Bogaerts, L. 1 , Siegelman, N. 2 , Benporat, T. 2 , Szmalec, A. 3 , Duyck, W. 1 & Frost, R. 2, 4, 5
1 Ghent University, Ghent, Belgium
2 Hebrew University, Jerusalem, Israel
3 Université Catholique de Louvain, Louvain-La-Neuve, Belgium
4 Haskins Laboratories, New Haven, USA
5 The Basque Center for Brain and Language, San Sebastian, Spain
The Hebb repetition task (HRT), an operationalization of long-term sequence learning, is hypothesized to provide a laboratory analogue for naturalistic vocabulary acquisition (Page & Norris, 2009). Recent studies have suggested that performance in the HRT is indeed related to linguistic (dis)abilities. However, in spite of the growing interest in sequence learning capacity as measured trough the HRT, no previous research has tested whether it provides stable and reliable measures of individual performance. In the present study we tested how the HRT relates to visual statistical learning and examined the reliability of verbal Hebb learning as an individual capacity. Our participants, Hebrew-English speakers, performed two verbal HRTs, one with English and one with Hebrew consonant letters as well a visual statistical learning task. They were retested after six months. We further assessed word- and nonword learning as well as reading skill. Test-retest reliability of the HRT was found to be close to zero, suggesting it has no predictive validity. Moreover, Hebb learning performance did not correlate with other measures of statistical learning or with linguistic capacities. These results and their theoretical implications are discussed in the light of recent findings regarding the reliability of statistical learning tasks (Siegelman & Frost, 2015).