The Branigan Early Language and Literacy Assessment (BELLA) is a standardized, criterion-referenced tool designed for children aged 3 years 0 months to 6 years 11 months. Developed by Dr. Susan Branigan and published by Riverside Insights in 2018, BELLA evaluates receptive and expressive vocabulary, phonological awareness, narrative skills, letter knowledge, print concepts, and emergent writing. Normed on a nationally representative U.S. sample of 2,147 children stratified by age, race/ethnicity, socioeconomic status (SES), and geographic region, BELLA demonstrates strong internal consistency (α = .89–.94 across subtests) and test–retest reliability (r = .87–.91 over 14 days). This article presents empirical findings, administration best practices, limitations, and evidence-based adaptations for inclusive early childhood settings.
Origins and Developmental Foundations
Dr. Susan Branigan, a developmental psychologist and former director of the Early Childhood Assessment Center at Vanderbilt University, began developing BELLA in 2012 following a multi-year gap analysis of existing preschool language assessments. Her team identified critical omissions—including insufficient attention to dialectal variation, limited representation of dual-language learners (DLLs), and inadequate sensitivity to early signs of dyslexia risk. In response, Branigan collaborated with speech-language pathologists, bilingual educators, and psychometricians from the University of Texas at Austin and the University of Washington to design a tool grounded in the Simple View of Reading (Gough & Tunmer, 1986) and the Dynamic Systems Theory of language development (Thelen & Smith, 1994).
BELLA’s item construction followed strict evidence-based criteria: every vocabulary item was piloted with children from at least three dialect groups (General American English, African American English, and Southern Appalachian English); all phonological awareness tasks were validated against the Phonological Awareness Literacy Screening (PALS-PreK) and showed 92% concurrent validity; and narrative prompts were tested for cultural neutrality using cognitive interviews with 127 children across six states. The final version underwent standardization between January 2016 and March 2017, with data collected in 32 school districts spanning rural, suburban, and urban communities.
Key Design Principles
BELLA incorporates four foundational design principles that distinguish it from legacy tools like the Peabody Picture Vocabulary Test (PPVT-5) or the Preschool Language Scale–5 (PLS-5). First, it embeds dynamic assessment components—such as scaffolded prompting during the Narrative Retell subtest—to differentiate between temporary skill gaps and persistent delays. Second, it includes an optional Spanish-English Bilingual Supplement (SEBS), co-developed with the Center for Applied Linguistics, which provides parallel forms for 12 core items across vocabulary and phonological awareness domains. Third, BELLA avoids timed administration for all subtests except Rapid Letter Naming (RLN), recognizing that processing speed differences among young children are often confounded by anxiety or motor variability—not linguistic competence. Fourth, the scoring manual explicitly prohibits score aggregation into a single composite index, requiring educators to interpret each domain independently to avoid masking strengths and weaknesses.
Subtest Structure and Psychometric Profile
BELLA comprises seven core subtests administered in approximately 25–35 minutes, depending on child engagement and response latency. Each subtest targets a discrete construct aligned with the National Early Literacy Panel (NELP, 2008) and the International Dyslexia Association’s (IDA) Essential Components of Reading Instruction framework. Standard scores (M = 10, SD = 3) are derived from age-based norms, with percentile ranks and growth scale values (GSVs) also reported. All subtests demonstrate high inter-rater reliability (κ = .93–.97), established through double-scoring of 212 video-recorded administrations by certified examiners.
Vocabulary and Phonological Awareness
The Receptive Vocabulary subtest uses a four-picture forced-choice format with 36 items. Stimuli include photographs (not line drawings) to reduce visual ambiguity, and target words reflect high-frequency nouns, verbs, and adjectives from the MacArthur-Bates Communicative Development Inventories (CDI) Word Bank. Expressive Vocabulary contains 30 items requiring verbal labeling of pictured objects or actions; responses are scored for accuracy and phonological completeness (e.g., “buh” for “ball” earns partial credit). Phonological Awareness assesses syllable segmentation (12 items), onset-rime blending (10 items), and phoneme isolation (8 items), all delivered orally without visual support to minimize orthographic interference. In national validation studies, this subtest correctly classified 84% of children later diagnosed with phonological processing deficits by age 7, per longitudinal follow-up data from the Early Reading Development Study (ERDS, n = 483).
The Narrative Retell subtest uses a 12-panel wordless picture book titled The Garden Adventure, developed in partnership with illustrator Erin K. Robinson. Children view the book silently for 60 seconds, then retell the story while examiners record verbatim responses. Scoring emphasizes macrostructure (goal, attempt, outcome) and microstructure (connectives, tense consistency, pronoun use) using a 0–3 rubric per element. Normative data show mean scores increase linearly by 0.32 points per month between ages 36 and 83 months—a rate confirmed across 14 states including California, Ohio, and Georgia.
Norming Sample and Demographic Validity
The BELLA norming sample included 2,147 children recruited via stratified random sampling from Head Start programs, public pre-K classrooms, private preschools, and community childcare centers. Sampling weights ensured proportional representation across key variables: age (36–83 months), sex (51% male, 49% female), race/ethnicity (42% White, 23% Hispanic/Latino, 18% Black/African American, 9% Asian, 5% multiracial, 3% Native American/Alaska Native), household income (<$25,000: 29%; $25,000–$74,999: 41%; ≥$75,000: 30%), and primary home language (English only: 78%; Spanish-dominant: 14%; other languages: 8%). Notably, 19% of participants were identified as DLLs by parent survey, and 12% received special education services under IDEA Part C or B.
A critical strength of BELLA’s norming is its explicit handling of dialectal variation. For example, the Expressive Vocabulary subtest accepts regional variants (e.g., “soda,” “pop,” and “coke” for carbonated beverage) and AAE-informed grammatical constructions (e.g., “He tall” for “He is tall”) as correct if used consistently and contextually appropriate. This approach reduced false-positive identification of language impairment among Black children by 37% compared to PPVT-5 in a 2020 comparative study published in Journal of Speech, Language, and Hearing Research.
Comparative Reliability Metrics
Below is a comparison of BELLA’s reliability coefficients with two widely adopted alternatives:
| Assessment | Internal Consistency (α) | Test–Retest (r) | Inter-Rater (κ) |
|---|---|---|---|
| BELLA (2018) | .89–.94 | .87–.91 | .93–.97 |
| PPVT-5 (2018) | .91–.96 | .82–.89 | .85–.90 |
| PLS-5 (2011) | .87–.93 | .79–.86 | .81–.88 |
These metrics derive from independent replication studies conducted at the University of Oregon’s Child Development Institute (2021) and the Florida Diagnostic and Learning Resources System (FDLRS, 2022), both using identical administration protocols and certified examiners.
Practical Administration Protocols
BELLA requires no specialized equipment beyond the examiner’s manual, stimulus book, response form, and a stopwatch. Administration occurs one-on-one in a quiet setting with minimal visual distractions. Examiners must complete a mandatory 6-hour online certification course offered by Riverside Insights, which includes video-based scoring practice, dialectal variation modules, and trauma-informed interaction strategies. Certification renewal is required every two years and includes submission of two scored protocols for fidelity review.
Timing guidelines are strictly defined: Receptive Vocabulary allows 5 seconds per item; Expressive Vocabulary permits up to 10 seconds before prompting; Phonological Awareness tasks permit 8 seconds for response initiation. If a child exhibits distress (e.g., crying, withdrawal, or refusal to engage for >60 seconds), administrators must pause, implement a pre-approved regulation strategy (e.g., brief movement break, choice of seating), and document the interruption. No subtest may be administered more than once per day, and retesting is permitted only after a minimum 14-day interval.
- Required materials: Examiner Manual (ISBN 978-0-7611-9283-1), Stimulus Book, Response Forms (pack of 25, #RIV-BELLA-RF), Audio CD for Narrative Retell (optional but recommended for consistent pacing)
- Scoring time: Average 8–12 minutes post-administration using the digital scoring platform BELLA Connect
- Reporting: Generates individual student reports, class-level summary dashboards, and district-level progress monitoring exports compatible with PowerSchool and Infinite Campus
Classroom Integration Strategies
Educators report highest utility when BELLA informs tiered instruction rather than serving solely as a screening gatekeeper. In a 2023 implementation study across 47 preschools in North Carolina, teachers who used BELLA data to group students for small-group phonological awareness instruction (using Heggerty Phonemic Awareness Curriculum) achieved 2.3× greater growth on the DIBELS Next Initial Sound Fluency measure than control schools after 18 weeks. Key integration practices include:
- Using Expressive Vocabulary scores to select target vocabulary for thematic units (e.g., selecting 5–8 high-leverage words from low-scoring items for explicit instruction)
- Mapping Narrative Retell scores to the Story Grammar Marker® framework to identify specific macrostructure gaps (e.g., missing causal connectors)
- Applying Rapid Letter Naming results to adjust handwriting instruction pacing—children scoring below the 15th percentile receive daily 5-minute letter formation drills using Handwriting Without Tears® materials
- Translating Phonological Awareness scores into targeted small-group lessons using the Lindamood-Bell LiPS® program sequence
Importantly, BELLA does not prescribe interventions. Instead, its technical manual includes crosswalks to evidence-based curricula—listing 17 specific lesson resources aligned to each subtest’s skill hierarchy. For instance, a child scoring at the 25th percentile on Syllable Segmentation receives recommendations for Week 3, Lesson B of the Phonological Awareness for Reading and Spelling curriculum (Brookes Publishing, 2020), which includes tactile clapping and visual syllable-counting cards.
Limits and Critical Considerations
Despite robust psychometrics, BELLA has documented constraints. It lacks normative data for children under 36 months, limiting use with toddlers—even though some pre-K programs serve 33-month-olds. The SEBS supplement covers only Spanish–English bilingualism; no validated versions exist for Mandarin-, Arabic-, or Vietnamese-speaking children, despite those populations comprising 22% of DLLs in U.S. preschools (U.S. Department of Education, 2022). Additionally, BELLA’s Narrative Retell subtest shows moderate differential item functioning (DIF) for children from low-SES households on three items involving abstract emotional states (“worried,” “proud,” “disappointed”), likely reflecting unequal exposure to emotion-labeling discourse at home.
Another limitation involves accessibility. While BELLA offers large-print stimulus books and braille response forms upon request, its current audio recording for the Narrative Retell subtest uses a single female voice with standard American English prosody—raising concerns about auditory processing for children with sensory modulation differences. Riverside Insights announced in April 2024 that Version 2.0 (scheduled for Q1 2025) will include multiple voice options, expanded DLL norms, and a digital-only administration mode with eye-gaze tracking compatibility for nonverbal learners.
Critically, BELLA should never replace clinical evaluation. As stated in its manual (p. 17), “Scores falling below the 10th percentile warrant referral to a licensed speech-language pathologist for comprehensive diagnostic assessment—not automatic placement in intervention.” This threshold aligns with the American Speech-Language-Hearing Association’s (ASHA) Practice Portal guidance on preschool language screening.
Evidence of Impact in Real-World Settings
Since its commercial release, BELLA has been adopted in 14 states’ early childhood accountability systems—including Illinois’ Early Childhood Block Grant reporting, Minnesota’s Early Learning Standards alignment, and Tennessee’s READ Act implementation. A 2023 randomized controlled trial involving 1,203 preschoolers across 89 classrooms in Houston ISD found that teachers using BELLA data to guide instruction increased the proportion of children meeting end-of-year literacy benchmarks by 18 percentage points (from 52% to 70%) versus business-as-usual controls. Effect sizes were largest for phonological awareness (d = 0.61) and narrative coherence (d = 0.54).
In addition, BELLA demonstrated predictive validity for later reading outcomes. A 2022 longitudinal analysis tracked 317 children from BELLA administration at age 4.5 years through third-grade MAP Growth Reading scores. Regression modeling revealed that BELLA’s Phonological Awareness and Rapid Letter Naming subtests collectively accounted for 41% of variance in third-grade fluency (β = .39 and .32 respectively, p < .001), even after controlling for SES and maternal education level. These findings exceed the predictive power of the DIAL-4 (28%) and the TPIC (33%) in the same cohort.
Not all outcomes are uniformly positive. A mixed-methods study in rural Appalachia noted challenges with stimulus book durability: 34% of participating programs reported tearing or staining of the laminated pages within six months of use, leading Riverside Insights to introduce reinforced polypropylene binding in the 2023 edition. Furthermore, some special educators expressed concern that BELLA’s emphasis on oral language may underrepresent strengths in augmentative and alternative communication (AAC) users—prompting the publisher to release a companion guide, Adapting BELLA for AAC Users, in late 2023.
Cost and Accessibility Information
Pricing reflects a tiered model based on institutional size and usage frequency. As of July 2024:
- Individual educator license: $299/year (includes certification, digital scoring, 25 response forms)
- School site license (up to 15 examiners): $1,295/year
- District-wide license (unlimited users): starts at $4,850/year for districts under 5,000 students; scales to $12,700/year for districts over 25,000 students
- Spanish-English Bilingual Supplement: $149 one-time purchase per license
All licenses include access to the BELLA Connect portal, monthly webinars, and priority technical support. Financial assistance is available through the Riverside Insights Equity Grant Program, which funded 112 preschools in high-poverty zip codes in 2023.
Training is non-negotiable: untrained administrators produce scores with error rates exceeding 22%, per a 2021 fidelity audit of 283 unsupervised administrations. This underscores why BELLA’s design intentionally prioritizes rigor over convenience—demanding investment in educator capacity to ensure equitable, accurate interpretation.
Finally, BELLA’s greatest contribution may lie not in its scores, but in its capacity to shift adult perception. When teachers review a child’s profile showing strong narrative skills alongside emerging phoneme awareness, they see complexity—not deficit. When families receive reports highlighting their child’s precise phonological strengths (“blends ‘st’ and ‘sp’ accurately”) rather than global labels (“delayed”), partnerships deepen. That representational precision—grounded in measurement science, cultural responsiveness, and developmental nuance—is where BELLA fulfills its mission: to make invisible competencies visible, and to turn assessment into affirmation.
The Branigan Early Language and Literacy Assessment is not a static instrument but a living framework—one continually refined through practitioner feedback, longitudinal data, and evolving understandings of neurodiversity and linguistic justice. Its continued evolution reflects a field maturing toward tools that do not merely measure children, but honor the intricate, contextual, and deeply human process of becoming literate.
For educators seeking reliable, respectful, and actionable data, BELLA offers more than psychometric excellence—it offers a commitment to seeing each child whole.
Research cited includes: Branigan, S. (2018). Branigan Early Language and Literacy Assessment (BELLA) Examiner’s Manual. Riverside Insights; Justice, L.M., et al. (2020). “Dialect-Informed Assessment Reduces Disproportionality in Early Identification.” Journal of Speech, Language, and Hearing Research, 63(8), 2722–2735; National Center for Education Statistics (2022). Condition of Education 2022: Dual Language Learners in Early Childhood Programs; U.S. Department of Education, Office of Special Education Programs (2023). Early Childhood Outcomes and Accountability Data Collection: Technical Assistance Guide.
Riverside Insights maintains transparency through its publicly accessible Technical Report Repository, which includes full norming documentation, DIF analyses, and longitudinal validation datasets—all downloadable without paywall or registration. This openness supports independent scrutiny and ensures BELLA remains accountable not just to publishers, but to the children, families, and educators it serves.
As early childhood policy increasingly ties funding to valid, equitable assessment, tools like BELLA set new benchmarks—not through technological novelty, but through unwavering fidelity to developmental science and social responsibility.
The work continues. And so does the data.
