Concurrent validity and classification accuracy of curriculum-based measurement for written expression.

Area of Science:

• Educational Psychology
• Special Education
• Language and Literacy Research

Background:

• Curriculum-based measurement (CBM) is a tool for assessing student progress.
• Written Expression CBM (WE-CBM) metrics are used to evaluate writing skills.
• The impact of different test lengths on WE-CBM validity and accuracy is not fully understood.

Purpose of the Study:

• To critically analyze WE-CBM metrics (Total Words Written, Correct Writing Sequences, Percent Correct Writing Sequences, Correct Minus Incorrect Writing Sequences) using 3- and 10-minute test lengths.
• To examine the criterion validity and classification accuracy of these metrics.
• To determine the optimal test length for WE-CBM screening.

Main Methods:

• Criterion validity was assessed by correlating WE-CBM metrics with a state achievement test's composition subtest.
• Classification accuracy was analyzed using discriminant function analysis and Receiver Operating Characteristic (ROC) curves.
• Fourth-grade students (n=109) participated in the study across 6 schools.

Main Results:

• All WE-CBM metrics showed moderate correlations with the state achievement test, with correlations increasing with longer test lengths (though not statistically significant).
• Correct Writing Sequences (CWS) and Correct Minus Incorrect Writing Sequences (CMIWS) demonstrated the highest accuracy in distinguishing proficient from non-proficient writers.
• Classification accuracy, sensitivity, and specificity improved with longer test durations (10-min vs. 3-min).

Conclusions:

• WE-CBM metrics, particularly CWS and CMIWS, are valuable for screening and identifying at-risk students.
• Longer test lengths (10-min) enhance the accuracy and reliability of WE-CBM for predicting writing proficiency.
• Findings support the use of longer WE-CBM probes for more effective educational decision-making.