A Development and Evaluation of Keywords to Automatically Categorize Research Methods for Research Synthesis

• In Event: Advances in Text Analysis Methods

Abstract

Systematic reviews and meta-analysis are known to be very costly in terms of time and labor when conducted by manual coders. Due to the increasing importance and the immense workload associated with research synthesis, scholars have recently started to use computational methods in order to ease the process. Accordingly we seek to develop, employ and to evaluate the application of a machine based content analysis via a dictionary approach. As such we aim to identify via a keyword search the methods used in communication science studies. Our results reveal that some methods are better suited for detection by a keyword search than others. Thus, we achieved good recall for quantitative studies, experiments and surveys. However, content analysis and observations are not well enough detectable (yet). Possibilities to further improve the approach are discussed.

Authors

• Emese Domahidi, Leibniz-Institut fuer Wissensmedien

• Elisabeth Guenther, U of Muenster