A New Look at Dual Language Learner Data Gaps

Nationally, one in three children from birth to age five are DLLs or learning English in addition to their home language.[1] States like California, Texas, and Florida are known for having large DLL populations, but a growing number of immigrants and refugees settling in new destinations across the nation means that every state is home to DLLs. In Alabama, Arkansas, and South Carolina, they represent 10 percent or more of the birth to age five population. As a result, early care and education (ECE) programs in every state should expect to welcome DLL children who speak many different languages.

DLLs bring cognitive, social, and cultural strengths that contribute to lifelong learning and development. Research shows that strong home language skills enhance bilingual children’s development of English, which is important because English proficiency at kindergarten has been associated with long-term academic achievement. Over time, education policy has shifted to recognize home language as an asset, rather than a deficit, and as a critical part of a child’s learning experience.

DLLs benefit greatly from high-quality early learning experiences and yet, compared to non-DLLs, they are less likely to participate in early childhood programs. State leaders have taken steps to encourage their enrollment in public ECE programs. Texas prioritizes enrolling DLLs in its state pre-K program, for example, and Colorado recognizes DLL status as a qualifying factor for additional hours of pre-K.

However, adequately and equitably serving DLLs requires more than enrolling them in ECE programs. To start, state leaders need access to information about the languages DLLs speak, which program(s) they attend, and their language proficiency over time to adequately prepare the workforce to support them and their families. Currently, census data provide the best approximation of the young DLL population, but this information lacks the detail needed to inform strategic decisions and track impact.

In 2018, New America provided nine recommendations for how state leaders can improve DLL data practices in three areas: (1) screening, identifying, and tracking DLL enrollment; (2) evaluating program quality; and (3) assessing learning outcomes.

Since then, no state has addressed all of the recommendations, and no single recommendation has been adopted by all states. This brief synthesizes recent interviews with state and national ECE leaders who are actively working to address DLL data gaps. Each section includes a summary of the original brief, an update on each recommendation, and emerging research or solutions in the topic area. The final section discusses three challenges to closing data gaps.

Serving DLLs equitably in any ECE program starts with understanding who they are, what languages they are exposed to, and their proficiency in those languages. The previous brief discussed how the fragmented and mixed-delivery nature of the ECE system presents challenges for gathering this information in a systematic way. While federal policy mandates the identification and data collection of English learners (ELs) in K–12 education, there is a lack of standardized protocols for DLL identification across the ECE system. Different policies, standards, and regulations govern the variety of programs that comprise the ECE system.

Our 2018 brief offered two recommendations to understand DLL enrollment in ECE programs:

• Adopt a uniform protocol to identify DLLs, understand home language experiences, and share these data across state ECE programs.
  • Progress toward Recommendation: Some states have made progress in collecting data on DLLs’ home languages, although the policies sometimes cover only one of the many ECE settings. According to a 2021–22 national survey of state-funded pre-K, almost two-thirds of programs (63 percent) can report enrollment data by home language, compared to 43 percent in 2018. However, it is unclear what systems and processes are being used by these programs to identify DLLs.
• Screen for language abilities in both English and the home language.
  • Progress toward Recommendation: One-third of state-funded pre-K programs screen children in their home language when possible. Data reported between 2017 to 2020 show variations in which state programs screen children in their home language and which do not.

Despite progress in this area, there has been a continued research effort to identify key steps to a comprehensive DLL identification process and foundational elements necessary for implementation. Below, we highlight two states and two federal programs that have taken steps to identify and gather information about DLLs.

State Policy Actions

California: Implementing DLL Identification across ECE Programs

DLLs comprise 59 percent of the birth to age five population in California. Yet, only recently did the state adopt a systematic approach to identifying them in state-funded ECE programs. On October 5, 2021, Governor Gavin Newsom signed Assembly Bill (AB) 1363 into law, establishing a process to identify and share information about DLLs enrolled in California State Preschool Programs (CSPP), which operate in schools or community-based settings. In addition to reporting child-level information, programs must report the languages used in the classroom and the languages spoken by staff members. The bill recognizes the linguistic and cultural assets of bilingualism and states that the purpose of identification is to prepare ECE programs and providers to support DLLs’ proficiency in both languages.

Previously, most programs relied on informal strategies to identify DLLs, such as asking the child’s caregiver about languages spoken at home or through classroom observations. A 2018–19 survey revealed that less than half of California’s ECE program directors reported collecting information about the number of DLLs enrolled in their programs. Programs participating in CSPP or receiving Head Start funds were more likely to collect information about DLLs and use tools such as home language surveys.

The impetus for AB 1363 came from research conducted through the First 5 California DLL Pilot Study and advocacy work led by Early Edge California and Catalyst California (formerly Advancement Project California). According to Carolyne Crolotte, Director of DLL Programs at Early Edge, the organizations partnered with experts and stakeholders to create a DLL Policy Platform, which included a call for a uniform process to identify DLLs.[2] This recommendation was incorporated into Newsom’s Master Plan for Early Learning and Care released in December 2020.

The provisions of AB 1363 were modeled on the Language Learning Project, a multi-agency professional learning initiative in Fresno, California, to build practitioners’ capacity to meet the needs of DLLs and their families. The identification process includes two steps: (1) identifying the child’s primary home language through a survey; and (2) interviewing families to learn about the language and cultural experiences of the child.

Crolotte described the process as a “vehicle and opportunity for building family engagement” that preferably takes place between the child’s teacher and family in the language the family speaks. To balance the ideal with the current reality of staffing shortages, the bill permits other designated staff members to conduct the in-depth interview and share the information obtained with the child’s teacher. In these interviews, programs can share information about the benefits of bilingualism and strategies for supporting home language development.

A second bill, AB 393, currently at the governor’s desk, would require general child care programs and migrant child care programs, which are separate from CSPP, to implement a similar identification and data collection process. Expanding DLL identification to other ECE programs would reach more of the birth to age five population and represent another step towards achieving the Master Plan’s goal of identifying and gathering information about DLLs across all ECE programs. (Note: AB 393 was signed by Governor Newsom on October 8, 2023).

Illinois: Identifying DLLs in Community-Based Settings

About one-third of the birth to age five population in Illinois are DLLs. Currently, school-based preschool programs administered by a public school district are required to identify DLLs using a home language survey. DLLs identified through the survey are then screened for English proficiency. Community-based programs are exempt from this requirement (except for those subcontracted by a school district to provide preschool services). The resulting dynamic, described by Erika Méndez, Director of P-12 Education Policy at the Latino Policy Forum, is that the early childhood system is inconsistent in providing settings with common expectations and information to inform hiring and staffing to support DLLs and to advocate for resources and funding to meet changing classroom needs.[3]

Earlier this year, the Illinois State Board of Education was awarded a $4 million dollar federal Preschool Development Grant Birth through Five (PDG). The grant includes an initiative to improve the capacity of community-based providers to identify and serve DLLs by funding consultants to travel to areas where DLLs might be under-identified to provide needed training and capacity. These consultants will train community-based providers to interview families and administer language screeners; they will also help administer screenings on site, as needed. A portion of the grant will also support data collection and data systems integration.

Méndez and Rebecca Vonderlack-Navarro, Vice President of Education Policy and Research at the Latino Policy Forum, emphasized that the PDG grant is only the first step and more sustained funding is needed to prioritize this population. To effectively target money and resources to providers, there needs to be a common way to identify DLLs across programs. The information gathered from this initiative will help demonstrate the need for a workforce that is prepared to support the young multilingual learner population and, later down the line, an agency that is aligned on what it takes to ensure these students receive the same quality bilingual programming across the mixed-delivery system.[4]

Federal Policy Actions

Child Care and Development Fund

The Child Care and Development Fund (CCDF) supports families with low incomes through child care subsidies. States were required to report the primary home language of families receiving CCDF subsidies starting in fiscal year (FY) 2017. This information helps ensure CCDF reaches a proportionate number of DLLs and their families with low incomes, which is important as the program continues to be an integral part of the ECE infrastructure.

The home language categories shown on the CCDF website are “English,” “Spanish,” “other,” or “unspecified/invalid.” The website lists the multiple languages that may fall into the “other” category. While the website says that these data are a snapshot of states’ progress in meeting new reporting requirements, Alabama, Iowa, and Oregon continue to not report home language information as of FY 2020.

Head Start

Head Start’s performance standards are an example of how ECE policies can be written in support of DLLs. These asset-based policies enable the collection of language information about DLLs and staff for all Head Start programs. This includes the home languages of the children enrolled, the percentage of DLLs enrolled, and staff members’ proficiency in languages other than English.

Head Start defines a DLL as a “child who is acquiring two or more languages at the same time, or a child who is learning a second language while continuing to develop their first language.” The performance standards, updated in 2016, mandate that programs “recognize bilingualism and biliteracy as strengths” and serve as a model in all three areas discussed in this brief.

Starting with screening, identification, and enrollment, Head Start programs are encouraged to gather language information from DLLs and their families through a needs assessment process conducted in the family's home language.

In terms of program quality, the standards mandate that programs implement evidence-based teaching practices that support both English language acquisition and home language development. They also specify steps to support language development even when staff members do not speak the same language as the child.

When assessing DLLs, programs must, to the extent possible, designate a qualified bilingual staff member who knows and understands the child’s language and culture to administer screenings and assessments in the language that best captures the child’s skill level and to assess language skills in both English and the home language. The standards include guidelines for what to do if there are no bilingual staff members, contractors, or consultants available.

Head Start provides resources to ensure the full participation of DLLs and their families. The Dual Language Learners Program Assessment is one tool that allows programs to reflect on strengths and weaknesses, access resources to improve their practice, and track progress. Yet, even with these policies and resources, little is known about the implementation fidelity of these practices. One study found minimal home language use even in Head Start classrooms with large populations of DLLs.

Once state leaders identify DLLs and understand where they are in the ECE system, it’s important to ensure they receive high-quality early learning experiences that incorporate best practices for supporting their language development.

Quality rating and improvement systems (QRIS) emerged as a federal and state strategy to evaluate ECE program quality in a systematic way. QRIS provides a tiered rating based on an assessment of a variety of domains, including staff member qualifications, classroom environment, and child assessment. These data are used to support quality improvement efforts, including coaching and funding. Another purpose of QRIS is to signal to families the quality of ECE programs to help them select a program that best meets their needs.

Unfortunately, many state QRIS have historically failed to incorporate and prioritize DLLs’ needs. As noted in a previous brief, a program can receive a high rating even if it does not address any DLL-related indicators, such as collecting home language information. In addition, the process to participate in QRIS, which can be required for certain providers and voluntary for others, is time-intensive, costly, and less accessible for culturally and linguistically diverse providers. It is unclear whether QRIS ratings factor into families’ decision-making about ECE programs, especially for DLLs and their families.

Our 2018 brief provided the following recommendations for improving state QRIS for DLLs:

• Adopt and prioritize DLL-related indicators.
  • Progress toward Recommendation: At least 41 states and the District of Columbia have implemented QRIS. According to the Quality Compendium, an online resource that tracks key data on QRIS programs through self-reports by QRIS administrators, 20 out of 44 programs include indicators that assess cultural and linguistic diversity in categories such as administration, curriculum, learning environment, and family engagement.[5]
• Encourage participation of linguistically diverse providers in QRIS through technical assistance and outreach.
  • Progress toward Recommendation: Analysis of Quality Compendium data shows that 40 programs currently offer technical assistance (TA) services at no additional cost, with six of those 40 programs offering additional TA for a fee. A majority of these programs (28) offer language access support for non-English speaking providers, and 18 programs offer TA in multiple languages. Ohio’s Step Up to Quality program specifies prioritizing on-site TA to programs serving a high percentage of “high-needs children,” which includes DLLs.
• Translate QRIS ratings into other languages and create and publicly report a DLL subscore of all DLL-related indicators.
  • Progress toward Recommendation: A majority of QRIS programs (30) provide information in other languages, and 14 programs provide language access support to non-English speaking individuals to help boost public awareness of the system.

Research and Policy Actions to Support Program Quality for DLLs

Implementing the recommendations listed above would reduce the barriers for culturally and linguistically diverse providers to participate in QRIS. Doing so could also strengthen the purpose and relevance of QRIS for DLLs and their families. At the same time, critics of QRIS have raised questions about what high-quality means and who gets to define it. Often, the rating scale is biased against providers who are more likely to serve DLLs and preferred by families of DLLs.

Within QRIS, the common observation tools used to assess program quality may not capture components of a high-quality program for DLLs. The quality of the learning environment, especially teacher-child interactions, is an important predictor of children’s academic and social-emotional development. This is often measured through classroom observation tools such as the Classroom Assessment Scoring System (CLASS) or Environmental Rating Scale. Studies show that CLASS can assess the quality of teacher-child interactions regardless of the language composition in the classroom, but it does not assess the language environment or instructional strategies that are supportive of both English and home language development.

The Classroom Assessment for Support of Bilingual Emergent Acquisition (CASEBA) fills this gap by evaluating components of classroom quality specific to DLLs’ language development that other tools do not measure. This includes the quality and quantity of language instruction in English and the home language. Alexandra Figueras-Daniel, Assistant Research Professor at the National Institute for Early Education Research (NIEER), co-developed CASEBA to assess instructional practices and classroom quality based on the needs of DLLs and highlight opportunities to serve them better.[6]

Drawing from her research and time working in a dual immersion school, Figueras-Daniel saw the value in a teacher-facing version of CASEBA for the purpose of professional development. A complementary educator self-reflective tool, Self-Evaluation of Supports for Emergent Bilingual Acquisition (SESEBA), helps current teachers understand and improve language and literacy practices in English and the home language. These tools are being implemented with early childhood educators at the cohort and district level in New Jersey. They have yet to be adopted by any state QRIS.

Kindergarten readiness assessments (KRAs), completed at kindergarten entry, are an opportunity for state leaders to understand the impact of their investments in early childhood. The previous brief discussed the limited availability of validated bilingual KRAs and bilingual, bicultural assessors. Teachers wanted more clarity on guidelines for linguistic accommodations provided to DLLs when administering KRAs. Once assessments are complete, state leaders have to decide whether to publicly report KRA data and disaggregate the data by DLL status. The brief recommended that such reports should note if KRAs are administered in English only and how that impacts the conclusions one can draw given that DLLs’ knowledge and skills are spread across two languages.

Our 2018 brief included the following recommendations to incorporate the needs of DLLs in KRAs:

• Assess DLLs in both English and the home language using valid bilingual assessment tools and bilingual, bicultural assessors.
  • Progress toward Recommendation: There is no standardized way for how assessment tools are designed to incorporate the needs of DLLs. According to scholars from the Learning Policy Institute, only Oregon and Texas have a Spanish KRA that provides and accepts responses in Spanish. Versions of Teaching Strategies GOLD and Desired Results Developmental Profile (DRDP) that are used as KRAs accept responses in multiple languages and include domains to assess English development. DRDP–Kindergarten also has a supplemental domain assessing Spanish language and literacy development. States can customize commercially developed KRAs or develop their own, which may leave out language and literacy items, such as phonological awareness and letter knowledge, that may actually be important information to track.[7] There remains a gap in KRAs available in languages other than English and Spanish and a lack of research on the validity and reliability of state-customized KRAs. Researchers have described the need for flexible assessment protocols so children can receive and respond to items in the languages they speak. States have enacted different policies regarding the linguistic accommodations available to DLLs. These may include providing directions in English and the home language, translating specific item prompts during direct assessments, and allowing students to display their knowledge and skills using any of the languages they know.[8] States also have varying policies regarding teachers' linguistic capacity when using these accommodations.
• Provide teachers with professional development and guidance on how to administer KRAs with fidelity.
  • Progress toward Recommendation: To implement KRAs with fidelity, educators must be trained to distinguish between a student’s lack of English or cultural proficiency and inadequate content knowledge. While states may note in their standards that they acknowledge home languages, there is limited information on how teachers or other assessors are trained to do so, especially in instances where the assessment is not available in the child’s home language or when there is no one available to conduct the assessment in a particular language.
• Explain the limitations of publicly reported data on KRAs, especially if it is going to be disaggregated by DLL/EL status.
  • Progress toward Recommendation: Not all states require KRA data to be made publicly available. Some states that publish KRA data annually disaggregate the results by DLL status. Others acknowledge linguistic accommodations available during assessment. For example, the Illinois State Board of Education notes how the state’s assessment tool allows children to “demonstrate skills and abilities in a variety of ways, such as use of a communication board, sign language, or their native language.”

Research and Policy Actions to Support DLL Assessments

One widely used assessment tool is the Desired Results Developmental Profile (DRDP). DRDP has three versions spanning birth through 12 years old; the versions for infants and toddlers and preschool-aged children are available in Spanish and Chinese. The design process for DRDP included plans to ensure that the assessment tool and administration process were appropriate for DLLs. DRDP includes a foundational language and literacy domain that assesses skills in all languages as well as a supplemental domain to assess English development. Research suggests that DRDP is an appropriate measure for DLLs and provides information about language developmental milestones which may be relevant for guiding instruction with DLLs.

In general, more comprehensive assessment tools that are linguistically accessible, developmentally appropriate, and aligned from early childhood through the early elementary grades are needed. In the absence of assessment tools available in languages other than English and Spanish, researchers often translate assessment tools to other languages. However, direct translations do not ensure cultural relevancy or address differences in language structure and literacy concepts. Researchers say that new assessments must:

• Be normed and validated on populations of children who speak the target language;
• Be normed and validated on children who live in the U.S.;
• Reflect the cultures being represented; and
• Be equated to ensure equivalent difficulty levels and comparability of scores.

Aside from assessment tools, Lindsay Meeker, Early Childhood Access Consortium for Equity (ECACE) Project Director at Western Illinois University, said that state leaders also need to be “thoughtful about our assessment trajectories and schedules and how assessments talk to each other,” ideally starting at birth, but particularly from pre-K to kindergarten.[9] For example, California and Missouri preschool programs use DRDP, but both states do not require a KRA. Without this alignment, it is hard to build a comprehensive linguistic profile of a child over time.

Illinois state leaders are working to more closely align their statewide KRA, the Kindergarten Individual Development Survey (KIDS), with stronger literacy components such as those present in DRDP, plus the linguistic considerations developed through WIDA Early Years. Expanding KIDS to start from age three through the early elementary grades would be another valuable contribution to equitable assessment for DLLs, according to Meeker.[10]

Over the past five years, there has been minimal progress in addressing DLL data gaps. This finding is concerning, given that the DLL population is present in every state, and growing. We found that only two states have taken action to implement some of our recommendations—mostly in the areas of identification and enrollment. There remains wide variation in the level and type of information state leaders have about their DLL population, impacting their ability to implement ECE experiences and truly address DLLs’ learning needs.

During our interviews, state and national leaders highlighted the following challenges that may be perpetuating these data gaps.

Lack of Alignment in Terminology to Describe DLLs

In early childhood, the term dual language learner rather than English learner is preferred because young children are generally learning both English and their home language simultaneously. The distinction may assuage families wary of their child being labeled as an EL, as DLL identification typically does not necessitate EL designation.

The use of a broader definition in ECE acknowledges the developmental trajectory of language acquisition, but it also makes alignment across systems more complex, especially as children transition into the public school system where there is a focus on English proficiency. There is also the question of whether the different ways systems talk about these students, shifting from a broad term to a more narrow definition, impacts the subsequent supports and services they receive.[11] These differences may be most apparent between ages three and five, where there is overlap between the definition of DLLs and ELs and where preschool programs may have different policies for DLLs depending on where they operate.

One of the unintended consequences of the difference in framing, suggests Vonderlack-Navarro of the Latino Policy Forum, is downplaying the civil rights of DLLs.[12] Currently, ELs are afforded legal protections that don't apply to DLLs.

Minimal Resources to Operationalize Policies across the Mixed-Delivery ECE System and through the Early Elementary Grades

ECE professionals can't make changes to address data gaps without investments in implementation, such as through PDG Birth through Five grants. This includes guidance, technical assistance, professional development, data systems, and new assessment tools that measure what the research says is needed to support DLLs.

Despite research showing the effectiveness of teacher training on DLL outcomes, there are not many DLL-specific professional development options available. State leaders could copy California and fund DLL-specific professional development. One promising model currently being implemented in New Jersey is the Building Early Learning Latiné Educators initiative, where Figueras-Daniel trains Latinx preschool teachers on DLL teaching strategies using CASEBA and SESEBA.

State leaders could also centralize resources on best practices when assessing DLLs. However, these resources must be paired with appropriate assessment tools. The Competitive Grants for State Assessments program is a federal funding source that supports the development of assessments for ELs, in addition to other categories. In 2022, three out of 11 grant awards included plans for EL-focused assessments; none focused on the pre-K through early elementary grades.

Limited Designated Funding for DLLs in ECE

Implementing a per-child allocation for DLLs and equitably distributing the funds across the ECE system could result in better data, implementation support, and outcomes. In the past, the federal government has recognized the importance of the early years and funded certain special populations in early childhood.[13]

One available federal fund is Title III of the Every Student Succeeds Act. Title III funds activities promoting English proficiency and academic achievement of ELs, which can serve ELs as young as age three if local educational agencies operate ECE programs. Currently, states are less likely to invest Title III dollars in the implementation of preschool programs. Meeker says, “It’s a choice on how districts use their Title III money to fund their preschoolers. And some states choose to do it, some districts choose to do it [and] some don’t.”

How to spend Title III funds is constrained by the limited resources allocated for EL education. Funding for ELs over the past 20 years has actually decreased, when adjusted for inflation, despite the steady growth of the EL student population. For districts to invest these funds in preschool, Title III funding needs to be significantly increased. At the same time, ESSA reauthorization could include provisions that allow states to incorporate DLLs into the counts used to determine Title III state allocations. This recommendation would hinge on the availability of reliable and valid DLL data, which could be an added incentive for states to collect this information.

I am grateful to Patricia Chamberlain, Carolyne Crolotte, Alexandra Figueras-Daniel, Lindsay Meeker, Erika Méndez, Maki Park, and Rebecca Vonderlack-Navarro for their expertise. I would like to thank Amaya Garcia, Leslie Villegas, Carrie Gillispie, and Cara Sklar for their edits, and Sabrina Detlef, Katherine Portnoy, and Mandy Dean for their communication support.

We thank the Heising-Simons Foundation and the W. Clement & Jessie V. Stone Foundation for their generous support of this work. The views expressed in this report are those of the author and do not necessarily reflect the views of the foundations, their officers, or their employees.

[1] This brief uses the term “dual language learners” (DLLs) to refer to young children from birth to age five. In the K–12 years, federal law refers to these students as “English learners” (ELs).

[2] Interview, May 9, 2023.

[3] Interview with Latino Policy Forum's Rebecca Vonderlack-Navarro (Vice President of Education Policy and Research) and Erika Méndez (Director of P-12 Education Policy), May 11, 2023.

[4] Interview with Vonderlack-Navarro and Méndez, May 11, 2023.

[5] The online resource was last updated in 2021. Florida has three QRIS programs. These data count each program individually.

[6] Interview, May 4, 2023.

[7] Interview with Lindsay Meeker (ECACE Project Director, Western Illinois University) and Patricia Chamberlain (Consultant, Chamberlain Educational Consulting), August 14, 2023.

[8] Amaya Garcia, “Interview: New Study Takes a Closer Look at Kindergarten Entry Assessments for ELs,” EdCentral (blog), New America, November 1, 2018, https://www.newamerica.org/education-policy/edcentral/interview-new-study-takes-closer-look-kindergarten-entry-assessments-els/.

[9] Interview with Meeker and Chamberlain, August 14, 2023.

[10] Email with Meeker, September 1, 2023.

[11] Interview with Vonderlack-Navarro and Méndez, May 11, 2023.

[12] Interview with Vonderlack-Navarro and Méndez, May 11, 2023.

[13] Part B of the IDEA Act covers children with disabilities ages three through 21. During IDEA's reauthorization in 1986, Congress amended Section 619 of Part B to fund the Preschool Grants program specifically for children with disabilities ages three through five.