This data set contains suboptimal choice data from multiple published studies. The data was compiled by Dunn et al. (2024).
Format
A data frame with 330 rows and 44 columns:
- row
Row index.
- study
Short identifier for the study from which the data and parameters were extracted.
- year
Year the study was published.
- species
Species tested in the experiment (e.g., pigeon or starling).
- strain
Strain of rat used.
- sex
Sex of subjects (M = Male, F = Female, MF = Males and Females).
- sex_MF_ratio
Ratio of Males to Females.
- age
Age of subjects.
- exp
Experiment number as reported in the original study. Values joined by an '&' indicate that data were aggregated across experiments.
- condition
Short label used to identify conditions within each study, aiding in dataset assembly.
- n
Number of subjects in the condition.
- cp
Observed choice proportion reported by the study.
- il_dur_a, il_dur_b
Average programmed initial link (IL) durations for alternatives A and B, respectively. FR1 schedules are assigned a duration of 1; other fixed-ratio schedules are given a duration of \(1 + (\text{ratio}/\text{run rate})\).
- tl_dur_a1, tl_dur_a2, tl_dur_b1, tl_dur_b2
Average programmed durations of terminal links. For example,
tl_dur_a1corresponds to the first terminal link on alternative A.- tl_p_a1, tl_p_a2, tl_p_b1, tl_p_b2
Probability of entering each respective terminal link.
- tr_p_a1, tr_p_a2, tr_p_b1, tr_p_b2
Probability of terminal (i.e., primary) reinforcement following entry into each respective terminal link.
- il_sched_a, il_sched_b
Schedule type used for the initial links (e.g., VI, FR).
- tl_sched_a1, tl_sched_a2, tl_sched_b1, tl_sched_b2
Schedule type used for each terminal link.
- operant_response
Description of the behavioural response required to make a choice and thus obtain terminal reinforcement.
- forced_exposure
Ordinal variable describing the amount of forced exposure (FE) trials used in the study. See
Details.- deprive_lvl
Deprivation level for the subjects in each condition, expressed as a percentage of their free-feeding body weight.
- il_stim
Description of initial link stimulus used.
- tl_s_splus_stim
Description of S+ terminal link stimulus used.
- tl_s_minus_stim
Description of S- terminal link stimulus used.
- tl_unsig_stim
Description of the unsignalled (optimal) alternative's terminal link stimulus used.
- term_rein
Description of the terminal (i.e., primary or unconditioned) reinforcer used in the study (e.g., piece of cereal; images of female models).
- term_rein_cat
Broad category of the terminal reinforcer (e.g., food, sexualized imagery), providing a more general classification of
term_rein.- DOI
Digital Object Identifier (DOI) for the original study.
- ref
Full bibliographic reference for the study.
- data_version
A character column specifying the dataset version based on the date it was last manually curated or corrected (e.g., "2025-07-20"). Used for tracking updates to the compiled data.
Source
Dunn, R. M., Pisklak, J. M., McDevitt, M. A., & Spetch, M. L. (2024). Suboptimal choice: A review and quantification of the signal for good news (SiGN) model. Psychological Review. 131(1), 58-78. https://doi.org/10.1037/rev0000416
Dunn, R. M., Pisklak, J. M., McDevitt, M. A., & Spetch, M. L. (2024). The signals for good news (sign) model data and code repository. Open Science Framework. https://doi.org/10.17605/OSF.IO/39QTJ
Details
This data set reflects a merged and corrected version of the three separate data files originally provided by Dunn et al. (2024) via the Open Science Framework (OSF). The original files supplied to the OSF contained some numeric values that were inadvertently over-rounded. These values have been corrected in this version to restore their intended precision.
No FE: Did not include any FE trials; Only 1 FE session at the start of the condition; FE trials for only some subjects.
Few FE: 4 - 20 FE trials at the start of each session; 6 - 12 sessions of FE at the start of the first condition in the experiment; 2 sessions at the start of each condition.
Lots of FE: FE trial following every suboptimal blackout; At least 50% of all trials were FE.