Examining the antecedent factors and their influences on English language anxiety via the lens of control-value theory and data mining approaches

English learning anxiety (ELA) is often observed in the population of Chinese university students from STEM (Science, Technology, Engineering, Mathematics) subjects, which hinders their English acquisition and improvement. Most previous studies consider ELA as a regular emotion and measure it based on the foreign language classroom anxiety framework. Despite remarkable findings, we argue that the origins of ELA remain unclear. To address this concern, the present study revisits ELA from the lens of the control-value theory. By defining ELA as a type of achievement emotion, we assumed that ELA is the product of the high value in English learning and the low ability for learning. Hence, this study examined how students’ self-efficacy, self-regulated learning abilities, intrinsic goal orientation, and task value influenced ELA. The results of decision tree classification demonstrate that the four variables can positively predict students’ ELA. The feature importance analysis shows that task value is the most important predictor feature than the others, which confirms that ELA results from the high value of English learning and the weak SRL abilities. The study discusses the implications of addressing students’ ELA according to the control-value theory. Copyright © 2023 The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.