Automated method of content analysis: a device for psychotherapy process research.

Area of Science:

• Psychotherapy research
• Computational linguistics
• Natural language processing

Background:

• Content analysis of psychotherapy transcripts is crucial for understanding therapeutic processes.
• Manual content analysis is time-consuming and subjective.
• Developing automated methods can improve efficiency and objectivity.

Purpose of the Study:

• To introduce and validate the Automated Co-occurrence Analysis for Semantic Mapping (ACASM) method.
• To assess ACASM's performance in content analysis tasks.
• To compare ACASM's results with those of human coders.

Main Methods:

• ACASM employs a context-sensitive, trans-theoretical approach to map text meanings.
• The method was tested on verbatim psychotherapy transcripts.
• ACASM's performance was evaluated against independent human coders on two tasks: semantic similarity estimation and semantic classification.

Main Results:

• ACASM's semantic similarity estimates were consistent with human coder estimates.
• The agreement between human coders and between human coders and ACASM was of similar magnitude for semantic classification.
• ACASM demonstrated comparable performance to human coders in content analysis.

Conclusions:

• The Automated Co-occurrence Analysis for Semantic Mapping (ACASM) provides a valid and reliable method for psychotherapy transcript analysis.
• ACASM's automated content analysis is indistinguishable from human coder analysis.
• This method offers a promising tool for advancing psychotherapy research through objective and efficient text analysis.