Continuity in Vocabulary Development: Predictions from Decontextualized Vocabulary and Lexical Access

• In Event: 2-093 - Beyond Vocabulary Size: New Considerations for Receptive Vocabulary Assessment in the Preschool Years

Integrative Statement

Vocabulary knowledge and speed of word processing develop rapidly in the first two years of life (Fenson et al., 1994; Fernald, Perfors, & Marchman, 2006) and are the building blocks for preschool and school-age language and cognitive skills (e.g., Duff, Reen, Plunkett, & Nation, 2015; Marchman & Fernald, 2008). However, the variance accounted for by vocabulary size is generally modest, potentially owing to the use of measures that collapse across a continuum of word knowledge from weak to strong. Recently, we found that a direct assessment of vocabulary size, hypothesized to tap more stable, decontextualized knowledge at age two, predicted third-year general language skills and showed utility in distinguishing children with low- relative to average-to-high language ability.

The current research extends this finding by exploring prediction from both vocabulary size and word processing efficiency. Children were first tested at 23 months of age (N=68, Mage=22;29) and completed the Computerized Comprehension Task (CCT; Friend & Keplinger, 2003; 2008). From this task we estimated decontextualized vocabulary size (i.e., number of correct touches) and obtained two measures of word processing efficiency: visual RT (i.e., latency to execute a gaze shift from the distractor to the target image), and haptic RT (i.e., latency to touch the target). A subsample (N=42) returned for visits at ages three (Mage=37;21) and four (Mage=49;15) and they completed the Peabody Picture Vocabulary Test (Dunn & Dunn, 1997), a measure of receptive vocabulary size.

Zero-order correlations revealed that, at 23 months, vocabulary size was significantly correlated concurrently with visual RT (r(63)=-.42, p<.001) and haptic RT (r(67)=-.24, p=.048): children with larger vocabulary sizes processed words more efficiently. Early vocabulary and visual, but not haptic RT, were prospectively correlated with vocabulary at ages three (r(44)=.49, p=.001 and r(44)=-.37, p=.01, respectively) and four (r(44)=.39, p=.008 and r(44)=-.37, p=.01, respectively). We next performed two linear regressions predicting vocabulary at age three and four, respectively, with child sex and maternal education as control variables in Step 1 and 23-month vocabulary, visual RT, and haptic RT in Step 2.

Regression 1 (Year 3): Step 1 was not significant but Step 2 was (R2Δ=.42, F Δ=9.47, p<.001). Vocabulary size, visual RT, and haptic RT were all significant predictors in the final model (β=.40, p=.01; β=-.35, p=.02; β=.37, p=.01, respectively). In contrast to predictions, once the variance from early vocabulary was accounted for, children slower to touch the target image at 23 months had larger vocabularies at age three: this may suggest a benefit of more deliberate processing in a computationally-expensive response modality.

Regression 2 (Year 4): Step 1 was not significant but Step 2 was (R2Δ=.26, F Δ=4.68, p=.007). Vocabulary size was the only significant predictor of age four vocabulary (β=.35, p=.047).

Together this evidence indicates that vocabulary size and the efficiency with which word meanings are processed significantly predict downstream vocabulary outcomes. However, it also suggests that vocabulary size may be a more robust, unique predictor of outcomes further downstream with implications for early identification of children at risk for language delay.

Authors

• Erin Smolak, San Diego State University & University of California, San Diego
  Presenting Author

• Kristi Hendrickson, University of Iowa
  Non-Presenting Author

• Pascal Zesiger, University of Geneva
  Non-Presenting Author

• Margaret Friend, San Diego State University
  Non-Presenting Author