Research Article review

J Child Fam Stud (2007) 16:209–218 DOI 10.1007/s10826-006-9079-0

ORIGINAL PAPER

Assessment and Decision-Making in Early Childhood Education and Intervention

Paul S. Strand · Sandra Cerna · Jim Skucy

Published online: 23 August 2006 C© Springer Science+Business Media, Inc. 2006

Abstract Assessment within the fields of early childhood education and early childhood intervention is guided by the deductive-psychometric model, which is a framework for legitimizing constructs that arise from theories. An alternative approach, termed the inductive-experimental model, places significantly more restrictions on what constitutes a legitimate construct. In this paper, the utility of these two assessment models, one more generative and one more restrictive, are evaluated within the context of a Head Start setting. Given the pragmatic goal of informing instruction, we argue for the superiority of the more restrictive approach. Implications for early childhood intervention are also discussed.

Keywords Early childhood . Education . Intervention . Psychometrics . Curriculum-based assessment

The failures of the relief and rescue efforts in the wake of hurricane Katrina will undoubtedly generate a review of policies that guide how human services agencies serve their constituents. Questions will arise about the quality of the information they acquire and the extent that it contributes to effective decision-making and action. It is perhaps timely, therefore, to undertake such a review as it applies to agencies that serve young children and families. To preview, the dominant model for generating information within early childhood education and intervention is described and contrasted with an alternative model. The utility of these models are then evaluated with respect to ongoing attempts by one Head Start agency to develop assessment procedures that inform teaching. Although this case study concerns an agency devoted to educational outcomes, implications of these conclusions for all agencies serving children and families are discussed.

P. S. Strand (�) Department of Psychology, Washington State University Tri-Cities, 2710 University Drive, Richland, WA 99534, USA e-mail: [email protected]

S. Cerna · J. Skucy Benton Franklin Head Start, Richland, WA, USA

Springer

210 J Child Fam Stud (2007) 16:209–218

Traditional assessment and intervention

The traditional approach to assessment and intervention that predominates in early childhood education and intervention is based on the idea that more information is better information. This approach is illustrated by early childhood education and intervention models that en- courage “providing a complete view of the child’s strengths and capacities” (Meisels & Atkins-Burnett, 2000, p. 233), and also by mandates for increasingly comprehensive as- sessments within Head Start (DHHS, 2000). With respect to the latter, Head Start employs an outcomes framework that targets eight Developmental Domains that include: Language Development, Literacy, Mathematics, Science, Creative Arts, Social and Emotional Devel- opment, Approaches Toward Learning, and Physical Health and Development. Each domain comprises between two and five Domain Elements for a total of 27 Domain Elements, which are further broken down into a total of 100 Indicators. Head Start agencies are legislatively mandated to assess, three times yearly, four specified Domain Elements and nine specified Indicators. In addition, each agency is encouraged to assess at least five additional Domain Elements or Indicators within the unmeasured Domains. Evaluating 18 Elements and/or Indicators three times equates to 54 pieces of information gathered on each child per year. Given an average class size of 17, teachers are called upon to manage, directly or indirectly, 918 pieces of information arising from formal assessments alone. These assessments are in addition to mandated academic readiness and social competence screenings that occur for each child within the first 45 calendar days of the child’s attendance. The assumption under- lying this gargantuan assessment undertaking is that in order to offer effective educational and developmental programs, teachers need a great deal of detailed information about the children they serve.

But do they? Although the idea that more information is better information has intuitive appeal, it is contrary to 50 years of decision-making research (Grove, Zald, Lebow, Snitz, Nelson, 2000; Kleinmuntz, 1990; Sarbin, 1986). Beginning with the work of the eminent psychologist Paul Meehl (1954), numerous studies have illustrated that rather than improv- ing human decision-making, a plethora of information oftentimes impairs it. Computers running very simple algorithms have been shown to outperform well trained and experi- enced clinicians with respect to diagnostic decision-making. Importantly, the advantage held by computers lies not in the fact that they process more information, but in the fact that they are programmed to process less. Clinicians are outperformed not by supercomputers but by algorithms that reflect scores on very few variables.

What is it about attending to less information that makes it more effective than attending to more information? Computers outperform clinicians to the extent that clinicians take into account redundant information or information that is less relevant to the judgment being made. In many cases, the more information made available to clinicians the poorer their performance relative to the computer, although more information leads clinicians to perceive their judgments to be more accurate. Similarly, analyses of expert versus novice performance illustrates that experts outperform novices not because they can account for more information, but because they effectively limit the information they use. These analyses suggest that agencies might improve outcomes by implementing procedures that limit the information available to staff, as opposed to exposing them to a great deal of information and expecting them to sort it out.

It is difficult to imagine how a data gathering process like the one specified by Head Start, as an example, could improve teacher performance. Even if all the measured Indicators and Domain Elements account for unique variance with respect to child development (i.e., they are non-redundant), it is beyond human information processing capabilities to synthesize and

Springer

J Child Fam Stud (2007) 16:209–218 211

prioritize them all. Tellingly, Head Start requiress that these data be gathered, but provides no clear guidelines about how they are to inform educational practice. How did we get to the point where teachers, or early childhood interventionists, spend time and energy taking measurements that provide no clear path to action? Is there a model available that may limit the information we collect to that which is relevant to educational efforts?

Competing models of construct development and assessment

The traditional approach to construct development within early childhood education and in- tervention follows the psychometric tradition. According to Cronbach (1957), psychometric psychology is a deductive science in which constructs arise from theories about human abil- ities and are then given scientific form when methods for their measurement are generated. Importantly, the methods of psychometrics allow for testing the legitimacy of measurement tools, but are silent regarding the legitimacy of constructs. The legitimacy of constructs is established at the level of theorizing—prior to a psychometric analysis.

A theoretical rationale exists for the assessment of each variable targeted by Head Start (DHHS, 2000). Having derived these variables from theory, the goal is to identify measure- ment techniques that illustrate adequate reliability and validity. In addition, factor analytic techniques may be used to identify redundancy with respect to how the constructs relate to one another, thereby reducing the number of variables studied. Nevertheless, even if a manageable number of variables remain, will they be relevant to teachers? It is very possible that they will not. The reason is that the variables have been chosen largely for theoretical reasons rather than for practical reasons. For example, even if only one vari- able having to do with motivation, cognitive ability, fine motor coordination, and language skills, respectively, survives a factor analysis, it may be that none have clear implica- tions for teacher behavior. The measured variables may tell us a great deal about child development without providing information relevant to classroom design, management or instruction.

In contrast to psychometrics, Cronbach (1957) identified experimental psychology as the inductive side of psychology. Rather than accepting constructs that flow from the top (i.e., ideas and general observations), the experimental tradition permits constructs on the basis that they illustrate functional relationships in the context of experimental manipulation. That is, experimentalists are interested in manipulating variables and exploring the effects of such manipulations on other variables. The presentation of these relationships usually takes the form of graphical displays regarding how changes in one variable are a function of changes in another variable. Traditionally, therefore, experimentalists do not put great weight in the reliability and validity of some measure unless that construct has been legitimized with respect to a functional analysis (Donahoe & Palmer, 1994).

To summarize, many constructs that meet the requirements of a deductive-psychometric approach to construct formation do not meet the criteria set by an inductive-experimental approach. In this way, the inductive-experimental approach has natural limitations on variable creation that a deductive-psychometric approach does not.

Functional control as criterion for construct legitimacy

Most educators and clinicians influenced by the inductive-experimental tradition focus on behavioral repertoires that are component elements of desirable complex behaviors, such as

Springer

212 J Child Fam Stud (2007) 16:209–218

reading, problem-solving and rule following (Ducharme, 1996; Nelson, Benner, & Gonzalez, 2005; Shinn, 1989; Strand, Barnes-Holmes & Barnes-Holmes, 2002). A primary assumption arising from this tradition is that mastery in the form of fluency with respect to component skills gives rise to higher-order skills. Oftentimes, what constitutes a functionally important component skill is surprising in its simplicity relative to the higher-order skill (Johnson & Street, 2004). For example, few would suspect that something as seemingly mundane as improving the speed and accuracy of saying and writing numbers could improve performance with respect to higher-order mathematics. Nevertheless, this functional relationship was observed among college students struggling with calculus (Haughton, 1980, cited in Johnson & Layng, 1992). Similarly, compliance on difficult tasks such a picking up toys is improved to the extent that compliance is established with respect to easy tasks (Strand, 2000). Therefore, the inductive-experimental approach utilizes experimental analysis to identify functional relationships between what are thought to be component skills and higher-order skills. Importantly, a component skill is defined as such via the experimental method, and not based on appearances or a priori theoretical considerations (i.e., deduction).

An apparent drawback of this approach is the paucity and simplicity of variables consid- ered legitimate. Might such an austere approach lead to a poverty of legitimate constructs? After all, things that seem intuitively related to child development and welfare may be ignored within such a model. For example, the concept of attachment would be deemed irrelevant to the extent that it could not be manipulated or evaluated repeatedly over time (Gewirtz & Pelàez-Nogueras, 2000). Moreover, variables that may appear simplistic are sometimes thor- oughly investigated from this perspective. In this way, the inductive-experimental perspective takes seriously the primary proposition that guides the natural sciences—simple processes, occurring at certain frequencies and in certain combinations, give rise to complex phenom- ena (Novak & Pàelez, 2002; Thelen & Smith, 1994). In addition to being philosophically consistent with natural science, the approach is appealing because it may prove necessary to overcome problems arising from the limitations of human information processing.

Transitioning to an inductive-experimental assessment strategy

In our experience, attempts to utilize the Developmental Domains to inform teaching have thus far failed. Originally the challenges seemed primarily logistical and included how to gather, score, and distribute data to teachers in a timely manner. Relatedly, faced with so much data, we needed to prioritize it. This latter issue did not worry us too much because it was our intention to conduct factor analyses to identify a list of non-redundant variables. Despite having a strategy for identifying what data would be useful, we still struggled with the problem of how they would be useful. Unfortunately, because a variable explains variance in regression or structural equation models does not mean that it is useful with respect to clinical or educational planning and action.

Another activity mandated by Head Start involves the assessment of certain pre-academic skills, including letter recognition and one-to-one correspondence (object counting). The process for measuring letter recognition involves asking children to identify as many as they can of the 26 letters of the alphabet as they appear on a set of three cards. One-to-one correspondence is measured by having children count objects such as stars or triangles as they appear printed on a card in five rows of five. Both of these constructs meet the legiti- macy criteria of the inductive-experimental approach in that they: (1) represent functional, component skills of identified higher-order constructs, (2) have been shown in controlled situations to be sensitive to changes in environmental conditions (i.e. they can be manipulated

Springer

J Child Fam Stud (2007) 16:209–218 213

through practice), and (3) are amenable to repeated measurement and graphical display. In the next sections of the paper we will describe how, by beginning with only one of these two constructs—letter identification—we hope to generate a comprehensive assessment-based intervention program for preschool education. Importantly, it is our belief that the long-term success of such a program hinges on the extent to which it allows individual teachers max- imum freedom with respect to curriculum development, within the confines of achieving measurable progress on pre-defined pre-academic skills.

Assessment and educational decision-making

Having concluded that the Developmental Domains provide little guidance to teachers, we began to explore in more detail the implications of repeated assessments of constructs that represented components of desired higher-order skills (Shinn, 1989). Because of the increased emphasis at a nationwide level on literacy readiness, improving letter recognition scores became a focus for this Head Start agency. As part of this focus, a phonics-based instructional program was implemented in 4 of 22 classrooms at the beginning of the school year. The program was implemented in the remainder of the classrooms beginning in March.

Figure 1 presents Letter Recognition results over the two-year period 2003–2004 and 2004–2005. The x-axis reflects the number of letters known by children towards the beginning of each school year (in October), and the y-axis reflects the percentage of children who recognized 10 or more letters toward the end of the school year (May of the next year). The results show that for both academic years, the percentage of children who met the 10- letter goal in May was predicted by the number of letters the child recognized previously. Moreover, they illustrate that across all levels of initial letter recognition, children in the 2004–2005 school year had better year end letter recognition scores.

Initially, one might attribute the between-year differences to the introduction of the phonics program in 2004–2005. However, a comparison in January revealed no differences

24%

45%

79% 83%

11%

20% 20%

60%

70%

94%

0%

20%

40%

60%

80%

100%

0 Letters 1 Letter 2 Letters 3 Letters 4-9 Letters

C h

ild re

n R

ec o

g n

iz in

g 1

0 L

et te

rs

2003 2004

Fig. 1 Percent of children who identified 10 letters at the end of the school year as a function of the number of letters identified at the beginning of the school year for academic years 2003–2004 (n = 154) and 2004–2005 (n = 128)

Springer

214 J Child Fam Stud (2007) 16:209–218

in letter recognition scores for children not yet exposed to the phonics program and those exposed to it in October. Moreover, end-of-year scores did not differ across classrooms as a function of phonics exposure. We concluded, therefore, that the improved performance reflected an agency-wide commitment to improving letter recognition skills—of which the implementation of the phonics program was only a part.

Presentation of these data at an agency-wide meeting had an energizing effect for agency personnel. What was important to teachers was not whether the results supported or did not support a phonics-based intervention for preschoolers. Rather, they were taken by the potential usefulness of exploring the development, over the course of the school year, of a skill that was shown to be functionally related to changes in the classroom environment. Practical applications of these data were readily apparent. For example, the data revealed that initial assessments provide information about the probable learning trajectory of individual children, and that these trajectories are sensitive to the learning environment. Furthermore, they suggest that to attain certain goals some children would require more letter recogni- tion instruction than others—and this could be predicted in advance. Therefore, these data provide information relevant to individualizing instruction. As a result, teachers welcomed the suggestion that frequent assessments of letter recognition be conducted over the course of the upcoming school year because such assessments would allow them to titrate specific curriculum elements depending on the progress of individual children.

Of course, comparisons like those presented in Fig. 1 could also be made for repeated measures of Developmental Domain scores or other psychometrically-based variables. How- ever, the power of letter recognition as a variable involves the extent that it is functionally related to particular educational interventions. The assessment process inspires speculation about potential changes to the preschool environment, which is something that is under the control of teachers. Psychometric data, on the other hand, typically inspires speculation about innate differences and the adequacy or inadequacy of home environments—neither of which is under the control of teachers.

In addition to allowing for changes to the educational curriculum as a result of the progress made by individual children, frequent assessments of skills such as letter recognition allow for evaluating the effects of different curricula on groups of children. For example, Fig. 2 illustrates letter recognition scores as a function of time for children from different language backgrounds exposed to a phonics program over a 6-month period. These data provide information relevant to frequently asked questions about learning differences across groups and also the differential effects of direct instruction teaching methods. First, these data are consistent with previous studies reporting that phonics instruction may be most helpful for children at highest risk for school failure (Ehri, Nunes, Stahl, & Willows, 2001; McConnell, 1982; Nelson et al., 2005). In this case, we see that Spanish-speaking children “catch up” by the end of the term. Second, these data show how time factors into differential performance across these language groups. That is, the graph reveals that the curves for Spanish-speaking compared to English-speaking children look quite similar over the first seven assessments. That is, both groups improved steadily, and the slight differences observed toward the beginning of the year are maintained until sometime after Winter break (between T6 and T7). Interestingly, the performance decline brought about by Winter break is steeper for the English than it is for the Spanish speaking children. After Winter break, the curves show differences, with improvements for Spanish-language children being relatively steeper than for English-speaking children, to the point that, as a group, Spanish speakers out perform English speakers on letter recognition skills at T10.

These data were well received by teachers anxious to identify methods for improving the relative performance of non-English-speaking children, and also for assessing learning

Springer

J Child Fam Stud (2007) 16:209–218 215

0

5

10

15

20

25

T1 T2 T3 T4 T5 T6 T7 T8 T9 T10

A ve

ra g

e N

u m

b er

o f

L et

te rs

R ec

o g

n iz

ed

English

Spanish

Fig. 2 Letter recognition scores for English-speaking (n = 45) and Spanish-speaking (n = 24) preschoolers over the course of 6 months of phonics instruction

trajectories for all children over the course of the school year. Teachers recognized that these data could serve as a benchmark for exploring the letter recognition skills development of children in future years. Unlike normative data gathered on national samples, these data are clearly representative of the population served by this agency and, therefore, represent a valid point of comparison for future samples.

Given that teachers are most interested in data gathered at their own agency, it is our belief that data such as these may serve another important function with respect to improving the educational environment. Specifically, exposure to this type of data—more so than hours of in-service training—has the potential for dispelling the crippling effects on disadvantaged and minority children of low teacher expectations (see Rist, 2000). That is, after examining Fig. 2, it is hard to maintain the belief that the Spanish-speaking children must necessarily lag behind their English-speaking peers with respect to letter recognition skills development. Any agency for which minority children do so, and that implements a tracking process such as that described above, is in a position to intervene to reverse such trends. Stereotypes are best combated with data rather than rhetoric.

Feedback as intervention

A working hypothesis that will guide our future efforts is that, given clear feedback about child progress toward objective goals, teachers will oftentimes succeed in generating methods for achieving those goals. But unless teachers can see what is happening with respect to learning, they will be unable to act in effective ways. Said differently, given timely and relevant feedback about child progress toward specified goals, teachers will identify the means for achieving those goals. In all likelihood, however, different teachers will generate different methods. Nevertheless, while appearing to differ in terms of content, these different methods may be functionally equivalent in that they produce similar outcomes.

A corollary of this assumption is that only for children for whom a teacher’s chosen methods are failing to generate desired outcomes will agency personnel encourage alternative educational plans. Moreover, planning will occur with respect to the observable progress of individual children, and not with respect to fidelity to popular curricula. That is, to the extent

Springer

216 J Child Fam Stud (2007) 16:209–218

that children are meeting specified goals, teachers have great freedom with respect to how they structure classroom activities. Given productive feedback, teachers should naturally gravitate toward intervention programs that work. Moreover, when even the best curricula fail in a given circumstance, ongoing assessment data allow for generating alternative plans in a timely manner. By timely, we refer to curriculum changes that can be assessed for effectiveness within weeks rather than months of implementation.

That performance may be enhanced by simply providing timely feedback to teachers is a testable hypothesis. It is our plan to test it with respect to one-to-one correspondence skills by providing bi-monthly feedback to teachers for each child regarding this skill. No specific math curriculum or training has been implemented, although teachers have been made aware that local school districts and national Head Start have identified counting 10 objects as minimum competence for Kindergarteners. It is our belief that providing teachers with an expectation and frequent feedback about performance will result in improved performance. Such a finding would suggest that, as is the case with other aspects of human performance (i.e. weight loss, smoking, gambling), frequent feedback that tracks performance improves performance (Rachlin, 2000). Although this is likely not true with respect to the teaching of some skills or for poorly trained teachers, we believe that it will hold true with respect to the teaching of object counting by skilled teachers. Such a finding would illustrate that educational improvements may be achieved by implementing feedback mechanisms with respect to teachable skills.

Assessment and early childhood intervention

Although the example we have presented is concerned primarily with early childhood edu- cation, the principles of the inductive-experimental approach apply also to early childhood intervention. Indeed, the drawbacks of the deductive-psychometric approach to assessment have been noted within that field, and efforts are underway to establish approaches to assess- ment that are influenced more by ecological validity concerns rather than concerns about reliability and construct validity (Meisels & Atkins-Burnett, 2000). Nevertheless, there ap- pears to be little or no call for limiting assessment within clinical and agency settings to constructs that may be subject to manipulation. Take, for example, the description of what constitutes some of the abilities deemed relevant with respect to early childhood intervention:

“Can the child get around on different floor surfaces? Does the child play and explore on calm days but tend to sit, watch, and suck his or her thumb when things get confusing? Can the child draw a circle on a horizontal surface but not on an easel? Does the child need to master a simple gestural symbol system in order to communicate with peers, although family members respond to different, highly familiar cues from the child? Answers to these questions lead to a more differentiated view of the child’s skills and resources, and ultimately to a more individualized set of interventions” (Meisels & Atkins-Burnett, 2000, p. 235).

Clearly, the assumption of this quotation is that assessing these variables will lead to better intervention outcomes. But what assurance do we have that that is true? Our only guarantee comes from studies in which attempts are made to manipulate these variables. Therefore, to the extent that these variables have not been shown to be responsive to intervention, their measurement is of questionable value in clinic or agency settings. That does not mean, however, that their measurement is questionable in all settings. For example, such variables should be explored by researchers interested in identifying the extent to which they may be

Springer

J Child Fam Stud (2007) 16:209–218 217

manipulated, and with respect to the possibly wide-ranging effects of such manipulation. Such efforts represent attempts to evaluate functional utility, and should be undertaken within naturalistic settings. Nevertheless, because a construct is a legitimate focus of research does not make it legitimate with respect to intervention.

Feedback in early childhood intervention

The challenges of implementing an inductive-experimental assessment procedure for early childhood intervention may be greater than they are for early childhood education. That is so because the component skills underlying academic competence are easier to identify and measure than are the component skills underlying healthy social and family functioning. Nevertheless, developing an assessment program based on constructs that are functionally related to intervention may provide a framework for focused, evidence-based interventions in agency settings (Repp & Horner, 1999).

Fortunately, a broad research base points to a set of infant and toddler social behaviors that have been shown, within naturalistic contexts, to be functionally related to interventions (Novak & Pelàez, 2004). These variables include, but are not limited to, smiling, verbal behavior, aggression, tantruming, imitation, and social cooperation. In addition to child variables, caregiver and family variables may be important targets as well (Cavell & Strand, 2002; Wahler, 1997). To the extent that these variables can be frequently assessed, they could serve as the basis for an inductive-experimental approach to assessment and intervention.

Conclusions

A problem with assessment as it is currently practiced in educational and agency settings is that much information is collected and disseminated that has no clear implications for intervention. It is our belief that the treatment utility of assessment is maximized to the extent that information is limited to that which has clear implications for intervention. Moreover, frequent feedback regarding performance on a limited set of manipulable variables may improve teacher and clinician performance, in lieu of directly altering educational or clinical practices. In this way, a procedure that is restrictive with respect to assessment may afford teachers and clinicians greater freedom with respect to curriculum development and intervention. Under these circumstances, we may find that teachers and clinicians generate a wide array of effective, individualized, early childhood interventions.

References

Cavell, T. A., & Strand, P. S. (2002). Parent-based interventions for aggressive, antisocial children: Adapting to a bilateral lens. In L. Kuczynski (Ed.), Handbook of dynamics in parent-child relations. Thousand Oaks, CA: Sage.

Cronbach, L. J. (1957). The two disciplines of scientific psychology. American Psychologist, 12, 671–684. Department of Health and Human Services, Administration on Children, Youth, and Families/Head Start

Bureau. (2000). The Head Start Path to Positive Child Outcomes. Washington, DC: Author. Donahoe, J. W., & Palmer, D. C. (1994). Learning and complex behavior. Boston: Allyn & Bacon. Ducharme, J. M. (1996). Errorless compliance training: Optimizing clinical effectiveness. Behavior Modifi-

cation, 20, 259–280. Ehri, L. C., Nunes, S. R., Stahl, S. A., & Willows, D. M. (2001). Systematic phonics instruction helps students

learn to read: Evidence from the national reading panel’s meta-analysis. Review of Educational Research, 71, 393–447.

Springer

218 J Child Fam Stud (2007) 16:209–218

Gewirtz & Pelàez-Nogueras, (2000). Infant emotions under the positive-reinforcer control of caregiver at- tention and touch. In J. C. Leslie & D. Blackman (Eds.), Experimental and applied analysis of human behavior (pp. 271–292). Reno, NV: Context Press.

Grove, W. M., Zald, D. H., Lebow, B. S., Snitz, B. E., & Nelson, C. (2000). Clinical versus mechanical prediction: A meta-analysis. Psychological Assessment, 12, 19–30.

Harbin, G. L., McWilliam, R. A., & Gallagher, J. J. (2000). Services for young children with disabilities and their families. In J. P. Shonkoff, & S. J. Meisels (Eds.). Handbook of early childhood intervention. (2nd ed.; pp. 387–415). New York: Cambridge University Press.

Haughton, E. C. (1980). Practicing practices: Learning by activity. Journal of Precision Teaching, 1, 3–20. Hayes, S. C., Nelson, R. O., & Jarrett, R. B. (1987). The treatment utility of assessment: A functional approach

to evaluating assessment quality. American Psychologist, 42, 963–974. Johnson, K. J., & Layng, T. V. J. (1992). Breaking the structuralist barrier: Literacy and numeracy with fluency.

American Psychologist, 47, 1475–1490. Johnson, K. J., & Street, E. M. (2004). The Morningside Model of generative instruction: What it means to

leave no child behind. Concord, MA: Cambridge Center for Behavioral Studies. Kleinmuntz, (1990). Why we still use our heads instead of formulas: Toward an integrative approach. Psy-

chological Bulletin, 107, 296–310. McConnell, B. B. (1982). Bilingual education: Will the benefits last? Bilingual Education Paper Series, 5 (8).

Los Angeles: Evaluation, Dissemination and Assessment Center, California State University. Meehl, P. E. (1954). Clinical versus statistical prediction. Minneapolis: University of Minnesota Press. Meisels, S., & Atkins-Burnett, S. (2000). The elements of early childhood assessment. In J. P. Shonkoff,

& S. J. Meisels (Eds.). Handbook of early childhood intervention (2nd ed.; pp. 231–257). New York: Cambridge University Press.

Nelson, J. R., Benner, G. J., & Gonzalez, J. (2005). An investigation of the effects of a prereading intervention on the early literacy skills of children at risk of emotional disturbance and reading problems. Journal of Emotional and Behavioral Disorders, 13, 3–12.

Novak, G., & Pelàez, M. (2004). Child and adolescent development: A behavioral systems approach. Thousand Oaks, CA: Sage Publications.

Rachlin, H. (2000). The science of self-control. Cambridge, MA: Harvard University Press. Repp, A., & Horner, R. H. (Eds.) (1999). Functional analysis of problem behavior: From effective assessment

to effective support. Belmont, CA: Wadsworth. Rist, R. C. (2000). Student social class and teacher expectations: The self-fulfilling prophecy in ghetto

education. Harvard Educational review, 70, 266–301. Sarbin, T. R. (1986). Prediction and clinical inference: Forty years later. Journal of Personality Assessment,

50, 362–369. Shinn, M. R. (Ed.). (1989). Curriculum-based measurement: Assessing special children. New York: Guilford. Strand, P. S. (2000). A modern behavioral perspective on child conduct disorder: Integrating behavioral

momentum and matching theory. Clinical Psychology Review, 20, 593–615. Strand, P. S., Barnes-Holmes, Y., & Barnes-Holmes, D. (2002). Educating the whole child: Implications of

behaviorism as a science of meaning. Journal of Behavioral Education 12, 105–117. Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to the development of language and cognition.

Cambridge, MA: Bradford Books. Wahler, R. G. (1997). On the origins of children’s compliance and opposition: Family context, reinforcement,

and rules. Journal of Child and Family Studies, 6, 191–208.

Springer

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (None) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (ISO Coated) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Perceptual /DetectBlends true /ColorConversionStrategy /sRGB /DoThumbnails true /EmbedAllFonts true /EmbedJobOptions true /DSCReportingLevel 0 /SyntheticBoldness 1.00 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveEPSInfo true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts false /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 150 /ColorImageDepth -1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /ColorImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 150 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.76 /HSamples [2 1 1 2] /VSamples [2 1 1 2] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 600 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org?) /PDFXTrapped /False /Description << /DEU <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> /ENU <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> >> >> setdistillerparams << /HWResolution [2400 2400] /PageSize [2834.646 2834.646] >> setpagedevice