Ethical compromise, valuing life, and Fechner's Law

Abstract

Ethical compromise, valuing life, and Fechner's Law Authors: Nigel Harvey, Xiaozhou TanDepartment of Experimental Psychology, University College LondonDOI: https://doi.org/10.5281/zenodo.21940136Licence: CC BY 4.0 What this is Behavioural data from four online experiments on sacrificial moral dilemmas. In the scenario, a rocket will certainly kill 1 person but destroys a missile that would otherwise kill N people. N varies across ten items, and participants judge how likely they would be to act. Experiments 1 to 3 ask three questions at every item, giving 30 judgements per participant on a 0 to 100 scale. Experiment 4 replaces this with a choice between sacrifice and inaction followed by an odds judgement. All data were collected with the Gorilla Experiment Builder (gorilla.sc) and are deposited as raw exports, with no transformations applied to the responses. Start here CODEBOOK.md documents the study, the files and the traps. Section 2 is a lookup table giving every analysis variable, the columns that hold it, and the file it is in. dictionary_experiment1.csv to dictionary_experiment4.csv document every column of every file. Filter the used_in_analysis column to yes to see just the participant IDs, ratios, responses, attention checks and demographics. Three things are easy to get wrong, and all three are explained in the codebook: the ratio columns hold the number saved, not the number killed; in Experiments 2 and 3 a trailing b or c on a column name means the question, not a repeated block; and the plain Response column is not the analysis variable in Experiments 2 to 4. Files Experiment1/ expt1consent.csv Consent expt1demographics.csv Demographics expt1linear2.csv Dilemma task, ratio rising in steps of 2 expt1linear114.csv Dilemma task, ratio rising in steps of 114 expt1log.csv Dilemma task, ratio doubling (2^n) expt1reactionTimes.csv Debrief screen timings (not dilemma reaction times) Experiment2/ expt2R1.xlsx ... expt2R10.xlsx Dilemma task, one file per condition expt2demographics.xlsx Demographics. Use sheet `Sheet1`; the `demographics` sheet is one condition only. Experiment3/ expt3data.csv Dilemma task expt3demographics.csv Demographics Experiment4/ expt4data.xlsx Dilemma task expt4demographics.xlsx Demographics Experiment 1's files are one row per participant. Experiments 2 to 4 are one row per logged screen event, which is why they are large; responses accumulate in the Store: columns. Joining files Use Participant Private ID, Gorilla's internal anonymous key. In the Experiment 2 demographics workbook the same key is called pid. Private IDs are not reused across experiments. Ethics and privacy Collected under institutional ethics approval [insert approval reference]. Participant Private ID is retained as the anonymous linking key. Timing columns and the demographic variables (age, gender, country of residence, nationality) are retained as study data. TODO before publishing: confirm the paragraph above is true of the files as deposited. The following must be removed from every file first, and had not been at the time of writing: Participant Public ID, Schedule ID, and the device columns Participant Device Type, Participant Device, Participant OS, Participant Browser, Participant Monitor Size and Participant Viewport Size. expt3data.csv and expt3demographics.csv still contain live recruitment-platform IDs. Every affected column is flagged in the dictionaries with role set to identifier_remove or device_remove. See CODEBOOK.md Section G. Provenance Experiment 1 was collected by a member of the research lab who is not an author of this deposit. Experiments 2 to 4 were collected by the authors. Citation Harvey, N., & Tan, X. (2026). Ethical compromise, valuing life, and Fechner's Law. University College London. Zenodo. https://doi.org/10.5281/zenodo.21940136 Contact Prof. Nigel Harvey, Department of Experimental Psychology, University College London.