Epistemic Trespassing & Disagreement

This dataset presents the results of four studies conduced on Amazon Mechanical Turk where participants made judgments about the social acceptability of conversational disagreement in a number of different scenarios. Disagreement was examined in a number of differed knowledge domains (which are more or less central to the speakers) and between people with different relationships and levels of expertise. In Study 1&2, participants were presented with 32 two-line conversations that encompassed ‘core’, ‘mid’ and ‘peripheral’ knowledge domains. (See the ‘text’ column of Data_Stimuli_1_2). In Study 2, participants were provided with brief backstories that contextualized these scenarios. They were either meant to elevate the knowledge of speaker A (the “Epistemic A>B Backstory” column) or of Speaker B (the “Epistemic B>A Backstory” column). Study 1A collected participants’ acceptability judgments on a 7-point scale from “very polite” (1) to “very rude” (7) for each of these 32 scenarios (see Data_1A). In Study 1B, participants were asked to invent different contexts which would change the acceptability of these conversations (making them either more rude or more polite) by typing into open text fields. They were instructed not to change the wording of the conversation and data of participants that consistently did so were removed from the data. Data_1B_PoliteRaw and Data_1B_RudeRaw show the responses entered by subjects. Data_1B_CodingScheme shows the procedure for coding these responses. Data_1B_RudeCodes and Data_1B_PoliteCodes show the codes for these responses. Note that each response has both a superordinate and subordinate code. Studies 1A & 1B were collected in two parts (see the ‘batch’ columns). In Study 2, participants were asked to judge the acceptability of the same stimuli paired with one of four backstories. Two were ‘epistemic’ (see the Data_Simuli_1_2 file for these; P.EPI refers to Epistemic A>B and R.EPI refers to Epistemic B>A) and two were ‘social’ (participants were told either that the two speakers were best friends (P.SOC) or strangers (R.SOC)). Data_2 contains participants’ judgments for these backstories on the same 1-7 scale. There were four different counterbalanced lists presented to subjects (the “vers” column refers to these) In Study 3, participants were given similar stimuli (3 conversations about art, hockey and piano) and told the relative number of years of experience the two speakers had (see Data_Stimuli_3) Each subject only judged one conversation The Data_3 file shows subjects acceptability judgments on the same 1-7 scale.本数据集呈现了四项在亚马逊众包平台（Amazon Mechanical Turk）上开展的研究结果，参与者需针对多种不同场景下的对话分歧（conversational disagreement）的社会可接受性（social acceptability）作出判断。研究考察了多个不同知识领域（knowledge domains）中，以及不同社会关系、专业水平人群之间的对话分歧情况，这些知识领域对说话者而言的重要程度各有不同。在研究1与研究2中，参与者会接触到32组两行式对话，涵盖“核心”“中等”与“边缘”三类知识领域（详见Data_Stimuli_1_2的"text"列）。在研究2中，参与者会获得用于为这些场景补充背景的简短故事：一类用于提升说话者A的知识权重（对应"Epistemic A>B Backstory"列），另一类用于提升说话者B的知识权重（对应"Epistemic B>A Backstory"列）。研究1A针对上述32组场景，让参与者以1（极礼貌）至7（极粗鲁）的7级评分量表（7-point scale）完成可接受性判断（详见Data_1A）。研究1B中，参与者需通过开放文本框自行构思不同背景，以改变这些对话的可接受性（使其更粗鲁或更礼貌），且要求不得修改对话原文，不符合该要求的参与者数据将被剔除。Data_1B_PoliteRaw与Data_1B_RudeRaw展示了受试者提交的回复，Data_1B_CodingScheme说明了对这些回复进行编码的流程，Data_1B_RudeCodes与Data_1B_PoliteCodes则对应了这些回复的编码结果。请注意，每条回复均同时具备上位编码与下位编码。研究1A与1B的数据分两批次收集（详见"batch"列）。在研究2中，参与者需对搭配了四类背景之一的相同实验刺激对话进行可接受性判断：两类为认知优势型背景（详见Data_Simuli_1_2文件，其中P.EPI指代Epistemic A>B，R.EPI指代Epistemic B>A），另外两类为社会关系型背景（告知参与者两位说话者为挚友（P.SOC）或陌生人（R.SOC））。Data_2包含了参与者针对这些背景的1-7级评分量表判断结果。本次研究共向受试者呈现了四组平衡后的测试列表（"vers"列即对应这些列表）。在研究3中，参与者会接触到类似的实验刺激（3组分别涉及艺术、曲棍球与钢琴的对话），并会获知两位说话者各自的相对从业年限（详见Data_Stimuli_3）。每位受试者仅需评判一组对话。Data_3文件展示了受试者的1-7级可接受性判断结果。