The role of feedback and dot presentation format in younger and older adults’ number estimation

Abstract

Numerosity estimation, the rapid assessment of the number of items in a visual scene, is historically inaccurate. We assessed whether providing feedback regarding the correct numerosity on either 0%, 50%, or 100% of the trials would affect younger and older adults' estimation accuracy for randomized, clustered (i.e., groups of 3 or 7 dots), and stacked (i.e., column) dot formats. Participants provided estimates and confidence ratings in six blocks, each containing 48 trials (16 numerosities shown in each format). Feedback frequency was manipulated between participants during blocks 1-4; no feedback was provided during blocks 5 and 6, which contained old and new numerosities and previously estimated presentations rotated 90°. Estimation accuracy was age equivalent across blocks despite younger adults initially being more accurate than older adults. Feedback improved both age groups' accuracy. Stacked presentations were most accurately estimated but were more likely to be over-estimated than clustered and randomized presentations. Older adults gave lower confidence ratings than younger adults despite both age groups showing increased confidence across blocks, for more structured presentation formats, and as feedback frequency increased. These results expand our understanding of the role of presentation format and feedback in producing age equivalence or age-related differences in numerosity estimation.

If so many are “few,” how few are “many”?

“…They suggest that contextual influence such as explicit reinforcement can be a driving force to shift the quantifier's scope (cf. Price et al, 2014 for an extensive discussion of feedback mechanisms during the estimation of quantities). Moreover, there is a change in the reference of the related quantifier “few,” even though this was not explicitly mentioned during the adaptation phase.…”

Numerical Building Blocks: Exploring Domain-Specific Cognitive Predictors of Mathematics

“…It is uncommon to see identification tasks in the large range or comparison tasks in the subitizable range. Some research has included identification of large quantities (e.g., Price, Clement & Wright, 2014), however performance on this task was not examined in relation to other ANS tasks, quantitative abilities, or math outcomes. Nonsymbolic comparison tasks in the 1 to 9 range are not uncommon, but rather than deriving separate subitizing and counting measures from this task, they are typically scored overall and labeled an index of the ANS.…”