Abstract: To explore the stability and variation of the indicators of subjective well-being. the classic latent state-trait model, Trait-State-Occasion model(TSO Model)and Indicator-Specific Trait model(IT model)were employed to evaluate the stability and variation of life satisfaction, positive emotion and negative emotion, based on a one-year and three-wave longitudinal study on a sample of 360 college students. The results showed that the IT model had a good fiting, All variables′ consistency coefficients(CO)were significantly higher than occasion-specificity coefficients(Ospe)in this model, which indicated that subjective well-being was more likely to be a personality trait. At the same time, the research found that the consistency coefficients of positive affection and negative affection were much higher than the life satisfaction coefficient, which suggested that people′s positive emotions and negative emotions were more stable than life satisfaction. This can be attributed to the fact that positive and negative emotions are more related to the genetics and personality traits of the individual. Therefore, these two coefficients are quite stable within a certain period of time. Nevertheless, life satisfaction is not only influenced by the cognitive methods of individuals but also by their life situation and how they evaluate their lives. Thus, life satisfaction is more influenced by contextual elements.

Key words: subjective well-being, stability, variation, latent state-trait model.

摘要： 为了探究主观幸福感各项指标的稳定性及其变化的情况，在一年的时间内对360名大学生的主观幸福感进行了3次追踪测量，分别采用潜特质-状态模型（LST Model）中的经典模型、特质-状态-情境模型（TSO Model）以及测量指标-特殊特质模型（IT Model）对生活满意度、积极情感、消极情感等主观幸福感指标的稳定性和变化性进行评估。结果表明，只有测量指标-特殊特质模型与数据拟合良好，该模型中各变量的一致性系数都显著的高于情境特异性系数，这说明，主观幸福感更可能是一种稳定的人格特质。此外，研究还发现积极情感和消极情感的一致性系数远高于生活满意度的一致性系数，这表明人们的积极情感和消极情感比生活满意度具有更高的稳定性。

关键词: 主观幸福感, 稳定性, 变化性, 潜特质-状态模型。