| Title | Validation of ChatGPT content in the Process Management topic |
|---|---|
| ID_Doc | 23816 |
| Authors | Vergara, ES |
| Title | Validation of ChatGPT content in the Process Management topic |
| Year | 2023 |
| Published | Gecontec-Revista Internacional De Gestion Del Conocimiento Y La Tecnologia, 11, 2 |
| Abstract | The purpose of the study is the content validation, at an exploratory level, of the ChatGPT responses to queries on the topic "Process Management and Status", based on a Theoretical Control Model, tested in previous empirical studies with high reliability and validity indexes. Forty queries were made to the ChatGPT, corresponding to the categories and subcategories of the Theoretical Control Model. The variables that ChatGPT should include in its answer sentences were also defined, following a "deductive" process by defining these codes a priori from the theoretical frame of reference. ChatGPT responses were processed and analyzed using NVivo software. Validation of each of the responses was performed using SPSS statistical software, considering only the variables included in each response, bearing in mind that data from 126 surveys from the previous empirical study were available. The results of the study indicated that the ChatGPT responses, in addition to being timely, were reliable and valid at sufficient levels, with Cronbach's Alpha reliability indices between 76.9% and 92.3% and "AVE" validity coefficients (percentage of variance explained) between 61.9% and 74.9%. However, each ChatGPT response was not complete according to the theoretical frame of reference, in each of the two categories none of the 20 ChatGPT responses included all of the variables, although in the "Process Management" category 16 responses (80%) included 6 or more variables out of the total of 10 (60%) and in the "Process Status" category 18 responses (90%) included 4 or more variables out of the total of 7 (57%). |
No similar articles found.