ABSTRACT
This chapter addresses fundamental methodological problems in the analysis of text by computer-aided content analysis based on word counts. It reconceptualizes methodological problems of word-count content analysis in terms of structural equation models, and suggests a framework for future research based on measurement models for content analysis. Authors will be concerned with the reliability of the assignment of entries to dictionary categories and with the reliability of the classification of the words in texts into these dictionary categories. Should the data indicate the need for multiple or double classification, then within the limits imposed by identification problems, the consequences of double or multiple classification can be explored in conjunction with the three problems.
