Skip Navigation

icon of glasses and a book Early Learning Programs and Policies


- OR -


- OR -


- OR -

FY Awards

- OR -

Development of the Bilingual Assessment of Phonological Sensitivity

Year: 2016
Name of Institution:
Columbia University, Teachers College
Goal: Measurement
Principal Investigator:
Hammer, Carol
Award Amount: $1,400,000
Award Period: 4 years (9/1/2016-8/31/2020)
Award Number: R305A160081


Co-Principal Investigator: Shelley Scarpino (Bloomsburg University)

Purpose: In order to address the early literacy needs of young dual language learners (DLL), this research team will develop a comprehensive, reliable, and valid Spanish-English assessment of phonological sensitivity (PS) that can be used with 3- to 5-year old DLLs. Phonological sensitivity (PS), or the awareness that words are comprised of syllables and sounds, begins to develop during the preschool years and has been shown to be one of the best predictors of reading outcomes in both English speaking children and Spanish-English DLLs. The Bilingual Assessment of Phonological Sensitivity (BAPS), is designed specifically for teachers, and will efficiently measure DLL preschoolers' abilities across the PS developmental continuum in both languages. By providing information about where Spanish-speaking DLLs are on the developmental continuum of PS in both languages, the BAPS will help teachers to identify children who have difficulties with PS, and determine their instructional responses to help these children improve.

Project Activities: Over the course of four years, the researchers for this project will develop and produce preliminary operational forms for the BAPS. Using an iterative development process, they will conduct a prototype study, pilot study, and field test with increasingly large student samples at each stage. The research team will use the findings from each study to refine the instrument and administration procedures for the next phase, culminating in the field test, which researchers will use to establish a score scale and norm groups for operational use. In addition to establishing the technical quality of the assessment, researchers will also work to maximize the instrument's usability and utility for teachers, so that it is easy to administer and produces score information that is easy to interpret and use for instructional decision-making.

Products: The research team will produce operational forms of the Bilingual Assessment of Phonological Sensitivity (BAPS), as well as training and administration materials, technical documentation of the test's reliability and validity, a reporting platform, and guidance on how to interpret and use BAPS scores to inform instruction. In addition, the researchers will produce reports, conference presentations, and peer-reviewed publications to disseminate their findings.

Structured Abstract

Setting: This project sill take place in preschool and Head Start programs in urban and suburban areas in Pennsylvania, New Jersey, and Delaware.

Sample: This study will involve 1200 Spanish-English DLL children between 3- and 5-years of age. Researchers will include children from families from a range of language backgrounds including Mexican, Puerto Rican, Dominican, Cuban as well as Central and South American.

Assessment: Measures include the Bilingual Assessment of Phonological Sensitivity (BAPS), a norm-referenced measure that is appropriate for 3- to 5-year-old Spanish-English DLLs and assesses the full range of PS in both languages. The BAPS comprises separate versions in Spanish and English (the BAPS-S and the BAP-E), each of which has two forms (Level A and Level B) that target different age ranges between 3 and 5 years. Each of the four forms (i.e., BAPS-E Levels A and B; BAPS-S Levels A and B) contains five subtests that follow the PS developmental sequence: (a) Rhyming, (b) Syllable and Phoneme Counting, (c) Blending, (d) Segmenting, and (e) Sound Matching. Within forms, items are ordered by task complexity. The research team will individually administer the BAPS to children through a computer application. The application will produce individual and classroom score reports for teachers, with information on chronological ages, total test scores, and decile ranks for each language based on the normative sample of the test. The BAPS will also produce a user-friendly error analysis that provides teachers with information on which items individual children passed/failed by subtest. This information will allow teachers to identify skill areas in which a child is particularly proficient or unskilled relative to her overall PS abilities, and devise targeted instructional solutions.

Research Design and Methods: In Year 1, the team will generate test administration procedures and a pool of test items, and will conduct a prototype study involving 100 3- to 5-year-old children. They will complete qualitative and quantitative analyses to evaluate and refine the test items and administration procedures. Researchers will modify and reduce the initial pool of items to prepare for the first full test administration. A panel of expert consultants will meet to collect validity evidence based on the BAPS' content. In Year 2, the team will carry out a pilot study with 300 children. They will perform qualitative and quantitative analyses to evaluate the test items and the test administration procedures, with a particular focus on usability for the test, the test administration manual, and the computer delivery platform. They will modify accordingly. In Years 3 and 4, the research team will conduct a field test with a sample of 800 children, and complete statistical analyses to scale the test and collect evidence in support of the test's validity for its intended purpose. To collect evidence of the BAPS' internal consistency, they will also conduct a test-retest reliability study. To collect evidence of the BAPS' relation to other variables, the researchers will explore correlations with measures of vocabulary and other extant measures of PS. The team will create test documentation, training materials, and score interpretation guides, as well as a final version of the BAPS.

Control Condition: This measurement project does not include a control condition.

Key Measures: For analyses of the BAPS' relation to other variables, the research team will use the Phonological, Alphabetic Principal, and Concepts About Print subtests of the Early Literacy Skills Assessment (ELSA), and the Sound Awareness, Picture Vocabulary, and Letter-Word Identification subtests from the English version of the Woodcock-Johnson III Tests of Achievement and the Spanish version (Batería III Woodcock-Muñoz: Pruebas de aprovechamiento). Both the ELSA and the Woodcock-Johnson are standardized assessments available in both English and Spanish. To evaluate usability and fidelity of administration, trained data collectors will video record 25% of their sessions for review and keep notes and complete feedback questionnaires for each testing session.

Data Analytic Strategy: Throughout the iterative development process, researchers will perform qualitative and quantitative analyses. Statistical analyses include:

  • Factor analysis to establish the dimensional structure and appropriate scaling procedure for the test;
  • Classical item analyses to identify potentially problematic items (i.e., those with low discrimination) and revise or confirm item placement on the various forms;
  • Rasch measurement analyses to establish a fixed scale for scoring student performance;
  • Reliability analyses (coefficient alpha) to establish the internal consistency of each test form;
  • Correlational analyses to evaluate the BAPS' relationship to other measures of PS and measures of vocabulary;
  • Differential item functioning (DIF) to confirm that the items operate similarly for students with different background characteristics.