D5.1 Experimental Datasets

D5.1 Experimental Datasets Brief description This folder contains the experimental datasets for the ABHuBe bargaining experiments. It groups datasets from the initial bargaining experiment and from the later Turing-test experiment, including round-level bargaining traces, participant identifiers, role assignments, pie sizes, payoffs, AI/human guess fields, and participant feedback. Contents Initial bargaining experimental datasets Location: Initial bargaining experimental datasets/ BargainingData.csv: initial bargaining dataset from previous work with 12,864 rows and 218 columns. It includes session/period identifiers, uninformed and informed participant IDs, payoffs, pie size, deal status, time-indexed offer traces for both roles, initial offers, and derived fields. BargainingData_new.csv: Initial-experiment dataset of current work with 3,960 rows and 212 columns, including session/period fields, pie size, initial offers, deal and payoff outcomes, participant component IDs, and time-indexed offer traces. BargainingData_combined.csv: combined bargaining dataset with 10,392 rows and 213 columns, including harmonized session/period fields, role IDs, payoffs, pie size, deal status, offer traces, initial offers, squared pie value, and component IDs. EDA_original_dataset.zip: archive of exploratory analysis outputs for the original dataset, including payoff/deal-rate plots and player/state clustering summaries and figures. Turing test experimental datasets Location: Turing test experimental datasets/ ABHUBE_preparation_rounds_export.csv: preparation-round export with 90 rows and 224 columns. ABHUBE_practice_rounds_export.csv: practice-round export with 180 rows and 224 columns. ABHUBE_real_rounds_export.csv: real-round export with 885 rows and 224 columns. ABHUBE_FEEDBACK_export.csv: participant feedback export with 12 rows and 14 columns covering perceived AI/human behavioral differences and related post-experiment responses. The Turing-test round exports (experiment results) include session and round IDs, participant role IDs and host/IP fields, pie size, initial offers, deal status, payoffs, group and block metadata, whether an agent was used, participant guesses about agent/human status, and time-indexed offer traces for both roles.