Does Nonlinguistic Segmentation Predict Literacy in Second Language Education? Statistical Learning in Ivorian Primary Schools

Abstract

Statistical learning is a learning mechanism that does not directly depend on knowledge of a language but predicts language and literacy outcomes for children and adults. Research linking statistical learning and literacy has not addressed a common educational context in primary schools worldwide: children who first learn to read in their second language (L2). Several studies have linked statistical learning with childhood literacy in Australia, China, Europe, and the United States, and we preregistered an adaptation for Côte d'Ivoire, where students are educated in French and speak a local language at home. We recruited 117 sixth‐graders from primary schools in several villages and tested for correlations greater than .30 between statistical learning and literacy with 80–90% power. We found no evidence for these correlations between statistical learning and literacy, but visual statistical learning was correlated with L2 phonological awareness, a crucial emergent‐literacy skill. This finding underscores the need to include L2 acquisition contexts in literacy research.

Uncovering Sampling Biases, Advancing Inclusivity, and Rethinking Theoretical Accounts in Second Language Acquisition: Introduction to the Special Issue SLA for All?

“…This was visible in the present set of replication studies. Several research teams used or added more advanced statistical analysis techniques, most notably using linear mixed modeling techniques (Kim et al, 2023;Kremmel et al, 2023;Mifka-Profozic et al, 2023;Peters et al, 2023) and Bayesian statistics (Zinszer et al, 2023) to analyze their data sets. Sometimes, reviewers pushed for methodological changes.…”

“…Ryan and colleagues (2023) did not reproduce the positive effect that planning time was found to have on task performance by Foster and Skehan (1996). And finally, Zinszer and colleagues (2023) did not find the hypothesized effect between statistical learning and literacy that Arciuli and Simpson (2012) observed and that Qi et al. (2019) subsequently replicated; on the contrary, they adduced Bayesian evidence that there was no effect.…”

“…Zinszer et al. (2023) contributed an important data point to the statistical learning literature. Statistical learning is the ability to induce patterns of covariation from verbal or nonverbal input.…”

“…Often, multiple rounds of piloting were necessary for the experimental procedures to work as intended. A salient case in point is the research in rural Côte d'Ivoire that Zinszer and colleagues (2023) report in this special issue. Zinszer et al.…”

“…That is, they found no effects for planning time on any of their measures (complexity, accuracy, linguistic diversity, fluency) for their focal participants, Somali and Afghani former-refugee communities residing in Hamilton, New Zealand. Zinszer et al (2023) contributed an important data point to the statistical learning literature. Statistical learning is the ability to induce patterns of covariation from verbal or nonverbal input.…”