This collection consists of data from several experiments investigating two topics. Two files (conjunction_fallacy_letters.csv and conjunction_fallacy_incentivized.txt) contain data from experiments investigating the conjunction fallacy, particularly what combination rules participants are using to combine probabilities when estimating conjunctions and disjunctions. Three files (ice_cream_between.csv, ice_cream_mturk.csv, and ice_cream_shop_composition.csv) contain data from experiments investigating the dilution effect, particularly whether the effect is really due to dilution of evidence or inference of missing evidence. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% conjunction_fallacy_letters.csv Notes: Data is used in the Normative Some of the Time paper. Method overview Task "In a trial, participants were shown two sentences, such as ""People have attribute X"" and ""People have attribute y"", where the probability word was either ?rarely?, ?sometimes? or ?usually?, and each attribute was denoted by a random letter. One letter was upper case and the other letter lower case. Participants were told by the experimenter that they would be shown two statements describing the probability of people having a psychological attribute, denoted by an upper case letter, and a physical attribute, denoted by a lower case letter, which were completely independent of one another. Below the statement was the question ?How many out of 100 people have? with an ending either asking participants to estimate a single attribute, ?X??/?Y??, a conjunction ?X and y??, or a disjunction, ?X or y??." Independent variables Probability word for the uppercase letter: rarely, sometimes, or usually Probability word for the lowercase letter: rarely, sometimes, or usually Question: Conjunction, Disjunction, Left segment, or Right segment. Dependent variables Response t_sliderClick t_sliderToEnd t_trialTotal Column Details Participant Randomly generated participant number Trial_Idx Trial number Block Block number (from 0) Block_Idx Trial number within block La Letter identity of the uppercase letter Lb Letter identity of the lowercase letter Pa "Probability of upercase letter 0 = Rarely 1 = Sometimes 2 = Usually " Pb "Probability of lowercase letter 0 = Rarely 1 = Sometimes 2 = Usually " Question "Question shown to participants 0 = Conjunction 1 = Disjunction 2 = Left pie chart 3 = Right pie chart" Response Participant response. The raw slider values are here and participants saw rounded integers between 0 and 100. t_sliderClick Time in ms from the participant starting the trial to clicking on the slider (from browser). t_sliderToEnd Time in ms from the participant clicking on the slider to clicking on the submit button (from browser). t_trialTotal Time in ms from the start of the trial ? when the DOM elements are created ? to the participant clicking the submit button. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% conjunction_fallacy_incentivized.csv Notes: Control condition data is used in the Normative Some of the Time paper. Other conditions will be used in a follow up publication. Method overview Task In a trial, participants were shown two pie charts (described as spinners to the participant) with one segment of each occupying either 14%, 44% or 84% of the pie. These segments were the segments of interest. Participants were asked to estimate how many times the arrow of a spinner would land on either one, both or either one or the other segment of interest. Independent variables Incentive scheme: Control, Linear or Precise (see below). Size of left segment of interest: 15%, 44%, or 84%. Size of right segment of interest: 15%, 44%, or 84%. Question: Conjunction, Disjunction, Left segment, or Right segment. Dependent variables Response t_sliderToEnd t_trialTotal trialPoints pointsSoFar Column Details Participant Randomly generated participant number Trial_Idx Trial number Block Block number (from 0) Block_Idx Trial number within block color1 First color on colored spinner color2 Second color on colored spinner shape1 First shape on shape spinner shape2 Second shape on shape spinner Pa "Probability of upercase letter 0 = Rarely 1 = Sometimes 2 = Usually " Pb "Probability of lowercase letter 0 = Rarely 1 = Sometimes 2 = Usually " Question "Question shown to participants 0 = Conjunction 1 = Disjunction 2 = Left pie chart 3 = Right pie chart" Response Participant response. The raw slider values are here and participants saw rounded integers between 0 and 100. t_sliderClick Time in ms from the participant starting the trial to clicking on the slider (from browser). t_sliderToEnd Time in ms from the participant clicking on the slider to clicking on the submit button (from browser). t_trialTotal Time in ms from the start of the trial ? when the DOM elements are created ? to the participant clicking the submit button. condition "Incentive scheme condition 0 = Control ? No incentive 1 = Linear ? Payment equals abs (correctAnswer - response) 2 = Precise ? Payment if within 10 of correctAnswer" correctAnswer The correct response where rarely = 16%, sometimes = 44% and usually = 84% trialPoints Points accumulated on that trial pointsSoFar Cumulative total of points for a participant %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ice_cream_between.csv Method overview Task In a trial, participants were told that their task was to figure out whether a person has been to ice cream Shop A or Shop B based on the ice cream cone they are holding. They were shown two "shops": boxes on the screen that contained the ice creams and cones that were sold that day. In this experiment, the ice creams and cones were always separated from one another. Participants were then shown a test stimulus, the ice cream or cone or both that a particular person was holding. Based on this stimulus, participants were asked to judge the relative likelihood of the stimulus coming from each of the two shops. Independent variables Condition: whether the single items were presented first or the combined items were presented first Type: the type of item that was presented to judge Informative feature: whether the cone, ice cream, both or neither provided evidence as to the shop of origin Shop: which shop was the correct shop Dependent variables Score Column Details subj Randomly generated participant number trial Trial number for that participant condition "Order in which stimuli were observed DS = All of the trials where there are both ice creams and cones (inf_anti, inf_inf, inf_non, and non_non) appear before trials where there is an ice cream or cone alone (inf_miss) SD = All of the trials where there is an ice cream or cone alone (inf_miss) appear before trials where there are both ice creams and cones (inf_anti, inf_inf, inf_non, and non_non) SandD = All of the trial types are randomly intermixed" shop The identity of the correct shop score Response of participant on a likelihood ratio scale (e.g., if Shop A has 6 copies of the stimulus and Shop B has 2, then the correct answer is 3:1 in favor of Shop A). Positive numbers are answers that are in the same direction as the normative response, while negative responses are anti-normative. 10.1 and -10.1 were the extreme values. informative.feature "Both = Both ice cream and cone are informative about which shop the stimulus came from Ice cream = Only the ice cream is informative about which shop the stimulus came from Cone = Only the cone is informative about which shop the stimulus came from Neither = Neither cone nor ice cream were informative about which shop the stimulus came from" type "Type of trial inf_anti = Both halves of the ice cream cone are informative, but they point toward different shops inf_inf = Both halves of the ice cream cone are informative, and they point toward the same shop inf_non = One half is informative, the other is non-informative inf_miss = One half is informative, the other is missing non_non = Both halves of the stimulus are non-informative " normative Correct response in terms of likelihood ratio %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ice_cream_mturk.csv Method overview Task In a trial, participants were told that their task was to figure out whether a person has been to ice cream Shop A or Shop B based on the ice cream cone they are holding. They were shown two "shops": boxes on the screen that contained the ice creams and cones that were sold that day. In this experiment, the ice creams and cones could be combined or separated from one another depending on the between-participants variable "condition". Participants were then shown a test stimulus, the ice cream or cone or both that a particular person was holding. Based on this stimulus, participants were asked to judge the relative likelihood of the stimulus coming from each of the two shops. Independent variables Condition: how the shops were presented Type: the type of item that was presented to judge Informative feature: whether the cone, ice cream, both or neither provided evidence as to the shop of origin Shop: which shop was the correct shop Dependent variables Score Column Details subj Randomly generated participant number trial Trial number for that participant condition "How the shops were presented 211_joint = Ice creams and cones were presented combined (i.e., joint), the informative features always provided 2:1 evidence, and non-informative features were included 311_joint = Ice creams and cones were presented combined (i.e., joint), the informative features always provided 3:1 evidence, and non-informative features were included 210_joint = Ice creams and cones were presented combined (i.e., joint), the informative features always provided 2:1 evidence, and non-informative features were not included 211_separate = Ice creams and cones were presented separately, the informative features always provided 2:1 evidence, and non-informative features were included" shop The identity of the correct shop score Response of participant on a likelihood ratio scale (e.g., if Shop A has 6 copies of the stimulus and Shop B has 2, then the correct answer is 3:1 in favor of Shop A). Positive numbers are answers that are in the same direction as the normative response, while negative responses are anti-normative. 10.1 and -10.1 were the extreme values. informative.feature "Both = Both ice cream and cone are informative about which shop the stimulus came from Ice cream = Only the ice cream is informative about which shop the stimulus came from Cone = Only the cone is informative about which shop the stimulus came from Neither = Neither cone nor ice cream were informative about which shop the stimulus came from" type "Type of trial inf_anti = Both halves of the ice cream cone are informative, but they point toward different shops inf_inf = Both halves of the ice cream cone are informative, and they point toward the same shop inf_non = One half is informative, the other is non-informative inf_miss = One half is informative, the other is missing non_non = Both halves of the stimulus are non-informative " normative Correct response in terms of likelihood ratio %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ice_cream_shop_composition.csv Method overview Task In a trial, participants were told that their task was to figure out whether a person has been to ice cream Shop A or Shop B based on the ice cream cone they are holding. They were shown two "shops": boxes on the screen that contained the ice creams and cones that were sold that day. In this experiment, the ice creams and cones were always separated from one another. Participants were then shown a test stimulus, the ice cream or cone or both that a particular person was holding. Based on this stimulus, participants were asked to judge the relative likelihood of the stimulus coming from each of the two shops. Independent variables Condition: the composition of the shops, which determines the value of the evidence of each ice cream or cone Type: the type of item that was presented to judge Informative feature: whether the cone, ice cream, both or neither provided evidence as to the shop of origin Shop: which shop was the correct shop Dependent variables Score Column Details subj Randomly generated participant number trial Trial number for that participant condition "Composition of the shops Regular = Ice cream and cones can be informative or non-informative, informative ice creams and cones provide 2:1 evidence Inf = All of the ice creams and cones are informative (2:1 evidence), the shop composition means there are no non-informative stimuli Non = One or the ice creams or cones is either informative (2:1 evidence) or non-informative, the other is always non-informative" shop The identity of the correct shop score Response of participant on a likelihood ratio scale (e.g., if Shop A has 6 copies of the stimulus and Shop B has 2, then the correct answer is 3:1 in favor of Shop A). Positive numbers are answers that are in the same direction as the normative response, while negative responses are anti-normative. 10.1 and -10.1 were the extreme values. informative.feature "Both = Both ice cream and cone are informative about which shop the stimulus came from Ice cream = Only the ice cream is informative about which shop the stimulus came from Cone = Only the cone is informative about which shop the stimulus came from Neither = Neither cone nor ice cream were informative about which shop the stimulus came from" type "Type of trial inf_anti = Both halves of the ice cream cone are informative, but they point toward different shops inf_inf = Both halves of the ice cream cone are informative, and they point toward the same shop inf_non = One half is informative, the other is non-informative inf_miss = One half is informative, the other is missing non_non = Both halves of the stimulus are non-informative " normative Correct response in terms of likelihood ratio