Jean-Yves' story is set in 2002, the year the Commonwealth of Massachusetts hired me for the second time as a consultant to the Boston and Somerville public schools.* I was commissioned to address the difficulties associated with screening limited-English-speaking children for special needs identification and to devise "creative" ways of testing these students. Having been a special educator in the United States and abroad, I understood the limitations of formal testing and knew that language strongly affected a child's performance on such screenings.

I thought I had a simple agenda as I set off to observe kindergarten screenings in a Boston public school. Children and mothers were arriving for the scheduled preschool test of 3- to 5-year-olds. On the advice of a school specialist, I followed Jean-Yves, a 4-year-old Haitian child, and his translator through the screening process. Although I know French, not Creole, I could both see and understand that this child did not want to leave his mother, despite the adults cajoling him to come along.

The translator, a young male teacher, took the reluctant child by the hand while the mother said a short farewell and urged him to "listen to the teacher." (My French helped me to decipher this statement.) Jean-Yves held the translator's hand and listened to assurances, in Creole, that they were "just going to play games." He walked with his head and eyes down and was met by a female speech-and-language therapist. She ushered the child and the translator into a tiny room labeled Station #1, and they took their places at the room's two empty chairs. The therapist spoke in English to the translator, who said the child seemed to understand Creole.

The therapist began administering the language portion of the test to Jean-Yves, including questions about his name and about various body parts. The translator patiently repeated the questions in Creole, but Jean-Yves simply looked down and remained silent. The translator told the therapist that these questions were hard for the child. He suggested that Jean-Yves be asked to tell a story about his mother and family. The child remained quiet with eyes wide, but said nothing. The therapist gave the translator paper for Jean-Yves to draw his family. He still did not respond to the translator's suggestions, given in Creole, to draw or to talk. The score sheet was marked "untestable," and the boy and his translator went on to the next station.

At Station #2, the special education teacher sat in a large classroom with puzzles and blocks on the table. Again, Jean-Yves and the translator sat across from the specialist. The specialist and translator chatted for a minute, then a few toys were brought out to entice the child to play. No response. Both the translator and the tester decided to try some segments of the test.

The tester asked Jean-Yves for a red block, then a blue block. When the child did not respond, the translator offered him some blocks to hold. Jean-Yves handled a block or two, then started to make a tower. He did not follow through on the commands of the tester or the translator, who then tried to show him pictures and puzzles. The child looked away, shook his leg, and waited. Once again, the score sheet was marked "untestable," and the boy and his translator were ushered to the next station. The child traveled the circuit, "failed" the hearing test, and was deemed "untestable" on all other portions of the testing that day.

"Untestable" or Culturally Different?

I watched as the translator finished with Jean-Yves, then came out to the play area where other mothers were waiting. I asked the translator if this child's performance was typical of what he saw when he worked in Boston. He confirmed that it was and said he believed many Haitian children were frightened and were not used to doing what adults asked of them in school. He told me that many of the parents he meets often tell their children to "behave," and that means to "be quiet." So he felt that many of these children needed more time than they were usually given to get used to the teachers and to feel comfortable talking.

My next observation puzzled me further. I turned and saw Jean-Yves, who had just been labeled "untestable," playing in the waiting area. He was very animated with the other children. I heard them laughing. I saw the children handling toys and interacting with each other verbally while the mothers chatted and the school staff milled around. Who would see Jean-Yves speaking and happily interacting with other children while he played?

Jean-Yves' mother was given a slip of paper that said he needed to be retested in three months to determine if he had special needs. But the question remained in my mind, especially given the ease with which Jean-Yves interacted with the other children: Was he tested correctly this time? Or did some form of cultural misunderstanding impede the process?

Are We Finding Assets or Deficits?

Jean-Yves' story is not an isolated incident. Currently, one in five public school students in the U.S. is either the child of immigrant parents or is an immigrant her- or himself. Yet it is far too easy for these children, especially those whose dominant language is not English, to "fail" preschool screening and later testing and to enter school with the label "special needs."

Researchers have known for years that English-language learners are regularly misidentified for special education placement. "Once a referral is made, the likelihood of testing is high. Once testing takes place, strong gravitational forces toward special education placement are in motion," wrote Genevieve Fedoruk, author of numerous studies in the 1980s and 1990s. "Once a language-minority child is referred for testing, that same child is placed in special education about 85 percent of the time. Once a child is placed in special education, despite a mistaken assessment, it takes them on average six years to get out."

In addition, researchers have long agreed that standardized testing done in the context of a "child deficit model"—looking for deficiencies in a child's abilities—may do more harm than good, especially for language-minority students. "Children may be temporarily overwhelmed by difficulties or blocked in their expression, but that does not mean that they cannot be helped to develop their strengths," noted Anne Martin in a 1988 Harvard Educational Review article. "Predicting failure can be a way to ensure it."

Still, federal law mandates that language-minority children be screened for special needs before they enter school, and the screening formats used are often culturally biased, dated, or both. In Boston, for example, the Early Screening Test Instrument (ESI) is used to screen all children, despite serious limitations. The ESI, originally developed using a population of 700 white, middle-class children from Rhode Island, had no minorities in its sample. And, according to the manual, the reliability and validity testing of the measurement were done in 1972.

So how can language-minority students be assessed in ways that help them to "develop their strengths" rather than label them as deficient? The first step involves looking carefully at the interaction between teacher and learner. This suggests redefining the purposes, formats, and processes associated with the assessment of diverse language learners.
 

Complex Differences, Complex Solutions

The traditional psychometric model of formal assessment used to evaluate all children consists of standardized tests and other fixed techniques to diagnose language and learning problems. These tools assume that any deficit that might be detected is in the child. This model is in contrast to a sociocultural approach, which assumes that every child presents a unique example of difference and complexity, and that understanding the difference—not the deficit—is the role of educational assessment.

A sociocultural perspective assumes that children learn language in real-life situations that depend on social interactions, and that bilingual children display different knowledge and language uses depending on the social contexts in which they are learning and living. In addition, a sociocultural perspective rests on three premises:

• Bilingualism is a cognitive asset that enhances thinking and learning.
• Sociocultural factors affect learning, and the context, or learning environment, is key to understanding language use.
• Language proficiency and individual learning abilities should be assessed in context and over time.

Within this framework, classroom assessment becomes an interactive process where teachers "sit beside" children to assess and to teach them. This approach to assessment realigns the usual power relationship between the potentially dominant teacher and the dominated child. Both what teachers do and what students do during learning activities are examined, simultaneously and not in isolation. Thus, assessment becomes a four-step process whereby the teacher:

1. self-assesses and researches the child's language and culture;

2. assesses the language demands of the classroom;

3. probes for the child's individual learning strengths;

4. gathers data on the child by monitoring his or her daily interactions in various groups within the classroom.

A sociocultural approach to assessment examines both the process and the products of a language-minority student's learning. Teachers examine the complexity of a child's background social and cultural factors (who a child is and how they learn); political factors (how a child reacts in a particular environment or setting); linguistic factors (how a child uses both native and second languages); and academic/educational factors (how a child performs a given task).

Overall, this framework suggests that educators look not only at what is "wrong" with the child but also what is "wrong" about what they know about language and culture, as well as the learning environment itself (the school, the classroom, or the curriculum).

Portfolio Assessment: Process and Product

To address the need for a more comprehensive and complete assessment of language-minority children, educators in many urban schools have begun using a portfolio assessment approach. The concept is to combine formal and informal assessment materials with a variety of observational sources to create a more complete picture of the strengths and weaknesses of young children as learners.

For language-minority children, the process involves collecting a "portfolio of information" that draws on a variety of sources, from observations of classroom work and play to questionnaires completed by parents. By compiling a wide range of data about a child, educators can gain a more comprehensive understanding of the child and her or his skills (see sidebar "Assessing Language-Minority Students: A Sociocultural Approach").

Before a child is involved in the screening or testing process, a parent interview helps determine the child's preferred language. For limited-English-proficient children, a speech-and-language specialist performs language-dominance testing. All observations, tests, and interviews are then administered in the child's dominant language. In the Somerville and Boston programs, testing personnel are available to test students in English and Spanish. For children who speak languages other than Spanish, native-language-speaking translators accompany the testers.

Teams that include administrators, facilitators, special educators, speech-and-language pathologists, occupational therapists, parent liaisons, and kindergarten teachers work in groups to develop each system's customized portfolio approach to screening and assessment. The basic framework for each school system's portfolio assessment includes:* observation of play behavior by two or three teachers and/or other school personnel;

• observation of group interaction using a formal checklist;
• a preschool screening or testing instrument;
• a parent questionnaire and follow-up interview;
• research on the child's culture of origin.

In school systems where this approach has been used, administrators, teachers, and specialists agree that children's assets or strengths are identified more often than in single-screening test efforts. The portfolio process also provides multiple perspectives on each child, thereby validating learning areas that may show culturally biased differences.

Multiple Languages, Multiple Assessments

The use of a portfolio approach to preschool screening may offer a means of assessment that can more carefully discern language from learning problems. If language-minority children are not mislabeled as special education students early, they stand a better chance of thriving in a mainstream academic environment later on.

Screening and standardized testing practices in the United States are often discriminatory, and the testing of language-minority students represents a special case of bias. Limited-English-proficient children who are screened and later tested using standardized measures are too often labeled deficient. Respecting language and cultural differences, as well as variability in child development, requires looking carefully at the strengths of immigrant children. Respecting the variability of child development and the realities of language and cultural differences means looking carefully at immigrant children for their strengths. Sometimes early differences in child development may appear similar to characteristics of some disabilities, but a more careful look can often lead to different conclusions.

Language-acquisition issues are key variables that need to be ruled out before educators determine whether a child has a disability. But can realistic assessments be performed with children like Jean-Yves on the first day they enter a school building? We need to allow children time in school-time to develop language, and time to understand the expectations of their teachers and their schools.

We also need to consider alternative approaches to teaching, learning, and assessment so that we can find the best in all children who attend our schools. At a minimum, a sociocultural framework for teaching and learning and a portfolio approach to assessment can help educators gather a variety of information about language-minority children and how they learn. Approaches like these are the only ways we will truly be able to teach and reach all children.

*Jean-Yves is a pseudonym.

Evangeline Harris Stefanakis is director of the Evelyn Pitcher Curriculum Laboratory and a faculty member in the Department of Applied Child Development at Tufts University. The child of Greek immigrant parents, Stefanakis began school at age five speaking virtually no English.