An Expectancy-Value Model of Emotion Regulation: Implications for Motivation, Emotional Experience, and Decision Making

情绪调节的期望价值模型：动机、情绪体验和决策的启示

Maya Tamir & Yochanan E. Bigman（The Hebrew University of Jerusalem）

Emily Rhodes, James Salerno, & Jenna Schreier（Boston College）

《Emotion》 2014

研究目的、假设

• 根据自我调节的期望价值模型，人们倾向于做他们认为对自己有用的事情。比如，比如人们认为奔跑可以让他们逃避天敌，那他们被天敌追赶的时候就会选择逃跑。甚至当人们不需要逃离什么事物时，如果他们认为跑步是有用的（比如促进智力的表现）他们也会倾向于跑步。
• 基于这个理论，我们提出了【情绪的期望价值模型】，即人们倾向于表现他们认为有用的情绪。本研究检验了情绪结果的期望在情绪偏好、情绪调节和决策行为上的效应。

研究方法

Study 1

假设

若被试学到“焦虑是有效的”则更倾向于体验焦虑。

程序

（线上测试，为了防止被试猜测实验任务，假装告诉被试要完成一个金钱有关的任务，在随后测验中加入小测试cover。）

• 69名美国成年人被试（30%男性），任务完成越好，可以获得的奖励越多。被试从之前完成任务较好的被试那里获取3条信息，所有被试阅读第一条（任务过程中仔细阅读指导语），实验组被试（anxiety= useful condition）的另外两条信息指向焦虑是有效的，控制组被试的另外两条信息指向动机是有效的。
• 随后，被试阅读网上的一个文章片段，并对他们偏好阅读的程度打分（a.情绪偏好测量），以及对他们当前的情绪打分。
• 插入混淆试听的金钱小测验
• 给被试呈现之前阅读的文章的标题，并对多大程度期望感受到焦虑、激动和有趣(b.结果期望，反映动机)进行打分。

工具

共发情绪评分（c.）：

被试对他们当前的情绪体验打分（0-一点都不，6-及其）。

情绪偏好测量(d.)：

随机给被试呈现6个标题，并让被试依次对想要阅读的标题打分（1-一点都不想读，6-非常想读），其中2个标题是诱发焦虑，2个诱发激动，还有2个是中性的。

结果

• 为了测试，期望状态的操纵是否影响了被试想要的体验，进行了情绪3（被试间：焦虑，中性，激动）期望状态2（被试内：焦虑组=有效，控制组）性别2（男，女）重复测量方差分析
  2-1.png
  情绪和期望情境的交互作用显著，F(2,66) = 4.84, p = .009, η2 = .07.认为焦虑有效的被试显著倾向于阅读诱发焦虑的文章，但激动和中性情绪组不显著。更多地，通过评估共发的情绪，我们得出操纵对情绪结果的期待（如更多奖励）可以影响体验某种情绪的动机，而不是体验情绪本身。

Study 2

目的与假设

• 检验实验一的结果是否在愤怒情绪下也有相同的效果；假设:认为愤怒情绪有效的被试更倾向于体验诱发愤怒的活动。

程序

• 实验室实验，66名美国大学生（50%女性）
• cover:告诉被试他们要完成一个可以测试他们专业潜能的新任务。读5个伪装的关于专业潜能测试的效度重要性等信息。
• 与study 1 相同，假装在之前表现良好的被试中获得5条提示，其中所有被试均阅读2条与情绪体验无关，另外3条作为操纵，实验组1：anger=useful condition, 实验组2：calmness =useful condition，控制组：无情绪相关提示(a.)

工具

共发情绪评分(b.)：

被试对当前体验的情绪打分（0-一点都不，8-及其），eg:愤怒（anger）， 高兴，平静（calmness）.

情绪偏好测验（c.）：

被试分别对多大程度他们想要听音乐（6个音乐片段，每段20s，2个诱发愤怒的，2诱发高兴的，2诱发平静的）、回忆过去的事情（6个事件，随情绪（愤怒、高兴、平静）和内容（与学校相关，与学校不相关）变化）和看电影片段（9段对电影内容的描述，激发愤怒的、高兴的、中性的各三段）进行打分（0-一点都不，8-及其），最后取每类情绪活动的平均分。

结果

• 重复测量方差分析，被试内变量：活动3（音乐，回忆，电影）情绪3（愤怒、高兴、平静）被试间变量：期望状态3（anger = useful; control, calmness = useful）*性别2（男，女）。
  2-2.png
  情绪和期望状态交互作用显著，F(4,61) = 5.02, p = .001, η2 = .14。anger=useful condition的被试更倾向于体验愤怒情绪，calmness=useful condition 更少倾向于体验平静情绪。
• 情绪主效应显著，F(2,63) = 74.41, p < .001, η2 = .55，相比起诱发中性和愤怒的活动，被试更倾向诱发高兴的活动。活动的主效应显著，F(2,63) = 22.52, p < .001, η2 = .27，相比起看电影和回忆，被试更倾向于听音乐；情绪和活动的交互作用显著，F(4.61) = 8.98, p < .001, η2 = .13，喜欢中性音乐高于中性电影和回忆；情绪和性别交互作用显著，F(2,63) = 5.94， p = .003, η2 = .09，女性比男性更喜欢诱发高兴的活动，该结果在情绪活动（喜欢音乐）性别的三元交互作用中更显著。

Study 3

假设

1. 认为愤怒是有效（或无效）的被试更可能选择诱发愤怒情绪的活动并体验到更多愤怒。
2. 根据期望-价值模型，人们希望参与他们期望会产生激励的行动，且在某种程度上这些激励措施与个人价值有关。

程序

• 69名美国大学生（61%女性）
• 被试随机分为三组,condition 1：anger = useful/task =
  relevant)，condition 2：anger = useful/task = irrelevant,condition 3：anger = irrelevant/task = relevant
  *每名被试阅读6个建议，其中四条随condition变化，作为操纵
• 对当前的情绪打分→对他们喜欢的音乐和回忆打分→听三首他们最喜欢的歌→再次对当前的情绪打分→对接下来的任务的准备程度打分

工具

共发情绪量表

同study 2,被试自评愤怒（前后测）和高兴（前后测）

情绪偏好量表

同syudy 2，材料为音乐和回忆事件

结果

• 重复测量方差分析：Activity (music, memories) 和Emotion (angry, neutral, happy) 被试内变量，Condition (anger = useful/task = relevant, anger = irrelevant/task = relevant, and anger = useful/task = irrelevant) 和 gender 被试间变量。Emotion*Condition 交互作用显著, F(4, 64) = 3.74, p = .007, η2 = .11，3, anger = useful/task = relevant组更偏好诱发愤怒的活动。

  
  2-3.png

• 期望的有效性&情绪调节：重复测量方差分析，Time (pre- and postinduction) 和 Emotion (anger, happiness) 被试内变量，Condition 被试间变量.时间的主效应显著，前测比后测情绪强度强；情绪的主效应显著，被试报告更多的开心更少的愤怒；情绪状态的交互作用显著，anger = useful/task = relevant condition组报告更多的愤怒。EmotionCondition*Time 交互作用显著，简单效应检验，anger = useful/task = relevant condition组被试听音乐后报告更高的愤怒和更低的高兴评分。线性回归:Condition是愤怒情绪偏好t(61) =2.80, p = .007和愤怒情绪体验t(61) = 2.57, p = .013的显著预测变量.
  

  2-4.png

• Expected usefulness and perceived readiness
  significant Condition* Anger interaction, t(41) = 2.13, p = .039.简单效应不显著，but the association between anger experience and wad positive in the anger = useful condition.
  feeling prepared

  
  2-5.png

Study 4

假设

被诱导认为愤怒有效（或无效）的被试倾向在对风险敏感的赌博任务中输钱且更少的亲社会。

程序

• 63名美国大学生被试（52%女性）

工具

• 操纵检验
  被试对不同心理状态的有效性打分（0-8），包括愤怒和易怒等
• 共发情绪评估
  同study 3：测愤怒和高兴
• 哥伦比亚卡片任务（CCT）
  测试冒险行为
  *亲社会任务
  测量亲社会判断
• Berlin Numeracy Test 柏林计算能力测试
  代表了冒险文化？？？

结果

• Manipulation check、Expected usefulness and emotional preferences、Expected usefulness and emotional experiences与前几个研究结果一致
• Expected usefulness and decision-making behavior

1. expectancy manipulation influenced emotion experience, which in turn, was associated with behavior.
   participants in the anger = useful condition performed significantly worse than others on CCT(认为愤怒有用的人越冒险)

• the effects on the CCT were mediated by anger experience，which in turn, was associated with behavior.The results demonstrated the downstream effects of expected usefulness of anger on decision-making behavior.

2. participants in the anger = useful condition made significantly less prosocial judgments, compared to others.

• 总结：Compared to participants in the control condition, participants who were led to expect anger to be useful behaved in a more anger-consistent manner, by taking greater financial risks and losing more money in a gambling task and by allocating smaller sums of money to others in hypothetical dictator game scenarios

Study 5

假设

Participants in the anger = useful priming condition to show stronger preferences than other participants for anger-inducing music.

程序

• 60名美国大学生（48.3%女生）
• condition 1:anger words (e.g., fury) were associated with high utility words(e.g., helpful).condition 2:anger
  words were associated with low utility words (e.g., harmful), condition 3: calmness words (e.g., calm) were associated with high utility words.(priming)
  *####工具

Implicit expectancy manipulation task

2-6.png

• Indicators of emotional preferences、Concurrent emotion ratings

结果

• significant Emotion * Condition interaction, F(2, 57) = 3.29, p=.045, η2 = .11.participants in the anger = useful condition tended to express stronger preferences for anger-inducing music and weaker preferences for happiness-inducing music compared to participants in the other conditions.

  
  2-7.png
• 总结：Expected usefulness of anger can influence emotional preferences even when primed outside of conscious awareness.By including an anger =
  harmful condition, we demonstrated that the effect was not due to priming the concept of anger per se(本身). By including a calmness =useful condition, we demonstrated that the effect was not due to priming the concept of usefulness per se. Instead, stronger preferences for anger were found only among participants who were primed with an association between anger and usefulness, in particular.

总结

In five studies, we motivated participants to increase an unpleasant emotion simply by making them believe that it may be useful to them. This effect occurred when we manipulated the expected usefulness of anxiety (Study 1) and anger (Studies 2–5). The expected usefulness of anger influenced the emotion-inducing activities people selected and how angry and happy they felt upon engaging in these activities (Studies 3–4). These findings show that the expected usefulness of anger does not change the spontaneous experience of anger, but alters anger experience through active regulation. We further demonstrated that the expected usefulness of anger has downstream effects on anger-consistent behavior (Study 4), such that people who were led to expect anger to be useful lost more money in a gambling task and reported they would give less to others. Finally, we demonstrated that the expected usefulness of anger can motivate people to increase their level of anger even when it is activated outside of conscious awareness (Study 5). Taken together, these findings provide strong support for the validity of the expectancy-value model of emotion regulation.