Analyzing responses to open ended questions for SPIRIT using aspect oriented sentiment analysis
Abstract
Open ended questions provide an effective way of measuring the attitude and perception of respondents towards a topic. Surprising Possibilities Imagined and Realized through Information Technology (SPIRIT) was a program (2008-2012) that employed open-ended questions to gauge program participants' attitudes related to computing. SPIRIT sought to increase the interest of high school students, especially female students, towards computing courses and careers. Pre- and post-attitude surveys were used during the program to measure the changes in attitudes of the participants towards IT and also to analyze the impact different sessions had on different demographic groups of participants. The open-ended survey questions from SPIRIT provide the data needed for this study's analysis. SPIRIT's external evaluator employed the constant comparison method to analyze the participant data. This study analyzed those same responses using aspect-oriented sentiment analysis to make reporting and decision making for such programs easier and more objective than human evaluation. The approach identified the aspect of each phrase or statement made in the responses and then quantitatively classified the sentiment of each aspect. Thus, the study's approach not only solves the problem of objectively analyzing the open-ended responses of participants of short term educational programs similar to SPIRIT but also may help mine new information from the surveys that would help make decisions in order to make future programs have a better impact on the participants.
Degree
M.S.
Advisors
Harriger, Purdue University.
Subject Area
Educational tests & measurements|Information Technology|Gender studies
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.