Assessing Quality of Life through Natural Language Use: Implications of Computerized Text Analysis

Quality of life (QOL) is a broad, multiply determined construct with both objective and subjective components. Objective indicators of QOL can include physiological markers of fitness and disease, a nation’s Gross Domestic Products, marital status, church attendance, or even the number of minutes per day spent laughing. But objective indicators do not always reflect one’s own assessment of or satisfaction with life: Cancer patients can exude vitality, rich people can think that they don’t have enough, married people can feel isolated, church goers can experience meaninglessness, and laughter can be a sign of anxiety. It becomes clear, then, that emotional state, belongingness, and cognitive judgments are important in the assessment of QOL. These subjective components of QOL are typically measured by self-report methods, including surveys or interviews. However, self-report methods suffer from their own shortcomings. Surveys are susceptible to response bias. The reference group on which a respondent heavily relies to makes comparative QOL judgments is often unknown. Furthermore, surveys are constrained by what the researcher assumes constitutes QOL. Standardized or open-ended interviews can be appealing in getting participants’ broader world views. The recurring frustration with interviews, however, is in finding reliable and efficient ways of coding them. In this paper, we bring together empirical evidence showing how computerized text analysis programs can assess natural language in order to reveal emotional state, belongingness, and cognitive judgments. We also present data that has linked natural language use to objective indicators of QOL. Our understanding of natural language use within the context of QOL is just in its infancy. Nevertheless, preliminary analyses are beginning to yield some important insights.