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.