Conditional error variance in the WISC-IV

<p>Measurement error at different ability levels in the WISC-IV was studied to empirically test the conditional error variance hypothesis. Graduate students in clinical psychology at a Midwestern university scored fictitious WISC-IV Vocabulary subtests constructed to yield actual scaled scores of 4, 10, and 16. Classical measurement theory assumes error rate will be constant across the three conditions. Modern test theories (Item Response Theory), however, predict that the precision of a measurement instrument will change as a function of the examinee's ability level. Data supported the conditional error variance hypothesis. Scorers made significantly more errors in the low- and high-abilitylevel conditions than they did in the average ability condition. Implications of these findings for intelligence testing are discussed.</p>

Name thesis:degree_name

Master of Science (MS)

Level thesis:degree_level

Open Access Thesis

Discipline thesis:degree_discipline

Psychology

Year

2006

Author dc:creator

• Erdodi, Laszlo Attila

Contributors dc:contributor

• John Knapp, PhD, Chair
• Renee Lajiness-O’Neill, PhD
• Dean Lauterbach, PhD

• Standard deviations
• Psychometrics
• Educational tests and measurements
• Psychology

Repository record dc:identifier

https://commons.emich.edu/theses/52

OAI identifier oai:identifier

oai:commons.emich.edu:theses-1051