Integrating Interpretative Phenomenological Analysis and Sentiment Analysis for an Enhanced Understanding of Emotions in Human Experience

Date of Award

8-2024

Degree Name

Doctor of Philosophy

Department

Evaluation

First Advisor

Brooks E. Applegate, Ph.D.

Second Advisor

Regina L. Garza Mitchell, Ed.D.

Third Advisor

June E. Gothberg, Ph.D.

Fourth Advisor

Kevin Lee, Ph.D.

Keywords

Interpretive phenomenological analysis, IPA, mixed-methods, sentiment analysis

Abstract

Existing qualitative methods for understanding human experiences, such as phenomenological interviews, primarily aim to explore and describe cognitive and emotional content within the text corpus. During the interview process, emotions are often elicited through probing questions—yet various factors, including lack of rapport or sensitive topics, can hinder the full expression and capture of emotional components within lived experiences (LEs). Additionally, conventional methods for analyzing emotional aspects often focus on the presence and general nature of emotions within narratives, relying on subjective researcher interpretations. This can result in an incomplete understanding of the LE under study.

Sentiment Analysis (SA), an algorithmic method for extracting emotional content from text, may complement conventional qualitative methods in this methodological challenge. SA introduces a quantitative dimension to the analysis by measuring both the valence (positive, negative, or neutral) and the intensity of emotions expressed in the text.

This study explores the integration of Interpretative Phenomenological Analysis (IPA), a qualitative methodology aimed at understanding LEs, with SA to enhance the understanding of emotions contributing to the LEs of participants who have lived through a severe spinal cord injury. The study assesses the extent to which SA can identify and contribute new insights into the emotional characteristics of LEs documented in a text corpus and how this can augment established IPA analytical methods. A novel mixed-methods research design was proposed, adopting a qual-QUANT convergent approach that combines qualitative IPA with algorithmic SA. By analyzing data across three dimensions—theme-based, case-based, and cross-case—this approach provides a more comprehensive perspective, revealing nuanced emotional patterns.

The findings indicate that SA can inform the intensity and polarity of themes, participant emotions, and themes within participants in an IPA study by providing a weighted metric that facilitates relative affective comparisons and influence theme-tagged sentences. Furthermore, SA can enhance the iterative IPA analytical process, inform trustworthiness, and aid the researcher with practical aspects of the study, such as selecting interview extracts that best illustrate themes in the final report.

The study concludes that SA complements IPA analysis, providing a fuller representation of the LEs studied. The integration of SA into IPA has the potential to equip social science researchers with a new tool capable of measuring potentially unrecognized emotional content within interview data that was not explicit, particularly when studying sensitive topics. SA facilitates a deeper understanding of the emotional components of LE, enabling researchers to use it as a tool to find emotional content within cognitive descriptions. This practical application of SA allows for a more comprehensive and nuanced understanding of the emotional landscape within the data. For example, while IPA might identify a theme of “loss” in participant narratives, SA can quantify the negative emotions associated with that loss and compare it with the intensity of other negative emotions foundational to other themes or among participants. Ultimately, this innovative approach advances mixed-methods research and offers a tool for understanding the essential emotional components of human experiences.

Access Setting

Dissertation-Abstract Only

Restricted to Campus until

8-1-2034

This document is currently not available here.

Share

COinS