Content analysis and computational linguistics: its quickness, reliability and perspectives

Abstract

Content analysis is a technique that converts open-ended responses into categories. This process is of great value since it defines the categories of a study based on the perception of the sample, avoiding imposed categories created by the researcher. However, this type of analysis involves extensive use of time, resources, and expertise. Programs such as ATLAS.ti or NVivo do not constitute an effective nor efficient solution. New software based on computational linguistics offers a different scenario, as it allows the “understanding and interpretation” of categories. In order to prove its effectiveness and efficiency, content analysis made by experts is compared with analysis using SPSS Text Analytics for Surveys (TA). We conclude that under the supervision of a specialized researcher, TA allows for an important time saving, increased reliability, and opens up possibilities for qualitative analysis of large samples.