An investigation of how formal features in audio-visual narratives influence audience members' empathy and prosocial behaviour

视听叙事对观众对世界的感知和对他人的态度有着强大的影响。引导叙事对受众产生影响的一个关键心理过程是移情，它被定义为对被观察者的心理状态的理解和体验。唤起同理心对于激发听众的亲社会（帮助）反应是有效的，这是一个运转良好的人类社会的基础。尽管视听叙事在叙事加工中很重要，但视听叙事的组成部分如何影响观众的共情反应仍不清楚。之前关于共情的研究只关注虚构、叙事和内容的影响，而忽视了形式特征的重要性。由于这一差距，媒体接收过程的基本部分仍然缺乏理论。本研究建议传播学、心理学和计算机科学的学者进行跨学科合作，以确定视听虚构叙事中能够高精度预测观众共情和亲社会行为的形式特征，并测试个体共情相关差异的调节作用。(1)研究1：我们从电影中随机抽取100个电影场景，描述人类角色之间引起共情的社会互动，并以定量的方式分析它们的形式特征。对形式特征的分析包括色彩、灯光、运动、镜头比例、镜头长度、镜头角度和音乐等指标，这些指标被认为对电影场景的情感认同很重要。对于视听分析，我们采用情感计算中开发的计算机算法，使我们能够以高分辨率检查这些特征。接下来，我们进行了一个实验，在这个实验中，每个参与者都看到了100个电影场景中的一个。观众的同理心是通过多种方法评估的。每个电影场景的数据集通过其形式特征、引起共情的内容以及观众引起的共情反应的度量来描述，并提交多层回归来测试形式特征与观众共情之间的因果关系。(2)研究二：通过实验检验个体差异（特质共情、情感需求、读心动机）的调节作用。根据共情效应的显著性和效应量，从研究1中选取两个视听形式特征，考察个体特质是否会影响形式特征对共情和亲社会行为的影响。研究结果为媒体加工中形式特征和个体差异的主要作用和交互作用提供了新的理论知识。基于这一结果，也将有可能设计和制作在未来更有效地对观众的共情和亲社会行为产生高影响的视听信息。

Audio-visual narratives have a powerful impact on audience members' perception of the world, and attitudes towards others. A key mental process that channels the impact of narratives on audiences is empathy, defined as the understanding and experiencing of mental states of an observed other person. Evoking empathy is effective in eliciting audiences' prosocial (helping) responses, the occurrence of which is fundamental for a well-functioning human society. Despite its importance in narrative processing, it is still unclear how the components of audio-visual narratives influence viewers' empathy response. Previous research into empathy focused only on the impact of fictionality, narrativity, and content but neglected the importance of formal features. Due to this gap fundamental parts of the media reception process remain undertheorized. The present project proposes an interdisciplinary collaboration among scholars in communication science, psychology, and computer science to identify those formal features in audio-visual fictional narratives that are capable of predicting viewers' empathy and prosocial behaviour with high accuracy, as well as to test the moderating effect of individual empathy-related differences. (1) Study 1: we select a random sample of 100 film scenes from movies depicting empathy-eliciting social interactions among human characters and analyze them for formal features in a quantitative way. The analysis of formal features includes metrics of color, lighting, motion, shot scale, shot length, camera angle, and music, assumed to be important to the emotional identity of a film scene. For the audio-visual analysis we employ computer algorithms developed in affective computing that enable us to examine these features at high resolution. Next, we conduct an experiment in which each participant is presented one of the hundred film scenes. Viewers' empathy is assessed by mixed methods. The dataset in which every film scene is described by metrics of its formal features, empathy-eliciting content, as well as viewers' evoked empathy response is submitted to multilevel regression to test the causal relationships between formal features and viewers' empathy.(2) Study 2: the moderating effects of individual differences (trait empathy, need for affect, mindreading motivation) will be tested in an experiment. Two audio-visual formal features will be selected from Study 1 based on significance and effect size of effects on empathy to examine whether individual traits modify the effect of formal features on empathy and prosocial behavior.Findings generate new theoretical knowledge on the main and interaction effects of formal features and individual differences in media processing. Based on this result it will also be possible to design and produce audiovisual messages that have a high impact on audiences' empathy and prosocial behavior more effectively in the future.