Examination of an Automated Procedure for Calculating Morphological Complexity

Area of Science:

• Linguistics
• Educational Psychology
• Computational Linguistics

Background:

• Assessing written language complexity is crucial for understanding student development.
• Manual analysis of morphological complexity (MC) is time-consuming and labor-intensive.
• Automated tools offer a potential solution for efficient and objective analysis.

Purpose of the Study:

• To validate an automated scoring procedure for calculating morphological complexity (MC) from written transcripts.
• To assess the utility of an open-access tool, Morpholex, for measuring MC.
• To compare automated MC measures with traditional hand-coding and teacher ratings of writing quality.

Main Methods:

• 146 written responses from fifth-grade students were analyzed.
• Morphological complexity (MC) was assessed using both hand-coding by trained scorers and the automated Morpholex tool.
• Correlational analyses examined the relationship between hand-coded and automated MC measures, and their predictive validity for writing quality.

Main Results:

• Automated MC measures showed a strong correlation (r = .63) with hand-coded derivational morpheme counts.
• Both derivational and inflectional morphemes significantly predicted teachers' overall ratings of writing quality.
• The Morpholex tool demonstrated strong agreement with manual assessment methods.

Conclusions:

• Automated scoring of MC is a valid and potentially valuable alternative to manual analysis.
• This automated approach can aid in monitoring student writing growth and assessing factors influencing perceived academic writing quality.
• The Morpholex tool offers a practical solution for researchers and educators evaluating written language complexity.