• Across all forms of reliability, the reliability of the composite index is in the mid to high 90s.

Validity

• Extensive content validity data are presented.
• Criterion-prediction validity is evidenced by correlations with, and comparison of means and standard deviations of, the Woodcock-Johnson III, the Wechsler Individual Achievement Test–Second Edition, and the 6 + 1 Writing Trait Rubric.
• Using the TEWL–3 to predict scores on the Woodcock-Johnson III and the Wechsler Individual Achievement Test–Second Edition resulted in sensitivity indexes of .86 and .91 respectively, specificity indexes of .89 and .88, and ROC/AUC of .96 and .96. Classification accuracy was .88 for both tests.
• Construct-identification validity is presented based on correlations with age, grade, cognitive ability, achievement, and group discrimination, and on the correlation of individual test items with total test scores.

The TEWL-2 norms represent more than 1,400 children from 33 states. Characteristics of the normative group correspond to those for the 1990 U.S. Census data relative to gender, geographic region, ethnicity, race, and urban/ rural residence.