Participants first learned a set of syllables arranged in either a single row (one-dimensional condition) or a grid (two-dimensional condition), then produced two random sequences for the same display. These data are licensed under CC BY 4.0, reproduced from materials in OSF.
participant id
participant's gender (self-reported)
participant's age (self-reported)
position of the item in the sequence, 0 indexed
unique identifier for the participant
whether the item belongs to the first sequence the participant uttered (A) or the second (B)
syllable uttered
timestamp of when the utterance starts, in seconds.
temporal difference with the start of the previous item (i.e. starts[index] - starts[index - 1])
whether the participant was allocated to the one-dimensional or two-dimensional condition
Which of five possible configurations the participant learned
The position of the syllable in the array. For 1D arrays, position is left to right. For 2D arrays positions 1-2 correspond to the top 2 cells; 3-5 to the middle 3 cells; and 6-7 to the bottom three cells (always left to right)
whether the item is a repetition of the last
whether the item is adjacent to the last in the display (after removing repetitions)
whether the item is a turning point, considering all items (after removing repetitions)
the Euclidean distance to the previous item (after removing repetitions)
a measure of how likely the item is in a uniform or gaussian distribution (see text)
the expectation for measure * derived from reshuffling the participant's sequence 10000 times
castillo2024.rgmomentum.e2An object of class data.frame with 28483 rows and 20 columns.
Castillo L, León-Villagrá P, Chater N, Sanborn AN (2024). “Explaining the Flaws in Human Random Generation as Local Sampling with Momentum.” PLOS Computational Biology, 20(1), 1--24. doi:10.1371/journal.pcbi.1011739 .