From Readers to Data - JCLS 2025

data (EXCEL) and code (Matlab 2024b) for JCLS submission Data 240813 - Key Novel Dataset - 9 - removed pilot entries.xlsx This file has been manually pre-processed to remove pilot questionnaires (that were incomplete) and fix other errors such as incorrect spelling of book names and authors. In addition, we manually added a pair of columns (author 2 and gender of author 2) to enable us to deal with books with two authors, in which it was impossible to identify either because they were combined. Data extraction code extractTables2.m This code reads the excel file and generates the necessary data structure in variable D, then saved in the Matlab data file allData250203.mat fig1_ambigScale.m Further analysis of scaled items and generation of figures 1A-C. fig2_modesOfAmbivalence.m Futher analysis of ambivalent items and generation of figures 2A-C. fig3a.m Analyses whether the identity of the book/title affects the degree of ambivalence reported by readers. fig3b.m Analyses whether the reader identity affects the degree of ambivalence reported across all questionnaires of that reader. Accessory code directPoissonBinomial.m A self-written function to calculate the Poisson-Binomial distribution. Input: p - a vector of n elements containing the probability of success for each Bernoulli trial. k - the number of successful Bernoulli trials (irrespective of position). Output: prob - the probability of achieving exactly k successful trials cellflat A helper function to flatten nested cell arrays. Input: celllist - a cell array to be flattened n - an optional input, limiting the number of flattened levels to n. Output: out - a flattened cell array Accessory data structures These are used to analyse some of the questionnaire items, by providing a table that allows translation between answer and some number / vector that can be used for analysis. Most of them are irrelevant for the current manuscript, but are necessary for the code to run. Included are: charNumKey.mat Used for transforming the verbal answers into a number of main characters (first column) and secondary characters (second columns). Numbers in the range 0-4 should be interpreted literally. 10 encodes "several" and variations thereof, and 100 encodes "many" and variations thereof. defaultCitations.mat The item asking about sources cited in the novel has both multiple choices and an open field. This data structure contains the pre-set multiple choices. defaultGenres.mat The item asking about genre types has both multiple choices and an open field. This data structure contains the pre-set multiple choices. evtNumKey.mat Number of key events in novel. This item contains both pre-set multiple choice and free text, and has to be translated into numbers. As before, 10 denotes "several" and variations thereof, 100 denotes "many" and variations thereof. geoData.mat Contains a data structure generated manually with all the 63 geographical entities that are given as answers in the item about geographical locations mentioned in the novel. Contains two data structures: geoEntity - a 5x63 cell array, with each column providing the continent/region/country/city/entity corresponding to one possible answer. Region: e.g. SE Asia, W Europe, ... Entity: e.g. military base. If the answer included only a continent, rows 2-5 will be empty. But if only a city was mentioned, the column will contain rows 1-4 and only row 5 will be empty. geoHier - a 9x63 binary matrix. Rows 1-5 indicate whether the entity corresponds to one of the above positions in the hierarchy. Rows 7-9 correspond to the following 3 categories: (7) undefined territory; (8) historic entity (e.g. Babylonian empire); (9) unrealistic entity (e.g. fictional island) importData.mat Translates multiple choice answers about the impotrance of the novel into binary categories defined by us. languagesUsed.mat Translates user free text answers into a code that can be analysed. anonID.mat Contains the anonymised ID of the reader, encoded as a number.本数据集（EXCEL格式）与代码（Matlab 2024b）用于提交至JCLS的研究工作。 Data_240813 - Key Novel Dataset - 9 - removed pilot entries.xlsx：本文件已完成人工预处理，剔除了未完成的预调研问卷，并修正了书籍名称、作者姓名拼写错误等各类问题。此外，我们手动新增了两列数据（第二作者及第二作者性别），以处理存在两位作者且原数据中作者信息合并导致无法区分的书籍样本。 extractTables2.m：该代码读取上述Excel文件，生成变量D所需的必要数据结构，并保存至Matlab数据文件allData250203.mat中。 fig1_ambigScale.m：用于对量表条目开展进一步分析，并生成图1A至1C。 fig2_modesOfAmbivalence.m：用于对矛盾态度条目开展进一步分析，并生成图2A至2C。 fig3a.m：用于分析书籍/标题的标识是否会影响读者报告的矛盾态度程度。 fig3b.m：用于分析读者自身的标识是否会影响该读者所有问卷中报告的矛盾态度程度。 ### 附属代码 directPoissonBinomial.m：自主编写的用于计算泊松二项分布（Poisson-Binomial distribution）的函数。输入参数：p为包含n个元素的向量，代表每项伯努利试验（Bernoulli trial）的成功概率；k为成功的伯努利试验次数（与试验顺序无关）。输出参数：prob为恰好获得k次成功的概率。 cellflat：用于展平嵌套元胞数组（cell array）的辅助函数。输入参数：celllist为待展平的元胞数组；n为可选输入参数，用于限制展平的层级数。输出参数：out为展平后的元胞数组。 ### 附属数据结构 此类数据结构用于分析部分问卷条目，通过提供可将答案转换为可用于分析的数值/向量的对照表。其中大部分数据结构与当前手稿无关，但为代码正常运行所必需。具体包含以下文件： 1. charNumKey.mat：用于将文本答案转换为主要角色数（第一列）与次要角色数（第二列）。数值0至4应按字面含义解释；10代表“若干”及其变体表述，100代表“许多”及其变体表述。 2. defaultCitations.mat：针对“小说引用来源”的调研条目同时包含多项选择与开放填写字段，该数据结构存储了预设的多项选择选项。 3. defaultGenres.mat：针对“小说体裁类型”的调研条目同时包含多项选择与开放填写字段，该数据结构存储了预设的多项选择选项。 4. evtNumKey.mat：用于将“小说关键事件数量”的调研答案转换为数值。该条目同时包含预设多项选择与自由文本选项。同前述规则，10代表“若干”及其变体表述，100代表“许多”及其变体表述。 5. geoData.mat：包含手动生成的数据结构，涵盖了小说地理位置调研条目中所有63个可能的回答的地理实体。该文件包含两个数据结构： - geoEntity：5×63的元胞数组，每一列对应一个可能回答的大洲/区域/国家/城市/实体。例如：区域可指代东南亚、西欧等；实体可指代军事基地等。若回答仅包含大洲，则第2至5行留空；若仅提及城市，则该列的第1至4行将被填充，仅第5行为空。 - geoHier：9×63的二值矩阵。第1至5行用于标识该实体对应层级中的具体位置。第7至9行对应以下三类：(7) 未明确领土；(8) 历史实体（如巴比伦帝国）；(9) 非现实实体（如虚构岛屿）。 6. importData.mat：用于将关于小说重要性的多项选择答案转换为我们预设的二元分类结果。 7. languagesUsed.mat：用于将用户自由文本填写的语言答案转换为可用于分析的编码。 8. anonID.mat：包含读者的匿名ID，以数值形式进行编码。