Bilingual Assessment in Early Childhood & Prevention Science

Introduction:

Prevalence rates of young bilingual children, in general and within prevention science initiatives, continue to rise dramatically.  In turn, accurate assessment is a critical component throughout prevention science, from the development of valid screening approaches to precise estimation of effects. Yet, there generally has been limited focus on the features of early childhood measures developed in multiple languages. As such, prevention science researchers, program staff, and policymakers may select bilingual measures based on their experience with the English version, popularity, and/or convenience.  Additional challenges include varied approaches for sampling, standardization, administration, scoring, and dialectical inclusion.

Methods:

This paper comprehensively examines bilingual measures to illuminate the present state of the prevention science field and contribute to improved efforts.  Seventeen early childhood measures within and across the English and Spanish languages were analyzed for their psychometric, cultural, and linguistic properties. This included attention to content equivalence of items and approach, semantic equivalence of translations, and structural consistency. 

Results:

Analyses revealed extensive variability among bilingual measures for young children, with some Spanish measures even inappropriately presenting data.  A number of assessments continue to be rudimentary translations, while others have undergone advanced procedures (e.g., language and content expert panels, item and latent ability analyses to identify item or order biases, international standardization or statistical equivalence).  While the English edition is often better developed than its Spanish counterpart, the opposite pattern was also found.  Such discrepancies across and within measures can produce variable results within prevention science. The paper will present the following to conference participants: 1) the percentage (and identification) of measures not yet well-developed in English, Spanish, and across both languages; 2) which early childhood domains are most susceptible to misestimation; 3) the statistical impact of less-valid bilingual measures on program effect estimates; 4) the approaches leading to optimal bilingual assessment development; and 5) the outcome of the author’s communications with North American publishers to promote continued improvement in the field.

Conclusions:

While bilingual assessments have improved over past decades, the identification, utilization, and development of sound measures for the growing bilingual population continues to necessitate attention. This paper can serve as a guide for prevention scientists in considering the key cultural, linguistic and psychometric facets of bilingual assessment.