Data Code Book

Authors
Affiliations

Elias Bouacida

University Paris 8

Renaud Foucart

Lancaster University

Published

October 7, 2024

We provide here the description of all the raw data created in the experiments.

The raw output from the experiment are not shared, as they contain non-anonymous Prolific or AMT identifiers. Once anonymized, files from all the experiments are called ExperimentN_XTYPEOFDATA.csv. N is the Experiment number in the paper. 1 therefore corresponds to the experiments run on AMT, while 2 are the experiments run on Prolific. X gives the different waves on AMT. We also have data files called Feedback_YYYY-MM-DD.csv and PagesTImes-YYYY-MM-DD.csv which contains additional data exported from oTree.

Note:

Folder OriginalData

ExperimentN[|_X][RawData|Comments].csv and Feedback_YYYY-MM-DD.csv

Session Variables

Session variables have the same value for all the subjects in one session.

Variables Description Kind of values Notes
session_code Unique identifier for the session string of 8 characters There may be several session_code per experiment, as it corresponds to the session as launched in oTree, and not necessarily how we paid subjects.
session_label Name given to the session empty Not used in this experiment
session_comment comment on the session empty Not used in this experiment
session_is_demo Is the session a test one? Boolean Not used in this experiment
session_config_criteria Configuration particularity of the session empty Not used in this experiment
session_config_real_world_currency_per_point Conversion rate between experimental currency unit and money integer Value is always 1, there was no point in the experiment.
session_config_participation_fee Value of the participation fee in the experiment (in $) integer 0.8 in AMT (Experiment 1) and 1 in Prolific (Experiment 2) experiments
session.lottery_time Which DC-5 lottery was used in the experiment Time and Date Each DC-5 lottery can be uniquely identified by the time and date it happened.
session.result_time When will the results be available to subjects? Time and Date Typically a few days after the experiment is run.

Participant Variables

Participant variables may differ between different subjects in the experiment but remain constant in all the experiment. When in the file starting with Feedback, the variables do not have the participant moniker.* In the ProlificRawData.csv, they do. It is the same variable.

Variables Description Kind of values Notes
participant_id_in_session Unique identifier for the participant in the session Integer
participant_code Unique participant identifier, in theory by session, in practice over all the experiment Random string of 8 letters and numbers Used by subjects to see their results on the website
participant_is_bot Is the participant a computer or not? Boolean Not used in this experiment.
participant_index_in_pages Which page number the row refers to in the experiment integer
participant_max_page_index The last page subjects reached in the experiment integer
participant_current_app_name Name of the App the subject is currently in for this row of data string
participant_current_page_name Name of the page the subject is currently in for this row of data string
participant_time_started_utc The date and time the subject entered oTree (which is slightly later than the AMT/Prolific entry time) Time and data value following ISO conventions
participant_visited Has the participant entered otree? Boolean Value is 1.
participant_payoff Payment computed in oTree (in $) integer Value is 0.1 or 0 in the AMT experiments, 0 in the Prolific one: the only payoff computed by oTree directly was the ambiguity aversion elicitation.
participant.criteria Which procedure was the non-lottery one string Value is Guessing the Paintings, Rock, Paper, Scissors, Arrival Time or Time
participant.criteria_first Was the non-lottery shown first? Boolean if false, the lottery was shown first
participant.criteria_choices What choices subjects made in the non-lottery? string that is a vector of figures Repeat the value of Algorithms_1_player_criteria_choiceX in order.
participant.lottery_choices What choices subjects made in the lottery? string that is a vector of boolean Repeat the value of Algorithms_1_player_lottery_choiceX in order.
participant.criteria_control Had subjects control on the non-loterry? Boolean Non-control on the non-lottery was only used in Experiment 1 C. The variable only exists for Experiment 1_C and 2, but is not used in the latter case.
participant.lottery_control or participant.control Had subjects control on the lottery? Boolean lottery_control replaced control in Experiment 1_C.
participant.mechanism Which procedure subjects chose to allocate the reward? string The string is the name of the procedure
participant.arrival_code Code created for subjects based on their arrival time figure with 5 digits Used for the Arrival time procedure, that is not in the paper, and to break ties if needed
participant.arrival_time When did the subject enter the experiment Time Used for the Arrival time procedure, may be redundant with participant_time_started_utc
participant.best_mechanism Did the subject think that they are more likely to win in the lottery or the non-lottery? Boolean Only elicited in round 2 of the Belief app, so mirrors Beliefs_2_player_best_mechanism
participant.number Number given to the participant in the Prolific experiment integer between 0 and 99 Used to group together subjects, as if they were chosen to implement the procedure, it only applied to a group of 100 participants.

“App” variables

Variables that may vary by page or block in the experiment and differ across subjects. In the following table, APP can take different values (Algorithms, Mechanisms, Beliefs, Questionnaire, Feedback and so on). n can take value 1 or 2. n is the round number if there are more than 1.

Variables Description Kind of values Notes
APP_n_player_id_in_group Identifier for the subject in the group for the given APP. integer Not used in this experiment
APP_n_player_role If we had groups, what role would the subject have in the group empty Not used in this experiment.
APP_n_player_payoff Payment received in this APP integer (in $) 0 for most App, except when ambiguity aversion is elicited.
Algorithms_1_player_criteria_choiceX Actions in the non lottery. integer The actions X are chosen by the subject in the treatments with control and by oTree without control. X goes from 1 to 5, in order.
Algorithms_1_player_hours The “hour” used to create the code in the Time algorithm integer
Algorithms_1_player_minutes The “minute” used to create the code in the Time algorithm integer
Algorithms_1_player_seconds The “seconds” used to create the code in the Time algorithm integer
Algorithms_1_player_lottery_choiceX Actions in the lottery. Boolean The actions X are chosen by the subject in the treatments with control and by oTree without control. X goes from 1 to 5, in order.
APP_n_group_id_in_subsession Identifier for the group in this APP. empty Not used in this experiment
Algorithms_n_subsession_round_number Round number in the APP. integer
Mechanism_n_player_mechanism Which procedure the subject chose? Boolean Each value correspond to the lottery or the non-lottery.
Mechanism.n.player.wta What is the given WTA of the subjet to change his choice? number Value between $0 and $2.
Beliefs_n_player_belief Did the subject think they will win or loose for the procedure given in round n integer (value of -1 or 1) -1: loosing, 1: winning
Beliefs_n_player_criteria What procedure was elicited in round n Boolean Lottery if false, non lottery otherwise.
Beliefs_n_player_best_mechanism Did the subject think that they are more likely to win in the lottery or the non-lottery? Boolean Only elicited in round 2 of the Belief app, so is empty for n=1.
Questionnaire_1_player_age Answer to the MCQ “How old are you?” Values are “<25”, “25-40”, “40-55”, “>55”, “Prefer not to say” Only in Experiment 1.
Questionnaire_1_player_gender_rev What the gender question order reversed or not between Male and Female Boolean 1 means Female first, 0 Male first, Only in Experiment 1.
Questionnaire_1_player_gender Answer to the MCQ “What is your gender?” Values are “Male”, “Female”, “Prefer not to say”, “Other” Only in Experiment 1.
Questionnaire_1_player_employment Answer to the MCQ “What is your occupation?” Values are “Employed”, “Self-employed”, “Student”, “Retired”, “Other” Only in Experiment 1.
Questionnaire_1_player_region “What is your country of residence?” String Only in Experiment 1.
Questionnaire_1_player_urn_rev Not used in the experiment. Boolean
Questionnaire_1_player_urn What urn did the subject choose? Values are “Urn Left” or “Urn Right” The left one is the risky one, the right one the ambiguous one
Questionnaire_1_player_colour_rev Not used in the experiment. Boolean It was supposed to invert Red and Black colour
Questionnaire_1_player_colour Which colour did the subject choose? “Red” or “Black”
Questionnaire_1_player_colour_drawn What colour did oTree draw when computing the ambiguity payments? “Red” or “Black”
Questionnaire_1_player_urn_winner Did the subject win the in the ambiguity elicitation task? Boolean If 1, the subject earned $0.10
Questionnaire.n.player.probability Choice between two statements “I think I have more chances of winning in this mechanism.” and “I prefer this mechanism, but for other reasons.” Boolean First statement is true, second is false.
Questionnaire_1_player_reasons Answer to the question “Why did you choose the CHOSEN PROCEDURE to get your reward?” String CHOSEN PROCEDURE was replaced by the procedure they chose in the experiment.
Questionnaire.1.player.merit Likert scale answer to the question “PROCEDURE 1 is more meritocratic than PROCEDURE 2” Likert scale values (integer) Only in Experiment 2.
Questionnaire.1.player.fair Likert scale answer to the question “PROCEDURE 1 is fairer than PROCEDURE 2” Likert scale values (integer) Only in Experiment 2. Procedure 1 and 2 varied between participants.
Questionnaire.1.player.easy Likert scale answer to the question “PROCEDURE 1 is easier to understand than PROCEDURE 2” Likert scale values (integer) Only in Experiment 2
Questionnaire.1.player.procedureX Which procedure is the procedure X string Values are Lottery, Time or RPS
unique_id Unique identifier (by session) created during the anonymization proces. integer Used in particular to make sure that an AMT participants did not participate more than once.
Feedback_n_player_completion_code Completion code given to subjects to fill in AMT. Used to check double participation. string Random sequence of 10 digits and letters.

Prolific added variables

As a reminder, the Prolific experiment is Experiment 2.

Variables Description Kind of values Notes
Started.at When did the subject start the experiment on Prolific Date and Time
Completed.at When did the subject finish the experiment on Prolific (i.e. entered the completion code) Date and Time
Reviewed.at When did the experimenter checked the participation in the experiment on Prolific Date and Time Not used in this experiment.
Archived.at When was the participation of the subject in the experiment archived on Prolific (i.e. entered the completion code) Date and Time Not used in this experiment.
Time.taken How long did it take to complete the experiment integer (seconds) May differ from the time completed on oTree, as it starts and finishes on Prolific
Completion.code What completion did subjects entered string It should have been CG1DN21H, but sometimes subjects used their uniqueid, which is fine as well. Used for the payments, but not in the analysis.
Total.approvals How many approvals did the subject received before on Prolific integer Not used in this experiment.
Fluent.languages What language do subject speak fluently string Self-declared to Prolific. Not used in this experiment.
Age How old is the subject integer Self-declared to Prolific
Sex What gender? string Self-declared to Prolific
Ethnicity.simplified What ethnicity does the subject decare string Self-declared to Prolific
Country.of.birth What country is the subject born in? string Self-declared to Prolific
Country.of.residence What country is the subject living in? string Self-declared to Prolific and is USA.
Nationality What is the nationality of the subject? string Self-declared to Prolific. Not used in this experiment.
Language What is the subject main language ? string Self-declared to Prolific. Not used in this experiment.
Student.status Is the subject a student? string Self-declared to Prolific.
Employment.status What is the subject’s employment status? string Self-declared to Prolific.

PageTimes-YYYY-MM-DD.csv files

These files allow us to compute the time spend on each page. Variables already introduced earlier are not repeated.

Variables Description Kind of values Notes
page_index Page number in the experiment integer Starts at 0
app_name The name of the App string Empty for the initial page, when subjets arrives in the experiment
page_name Name given to the page in the experiment string The name is given in the oTree code.
epoch_time_completed The time when the page is left integer Starts at the conventional epoch time for computer (January 1, 1970, midnight UTC).
round_number Which round is it? integer At most 2
timeout_happened Did the subject go to the next page because of a time limit? Bool If 0, then no timeout. In general there should be not timeout
is_wait_page Is the page a waiting page (i.e., a page where the subject is waiting for other subjects or an event) Bool 0 for everyone. Not used in this experiment.

ExperimentN_XComments.csv

In these files, we classified the literal comments provided by subjects as indicating that probability or preferences for a given procedure seem to matter more. We did the classification ourselves, and it is obviously subject to our own subjective interpretation of the commentary for subjects. We could not always classify the comments, so some subjects may belong to no or both classes. The participant_code is the same as in the other files.

Variables Description Kind of values Notes
probability Does the subject state that probability matters more? Bool
preference Does the subject state that the prefer one procedure over the other, for a specific reason? Bool
error Do we think that stated preferences and behavior coincide? Bool True means that we think there is an inconsistency

Rock_Paper_Scissors_Raw_data.csv

Data is coming from players playing on . We thank Lasse Hassing for giving us access to their data.

Variables Description Kind of values Notes
game_id Unique identifier for a game integer
game_round_id Unique identifier for the round in the game integer The identifier is unique at the round and game level. Players stay in the same round when they use the same strat.
player_one_throw What did player 1 choose? integer 1 means Rock, 2 Paper and 3 Scissors.
player_two_throw What did player 2 choose? integer 1 means Rock, 2 Paper and 3 Scissors.

Folder Input

In this folder, the data has been cleaned an pre-treated. Some variables are repeated from before, others are new. The treatment has been done by files ExperimentN[|_X]CleaningData.ipynb Experiment1CleanedData.csv merges together all the data cleaning from the AMT experiment. Its counterpart for the Prolific experiment is Experiment2CleanedData.csv

Experiment[N|N_X]CleanedData.csv

Variables Description Kind of values Notes
[participant_]code Unique participant identifier, in theory by session, in practice over all the experiment Random string of 8 letters and numbers Used by subjects to see their results on the website. Not used in the data analysis.
[participant_]time_started_utc The date and time the subject entered oTree (which is slightly later than the AMT/Prolific entry time) Time and data value following ISO conventions
criteria_choiceX What X action did the subject choose for the non-lottery. In the treatment without control, oTree gave the choice to subjects integer X goes from 1 to 5, in order.
lottery_choiceX What X action did the subject choose for the lottery. In the treatment without control, oTree gave the choice to subjects Boolean X goes from 1 to 5, in order. Value of 0 means odd, 1 means even.
age or Age How old is the subject bracket or integer Values are “<25”, “25-40”, “40-55”, “>55”, “Prefer not to say” in AMT a integer value in Prolific
employment Answer to the MCQ “What is your occupation?” Values are “Employed”, “Self-employed”, “Student”, “Retired”, “Other” Not asked in Prolific
Employment_status What is the subject’s employment status? string Self-declared to Prolific.
Student_status Is the subject a student? string Self-declared to Prolific.
Ethnicity.simplified What ethnicity does the subject decare string Self-declared to Prolific
colour Which colour did the subject choose? “Red” or “Black” Not asked in Prolific
Questionnaire_1_player_colour_drawn What colour did oTree draw when computing the ambiguity payments? “Red” or “Black” Not asked in Prolific
urn_winner Did the subject win the in the ambiguity elicitation? Boolean If 1, the subject earned $0.10, only AMT
reasons Answer to the question “Why did you choose the CHOSEN PROCEDURE to get your reward?” String CHOSEN PROCEDURE was replaced by the procedure they chose in the experiment
unique_id Unique identifier (by session) created during the anonymization proces. integer Used in particular to make sure that an AMT participants did not participate more than once.
time_in_experiment Total time spent between landing on the Welcome Page and arriving to the last page of the experiment.
time_page_X Total time spent on page X of the experiment integer The length (in pages) of the AMT and Prolific experiment differed.
criteria Which procedure was the non-lottery one string Value is Guessing the Paintings, Rock, Paper, Scissors, Arrival Time or Time
criteria_first Was the non-lottery shown first? Boolean if false, the lottery was shown first
criteria_choices What choices subjects made in the non-lottery? string that is a vector of figures Repeat the value of Algorithms_1_player_criteria_choiceX in order.
lottery_choices What choices subjects made in the lottery? string that is a vector of boolean Repeat the value of Algorithms_1_player_lottery_choiceX in order.
criteria_control Had subjects control on the non-loterry? Boolean Non-control on the non-lottery was only introduced in Experiment 1_C
lottery_control or control Had subjects control on the lottery? Boolean
mechanism Which procedure subjects chose to allocate the reward? string The string is the name of the procedure
arrival_code Code created for subjects based on their arrival time figure with 5 digits Used for the Arrival time procedure, that is not in the paper, and to break ties if needed
arrival_codeX Xth number in the code created for subjects based on their arrival time integer Used for the Arrival time procedure, that is not in the paper, and to break ties if needed. Together forms the arrival_code variable.
best_mechanism Did the subject think that they are more likely to win in the lottery or the non-lottery? Boolean Only elicited in round 2 of the Belief app, so mirrors Beliefs_2_player_best_mechanism
criteria_belief Did the subject think they will win or loose in the non-lottery? Boolean If true, the subject stated that they would win in that procedure.
lottery_belief Did the subject think they will win or loose in the lottery? Boolean If true, the subject stated that they would win in the lottery.
female Is the subject declaring she is a female? Boolean
male Is the subject declaring he is a male? Boolean
other Is the subject declaring nothing regarding their gender? Boolean
ambiguity_averse Did the subject choose the non-ambiguous urn? Boolean The measure does not elicit ambiguity neutral subjects, who could choose any urn (they are indifferent).
treatment To which treatment does the subject belong? string Give the non-lottery procedure and the control given or not on the both the lottery and non-lottery one.
country Normalized country names from their previous answers string
prediction_payed Which belief elicitation is used for payment? string Three possible values: lottery, criteria or best_mechanism
best_mechanism_winner Did subjects correctly guessed their relative ranking between the lottery and non-lottery procedure Boolean
payment How much did subjects earn in the experiment (excluding show-up fee) number
lottery_winner Is the subject winning in the lottery procedure? Boolean
lottery_ranks What is the rank of the subject in the lottery procedure (in their group)? number Given between 0 and 1 (in percentage), to normalize because not all groups where exactly made of 100 subjects.
lottery_score How many odd/even guess were correct? integer
lottery_belief_winner Is the subject correctly guessing if they won or not in the lottery? Boolean
criteria_winner Is the subject winning in the non-lottery procedure? Boolean
criteria_ranks What is the rank of the subject in their non-lottery procedure (in their group)? number Given between 0 and 1 (in percentage), to normalize because not all groups where exactly made of 100 subjects.
criteria_score Score in the non-lottery procedure. integer The way the score is computed depends on the procedure. If there was an objective answer, it is the number of correct guesses (guessing the paintings), the number of wins (RPS) or the number of evens (Time).
criteria_belief_winner Is the subject correctly guessing if they won or not in the lottery? Boolean
probability Does the subject state that probability matters more? Bool
preference Does the subject state that the prefer one procedure over the other, for a specific reason? Bool
error Do we think that stated preferences and behavior coincide?Not used in the data analysis. Bool True means that we think there is an inconsistency
session To which experiment did subjects belong integer values are 1, 2 or 3
number Number given to the subject in their group integer between 0 and 99 Used to determine which participant in the group is used for payment.
wta What is the given WTA of the subjet to change his choice? number Value between $0 and $2.
bonus How much subjects received as bonus integer Always be 0, it was directly in their payments.
group In which group of participants the subject is integer Values go from 1 to 6 (or NA). All subjects in one group have the same procedure used to compute payments. Only used in Experiment 2.
expe In which treatment were subjects integer 1 for RPS, 2 for Time. Only in Experiment 2.