This study investigates how to construct a valid text--‐based measure of the job affect dimensions boredom and enthusiasm using computer--‐aided text--‐analysis (CATA). Following a recent guideline for CATA research (McKenny et al., 2018; Short et al., 2010), the study aim was to conduct a systematic approach towards validation of the text--‐based measures and analyse the convergent and discriminant validity to job--‐ and life--‐related behaviour constructs. The text--‐based measures were developed from responses to a semi--‐open question in a job satisfaction survey with 364 English--‐speaking respondents. Interrator reliability tests resulted in scores that were indicative of high reliability. Convergence validity results showed that the convergent correlation of the text--‐based measure of boredom was significantly stronger than discriminant correlations. The convergent correlation of the text--‐based measure of enthusiasm, however, was not significantly stronger than discriminant correlations. The relationships offered substantial evidence for the discriminant validity of the constructed text--‐ based measures. In the end, the limitations of this study were discussed that could be addressed in future research. The study showed that semi--‐open questions with job affect words can be used to construct a context--‐specific text--‐based measure to analyse the relationships with job--‐ and life--‐ related behaviour constructs. Therefore, it might be interesting to expand this study field in further CATA research by analysing the added value of open questions in similar survey instruments which considered closed questions as the standard. The study could also function as an inspiration for scholars interested in content analysis studies of job satisfaction using CATA.

Additional Metadata
Keywords CATA, job affect, open questions, semi-­‐open question, text-­‐based measure, reliability, convergence validity, discriminant validity
Persistent URL hdl.handle.net/2105/49451
Series Economics
Citation
Validation of a Text-­‐based Measure of Job Affect. (2019, November 7). Validation of a Text-­‐based Measure of Job Affect. Economics. Retrieved from http://hdl.handle.net/2105/49451