Developing a Longitudinal Scale for Language: Linking Across Developmentally Different Versions of the Same Test

Area of Science:

• Psychometrics
• Developmental Psychology
• Educational Measurement

Background:

• Many language tests lack statistical linkage between versions or developmental scaled scores.
• This limitation hinders accurate longitudinal tracking of language development.
• Existing scoring methods can obscure true developmental trajectories.

Purpose of the Study:

• To illustrate the problems associated with unlinked and unequated language test scores.
• To propose and validate a statistical model for deriving a developmental scaled score.
• To enable more accurate longitudinal inferences in language assessment.

Main Methods:

• Utilized an accelerated cohort design with 890 students in Grades 1-5.
• Applied a confirmatory factor model to six subtests of the Test of Language Development.
• Linked subtests to a general language factor and equated measurement characteristics across grades and cohorts.

Main Results:

• The proposed models demonstrated good fit, supporting the validity of measuring a general language factor.
• Evidence suggests longitudinal consistency in measurement across different test versions and student groups.
• Language growth was found to be relatively linear among the participating children.

Conclusions:

• Standard scores on language tests can be problematic for longitudinal analysis.
• The developed model provides a method for obtaining developmental scaled scores for consistent language assessment.
• Researchers and clinicians should prioritize tests offering developmental scaled scores for robust longitudinal studies.