Individual Differences in Statistical Learning: Conceptual and Measurement Issues

Abstract

The ability to adapt to statistical structure (often referred to as "statistical learning") has been proposed to play a major role in the acquisition and use of natural languages. Several recent studies have explored the relationship between individual differences in statistical learning and language outcomes. These studies have produced mixed results, with some studies finding a significant relationship between statistical learning and language outcomes, and others finding weak or null results. Furthermore, the few studies that have used multiple measures of statistical learning have reported that they are not correlated (e.g., [1]). The current study assesses the reliability of various measures of auditory statistical segmentation, and their consistency over time. That is, do the generally low correlations observed between measures of statistical learning stem from task demands, the psychometric properties of the measures, or the fact that statistical learning may be a highly fragmented construct? Our results confirm previous reports that individual measures of statistical learning tend not to correlate with each other, and suggest that the somewhat weak reliability of the measures may be an important factor in the low correlations. Our data also suggest that aggregating performance across tasks may be an avenue for improving the reliability of the measures.

The Necessity of the Hippocampus for Statistical Learning

“…In Experiment 1, patient L. S. J. performed at chance and significantly worse than matched healthy comparison participants. However, although comparison participants in Experiment 1 performed better than chance at the group level, they demonstrated significant variability, with a sizeable minority performing at chance levels (this variability in performance in healthy adult participants is consistent with the statistical learning literature more broadly; see (Erickson, Kaschak, Thiessen, & Berry, 2016; Arciuli, Torkildsen, Stevens, & Simpson, 2014; Misyak & Christiansen, 2012; Saffran, Newport, Aslin, Tunick, & Barrueco, 1997). Therefore, Schapiro et al implemented Experiment 2, which consisted of repeating pairs of stimuli.…”

“…This performance variability in statistical learning tasks has been noted in recent work (Erickson et al, 2016; Siegelman & Frost, 2015) and points to the importance of testing a variety of stimuli and stimulus combinations. We observed striking differences in patient performance depending on the assignment of items to triplets and pairs; notably, patient performance was poorer on the triplet assignments that had been tested with patient L. S. J.…”

“…Furthermore, Patients 1846 and 2363 also have significant reductions in hippocampal volume and severe declarative memory impairment. Although significant variability in statistical learning tasks, both within and among participants, is well documented (Erickson et al, 2016; Arciuli et al, 2014; Misyak & Christiansen, 2012; Saffran et al, 1997), it is difficult to discern why the patients in this study performed consistently better than L. S. J.…”

Statistical learning is related to early literacy-related skills

“…To our knowledge, there are no published studies that explore the extent to which different statistical learning tasks assessing the same process (e.g., the Simon task and Serial Response Tasks both require integration processes) are in fact tapping into the same underlying abilities. A key future direction for studies of this sort will therefore be to develop our understanding of the psychometric properties of the various statistical learning tasks that have been used in the literature (see Erickson, Kaschak, Thiessen, & Berry, under review, for a discussion).…”

Statistical learning abilities and their relation to language

“…To do so, the relevant sub‐components of SL should be specified, and their (non)overlapping computations should be explicitly discussed (see Arciuli, ; N. Siegelman, L. Bogaerts, & M. H. Christiansen for detailed discussions; and see Arciuli & Conway, for a detailed discussion with a focus on clinical populations). To this aim, studies of individual differences can be particularly useful, as they can directly examine the correlations between tasks that capture different SL aspects as a means of estimating the degree of overlap between SL facets (Erickson, Kaschak, Thiessen, & Berry, ; Siegelman & Frost, ).…”