A pilot study to analyze descriptive questionnaire responses with natural language processing techniques

The main purpose of the research reported here is to show that the questionnaire responses in text format can be processed and statistically analyzed to some extent by computer programs. We employ unigram, bigram and trigram frequency calculations with the assistance of Python programming in order to analyze the questionnaire responses in text format.
Questionnaire responses chosen from options can be easily statistically analyzed whether they are on a ratio, interval, ordered or nominal scale. In contrast, descriptive questionnaire responses have long been read and interpreted by researchers themselves. They have been analyzed by human labor alone. This could to some extent lead to arbitrary and subjective interpretations and analyses. We demonstrate in this paper that natural language processing techniques in Python could help analyze questionnaire responses in text format more objectively and accurately.