Fig 1.
Static-cling stickers used in the experiment, along with verbally presented questions. Children were asked a series of questions about each bear–first in one order (e.g., choices in blue on the left) and later in the opposite order (e.g., choices in red on the right). Children were only shown the stickers after they responded to the question verbally.
Fig 2.
Robust recency bias across ages in Experiment 1.
Each dot represents an individual’s proportion of second choices during the trials in which they responded verbally. The error bars depict 95% confidence intervals. The dotted line represents chance. Twenty-two of the twenty-four participants exhibited significantly more second-choice responses than would be predicted by chance.
Table 1.
Experiment 1 generalized linear mixed effect model results.
Fig 3.
Recency bias present in natural parent-child setting.
(A) Proportion of second-choice responses in CHILDES binned by age of the children in years. (B) A smoothed LOESS curve fitted to the raw data from CHILDES (empty dots).
Fig 4.
Examples of novel toy pictures used in the experiment, along with the verbal prompt. Children were asked to name a series of 20 toys (e.g., choices in the blue on the left). Later they were shown the same toys and asked to name them again in the opposite order (e.g., choices in the red on the right). Names varied in syllable length.
Fig 5.
Recency bias present in preschoolers.
Each dot represents an individual’s proportion of second choices. The error bars depict 95% confidence intervals. The dotted line represents chance. Fourteen participants exhibited significantly more second-choice responses than chance would predict. Three participants exhibited significantly less second-choice responses than chance would predict.
Fig 6.
Bias is stronger when choice options have more syllables.
Proportion of second choices made separated by word length. Dotted line indicates chance. For words two syllables or more, the proportion of second choice selections is significantly above chance.
Table 2.
Experiment 2 generalized linear mixed effect model.