The paper presents the application of the metric methods to the analysis of the questionnaires used in various fields. The generic methodology is presented, including particular modules, responsible for the subsequent operations. They include generation of category patterns based on the available data, application of envelopes, dataset complexity assessment and performing classification of questionnaire results. Metrics applied in the presented research are then introduced. The methodology is tested on three data sets from the psychological, sociological and educational domains. Results show the advantage of our approach compared to the standard classification and decision making methods. Also, it may be used for the results interpretation, finding relations in data, or evaluation the test discriminating power (regarding each question separately).Proposed methodology may be found beneficial in all areas where questionnaire data is used - from classical diagnosis to HCI and big-data applications.
Digital Object Identifier (DOI)
Rafalak, Maria; Bilski, Piotr; and Wierzbicki, Adam
"Analysis of Questionnaire Results Using Metric Methods,"
Applied Mathematics & Information Sciences: Vol. 10:
4, Article 5.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol10/iss4/5