Abstract: Social anxiety is a state of excessive worry, nervousness, and anxiety that individuals experience in social situations. Examining the cognitive processes of individuals with varying degrees of social anxiety symptoms can help to further understand the cognitive mechanisms. Individuals with social anxiety tend to have irrational social evaluation expectations. The Violation of Expectation model describes the formation of expectations through two processes: general expectation and situational expectation. Additionally, an individual's mental illness symptoms (e.g. social anxiety) can influence the expectation process. Previous research has focused on the effect of social anxiety on expected evaluation outcomes, which may ignore the dynamic process of situational expectations. To address this issue, this study used the drift-diffusion model (DDM) to analyze the formation process of situational expectations. We not only verified the difference in valence (positive vs. negative), but also further explored the moderating effect of social anxiety. The aforementioned DDM allowed us to examine of the parameters associated with the process of situational expectations, including the starting biases, drift rates, non-decision time, and threshold. A total of 85 participants were included in the analysis and data collection was conducted online through Credamo. Social anxiety levels were measured using the short version of the Social Interaction Anxiety Scale and Social Phobia Scale, and general social expectation were measured using the adapted General Social Expectations Scale. To explore the formation of situational social evaluation expectations, we first elicited subjects' expectations through a structured interview in which we pretended that eight audiences of similar age were watching. A social evaluation expectation task was then designed in which subjects were asked to anticipate whether the audiences would describe them using some trait adjectives displayed. Pressing F represented yes and J for no. The experiment included two blocks, each containing 40 trials, with breaks set between blocks. In each block, there were 20 trials with positive social-related adjectives and 20 trials with negative social-related adjectives, and the adjectives were not repeated between the two blocks. The DDM model was optimized using the Kolmogorov-Smirnov method. According to a previous study, our research specified a DDM with starting biases and drift rates depending on the experimental conditions (i.e., adjectival valence). Then, we checked the model fit individually using a simulated study. SPSS 22.0 and R-based Jamovi software were used for statistical analysis. First, paired-sample t-tests were used to examine the differences of DDM parameters, accuracy rates, and response time across conditions. Second, correlation analyses were used to reveal the relationships between social anxiety, general social expectation, and DDM parameters. Third, general linear models were used to test the moderating effect of social anxiety on the relationship between general and situational expectations. The results indicated that positive evaluation expectations had higher drift rates and starting point biases than negative evaluation expectations. This suggests that participants were more likely to accumulate evidence confirming positive expectations and had a stronger prior bias toward positive expectations. However, there was no significant difference in the absolute value of the drift rate between positive and negative expectations, indicating that the direction of drift rate matters for the valence difference rather than the rate. Social anxiety and general social expectations significantly influenced the drift rate of positive and negative evaluation expectations. Social anxiety weakened the relationship between general social expectations and drift rate but strengthened the relationship between general social expectations and starting point bias, only in terms of the positive evaluation expectations. This indicates that high social anxiety may impair the formation of positive self-bias in social evaluation expectations, leading to a more negative overall evaluation. This study used the DDM to reveal the process of situational evaluation expectations. The results validated the positive self-bias of social expectation, and examine Violation of Expectation model in the field of social evaluation expectations. The moderating role of social anxiety in the formation of social evaluation expectations was demonstrated from a new perspective. This study provides new perspectives for understanding the process by which social anxiety influences the formation of social expectations.

Key words: social anxiety, social evaluation expectation, model of violated expectations, drift-diffusion modeling

摘要： 社交焦虑的个体具有异常的社会评价预期机制。预期违反模型认为预期包括一般性预期和情境性预期,一般性预期为情境性预期的形成提供信息且过程中受到个体差异的影响。在社会评价研究领域,尚未探索社会评价预期的形成过程和社交焦虑对这一过程的影响机制。研究使用漂移扩散模型来解析情境性社会评价预期,旨在探究积极与消极预期过程的差异,并考察社交焦虑和一般性社会预期对情境性社会预期过程的影响。结果发现,相对于消极评价预期,积极评价预期信息积累速度更快、起点更高;社交焦虑在一般性社会预期和情境性预期过程（信息积累速度和起点偏向）中发挥调节作用。研究发现为理解社交焦虑影响社会评价预期的形成过程提供新的视角。

关键词: 社交焦虑, 社会评价预期, 预期违反模型, 漂移扩散模型