Implementation of the Devereux Early Childhood Assessment with Spanish-Speaking Families in Child Welfare Contexts

Background and Purpose

The Devereux Early Childhood Assessment (DECA) is used in a midwestern state child welfare agency to assess developmental need among children entering foster care. In August 2024, bilingual (Spanish / English) screeners utilizing the DECA identified problems associated with use of the Spanish language version of the instrument. A workgroup of bilingual screeners from agency and university partners was engaged to study these problems.

Study objectives are to identify which characteristics of the DECA Spanish translation are most problematic; why; and what recommendations can be made for modification. The data are being used to develop a guidance document to offer evidence-based recommendations for adaptations to implementation.

Methods

A focus group representing users of the DECA instrument in bilingual contexts was implemented. In addition, surveys about the DECA were developed and fielded (data collection ongoing). The last component of the study is a scheduled second focus group to review the survey findings and finalize the guidebook.

Initial focus group participants included six screeners utilizing the Spanish language DECA (purposive sample). Survey participants include all 25 bilingual screeners statewide.

The focus group interview was developed by the workgroup. Surveys include DECA items and their published Spanish translations, queries about each item, and requests for recommendations for modification. Surveys were programmed into survey software, with links distributed via email.

Focus group responses were analyzed using thematic analyses. Survey data are being descriptively analyzed (closed-ended questions) and content analyzed (open-ended questions). The draft guidebook with specific Likert scale and item modification recommendations will be reviewed by a group of experts to establish Item-level Content Validity Indices (I-CVI) with regard to wording / presentation appropriateness. Analyses will assess universal agreement and Scale Content Validity Indices (S-CVI). Modified kappa statistics will assess agreement.

Results

Initial focus group results identified four, main themes:

• The Likert scale is easily misunderstood, particularly by recent immigrants, some of whom have no prior exposure to this form of inquiry. Respondents seek concrete anchors.
• Problematic questions incorporate Spanish translations that are overly formal and not readily understood by respondents.
• Respondents demonstrate high susceptibility to social desirability bias. They often convey indicators that they are seeking to provide the “right” answer.
• The need to deliver the surveys in a standardized manner inhibits screeners from offering explanations or examples that could increase validity of the data.

Conclusion and Implications

Themes identified based on focus group responses preliminarily validated the concerns broached by screeners. These findings suggest the potential for systematic error related to data validity for a subset of Spanish-speaking respondents. Additional data collection will provide detail and support recommended modifications.

The preliminary findings suggest that typical strategies involving translation and back-translation of instrumentation at a single point in time may limit the validity of data collected in multicultural or domain-specific contexts (e.g. child welfare). Continuous review of the appropriateness of translations, development of multiple translations (for instance, to manage dialectical variation), and development of specialized guidance to support implementation with specialized populations may improve data validity.