In the late nineteenth century, the European scientific community was consumed with a fervor and creativity best characterized by the studies of evolution, theories about the unconscious mind, and a growing concern for the mentally deficient and insane. The field of experimental psychology was largely concerned with the measurement of various perceptual abilities, and the science of child development was dominated by single-child case studies— often the children of the scientists. Although these “baby biographies” were insufficient for benchmarking normal child development, they provided the basis for the creation of development tests during the twentieth century.

The concept of measuring infant capacities grew out of the concern of scientists of the time to find a metric for human intelligence that would permit the creation of criteria for schools for children with intellectual deficiencies (5). In 1904 the minister of public education in Paris appointed Binet and Simon to be members of a commission studying the question of special education. In response to their charge, Binet and Simon developed the concept of mental age and a test, published in 1905, for measuring it (3). Terman (6) created an adaptation of the Binet-Simon test for use in the United States and revised the concept of mental age by creating the intelligence quotient (IQ), a ratio between mental age and chronological age multiplied by 100.

In the decades that followed the development of Terman’s IQ test, two basic approaches were used in exploring intellectual development in infants. One approach, the infant IQ model, sought to extend downward the IQ assessment model to children younger than school age, eventually including infants. A second approach, the Gesellian model, sought to create a new model of assessment that began with the newborn and extended upward. In contrast to the “baby biographers” of the previous century, the more behaviorally empirical scientists of the 1920s and ’30s relied on direct observation of child behavior across several children under highly structured conditions.

The best example of an infant IQ or infant cognition test is the Cattell Infant Intelligence Scale (7). Cattell’s test was conceptualized as a downward extension of the 1937 Stanford-Binet Intelligence Scale, the much-revised American version of the Binet-Simon test. Although some initial studies were encouraging, the ability of the Cattell and other infant IQ tests to predict an infant’s later IQ was weak (8). Overall, “infant IQ tests” were found to be inadequate at meaningfully extending the IQ assessment model into infancy and are now all but extinct in use.

At about the same time that first “infant IQ tests” were being developed, Arnold Gesell began his groundbreaking work at Yale University on systematically documenting normal maturational development in infants and toddlers, leading to the creation of the Gesell Developmental Schedules (GDS), first published in 1925. Several aspects of Gesell’s work distinguish it from those of his predecessors and contemporaries. First, whereas the infant IQ tradition grew from an interest in identifying deviant patterns of development, Gesell’s interest was mostly in documenting normal developmental trends. Second, Gesell’s model supported the conceptualization of development occurring simultaneously in many distinct but interrelated domains (9), as opposed to a singular factor of intellectual ability. Third, Gesell’s model of developmental maturation supported an understanding of the effects of the child’s environment in altering the course of development, as opposed to viewing IQ as a stable and static trait, opening the way for the more transactional and dynamic understanding of developmental processes that would later prevail.

Gesell’s work was farreaching, and subsequent researchers relied heavily on the developmental tasks and techniques he pioneered. Most notable was Nancy Bayley of the Berkley Institute of Child Welfare. Bayley applied testing concepts more familiar to the infant IQ model to the assessment techniques developed by Gesell. The result was the development of two scales of infant development that would later revolutionize and dominate the field of infant assessment: the California First-Year Mental Scale10,11 and the California Infant Scale of Motor Development 12,13.

In the 1940s and 1950s, the science of infant developmental assessment gained greater international attention, through the publishing of the Griffiths Mental Development Scale(14) in London. At that time, infant assessment techniques were used primarily for diagnosis and categorization purposes, e.g., preadoptive screening, testing for admission to special schools, or evaluating physical handicaps (15). The assessments for diagnosis and categorization were predicated on clinicians’ continuing adherence to the belief that intelligence is fixed from infancy. This belief in a fixed IQ still was so well rooted that evidence to the contrary (findings documenting an increase in IQ in high-risk infants after nursery school attendance) was being dismissed and attributed to poor standardization of the scales (3). In the context of this debate, American developmentalists first began to consider seriously the stage theory of Jean Piaget, which clearly favored the view of qualitative, rather than simply quantitative, differences in the abilities of children of various ages. It was largely on the basis of Bayley’s longitudinal work with her developmental instruments (16) that the concept of an immutable IQ that was fixed at birth was discredited. The use of infant IQ tests continued for a while, but are now all but extinct in use.

The 1960s and 1970s brought a new wave of infant development tests, more rigorously standardized on larger numbers of infants, with careful testing of interobserver agreement and test-retest reliability. The Bayley Scales of Infant Development (BSID)(17) consolidated Bayley’s original two scales into one assessment instrument with norms based on a nationally representative sample of infants. The BSID set a new and enduring standard of sophistication for the development and standardization of infant developmental assessment tools. Thoughtful research led to far more caution about the predictive validity of early assessment and to important conceptual revisions. Concomitant interest in newborn capacities and the rapidly emerging field of newborn sensory perception also led to the development of a number of scales to measure competency in newborns.

The 1980s and 1990s were dominated by increasing psychometric improvements; a renewed interest in diagnostic functionality, again primarily driven by the needs of public education; and the application of information-processing theory to the study of infant development. Federal mandates for the special education of young children created the need for an arsenal of assessment tools that encompassed all of the qualifying areas of delay, utilized information from both direct assessment of the child and parent report, and facilitated early intervention treatment planning and program accountability. Also, a plethora of brief developmental screening instruments
also were developed for use by professionals with relatively little training in formal assessment.

The advent of information-processing theories of human intelligence led to more elaborate models of infant cognition. Several studies have shown measures of infant information-processing (e.g., attention, recognition, stimulus habituation, and memory) to be significantly related to later IQ scores (several excellent reviews are available 18,19). However, these measures can be difficult to administer under usual clinical assessment conditions, and the model has not been translated to clinical assessment tools. Regardless, measures of information-processing may play an important role in the next generations of developmental assessment instruments.

It is axiomatic that skillful interviewing is central to a complete developmental assessment, since much of the data about infants’ daily functioning and their relationships with others come from interviews with the caregivers. Skillful interviewing techniques include letting caregivers begin their story wherever they choose; using directed, information-gathering questions in such a way as to clarify but not disrupt the parents’ account; and listening for affect as much as content. Importantly, nearly every step of the assessment process requires an alliance between clinician and caregiver, since infants usually perform better when they are in the company of familiar adults, and the initial interview between clinician and family is crucial in setting the tone for such an alliance. Moreover, establishing an alliance is central to evaluating infants’ interactions with the adults in their world. Indeed, infant assessments are quite compromised when there are no familiar adults available to meet with the clinician and be with the infant. Parenthetically, it is often in cases involving the most severe environmental disturbance that clinicians do not have access to caregivers that are able to describe the infant’s history.

Provence (21) has suggested that a productive method of gathering developmental and family data is to ask the caregivers to describe a day in the life of their child. Provence outlines how this question can be the framework for learning about daily activities, how the infant and caregivers interact throughout the day, and about interactions around mealtime, bedtime, or times of distress. When all major caregivers are present for the interview, this question provides a time for each of them to present descriptions of his or her time with, and perceptions of, the child. Additionally, clinicians may use structured interviews, such as the Vineland Adaptive Behavior Scales(22) currently under revision, in order to collect both quantitative developmental data and provide an opportunity to open new areas of clinical discussion. Also, the Infant-Toddler Developmental Assessment (IDA)(23), appropriate for infants from birth to 36 months, is particularly useful in providing schemata for organizing important information from caregiver interviews, medical/developmental records, and behavioral observations.

Implicit in this overview of interviewing caregivers as a part of the clinical assessment is the assumption that such assessments require several sessions. Minimally, one meeting with caregivers, two or more with the child and caregivers together, and another to present the results to the caregivers are necessary. The sessions with the infant also provide an opportunity to gather more interview information, as other questions will occur in the context of the child’s behavior and performance. For example, asking whether the child’s response to a particular situation within the evaluation context is usual for him or her may open up another area of information from the caregivers. As is likely clear from these suggestions, in our view, infant assessment is a process of constantly gathering information, revising impressions, and testing hypotheses— and that requires time.

On the other hand, clinicians asked to evaluate infants and young children will not always have the ideal situation outlined above. At times, consultants, caregivers, or both may insist that the evaluation be done in one session, or the clinician may have limited access to the child, as with many evaluations done in a hospital setting. It is important at these times for the clinician to be clear about the limitations of the evaluation findings. Another situation that occurs increasingly commonly and does not fit the ideal model just outlined is when no parents, family members, or other caregivers are available. Situations of severe abuse, abandonment, multiple placements, or seriously ill parents are examples of times in which the clinician will not have available certain critical sources of information. In these instances, certain hypotheses suggested by the child’s presentation and status may be left unconfirmed. As in situations where the time for the evaluation is brief, it is most important for clinicians to acknowledge which aspects of their diagnostic formulation are relatively certain, which are not, and what information would likely be clarifying, were it available.

Observation is the fundamental skill needed for measuring infants’ development. After all, most diagnostic evaluations are based on observation of physical signs and/or behavioral responses. However, what distinguishes the observational skill necessary for developmental assessment is that it occurs on many levels simultaneously and is perhaps the area in which the developmentalist’s dual role as both generalist and specialist is most evident. Moreover, the observational skills inherent in assessments of infants require a blend of free-floating attention bounded by a structure. In other words, while the clinician must be comfortable enough in the setting to attend to whatever occurs, he or she must also have a mental framework by which to organize the observations collected during the session. Such a framework entails at least four broad areas: a) predominant affective tone of the participants, b) involvement in the situation (curiosity and interest), c) use
of others (child’s use of the caregivers or examiner), and d) reactions to transitions (initial meetings, end of sessions, changes in amount of structure).

The process of synthesizing all the data gathered from the different sources during an assessment is a technique and skill unto itself. Moreover, how this synthesis, with its attendant recommendations, is conveyed to caregivers and other professionals is another essential step in the assessment process, and the assessment is not complete until the therapeutic alliance among all stakeholders is brought to fruition in a collaborative formulation. Infant assessments often involve referring pediatricians and other clinicians, all of whom need to be included individually in the clinician’s datagathering interviews and in the final synthesis.

The synthesis of information from an infant assessment differs from the synthesis involved in other medical diagnostic processes in that there are very few specific diagnostic categories that encapsulate all the findings of an infant assessment. The synthesis involved in an infant assessment requires a bringing together of all the data gathered from interviews, observations, and testing into a qualitative description of that infant’s capacities in different functional areas (motor, problemsolving, language and communication, and social) and of the infant’s current strengths and weaknesses. It also involves integrating the assessment information in the context of the infant’s individual environment. For example, an infant who has experienced multiple foster placements may be socially delayed, but such a finding may assume a different significance for an infant who has had a stable home environment.

Synthesizing the large amount of data obtained from a comprehensive clinical infant assessment can be quite daunting. As stated in the introduction, in performing the infant assessment the clinician must be both generalist and specialist. The clinician must draw upon and synthesize knowledge from child
psychiatry, pediatrics, neurology, developmental psychology, speech/language therapy, physical and occupational therapy, and often genetics and endocrinology (28). Increasingly, clinicians evaluating young children need also know about early childhood education programs and early intervention, as well as laws regarding child abuse, neglect, and domestic violence. Knowledge from these diverse fields allows a clinician to place the results of a developmental assessment in a meaningful context for the individual child and leads to a better conceptualization of treatment options. For example, understanding the physiological effects of prolonged malnutrition and episodic starvation in infancy 29,30 helps the clinician evaluate the relatively greater gross motor delays of a child with failure to thrive who has no other neurological signs. Similarly, understanding the effects of a parent’s affective disorder on a child’s responsiveness to the external world (31) adds another dimension to understanding the infant’s muted or absent social interactiveness, babbling, and smiling.

Finally, it is often during the synthesis process that the therapeutic effect for caregivers participating in the assessment is most evident. At the very least, caregivers often change their perceptions of their infant’s capacities. They may see strengths in their infant they had not previously recognized or become deeply and painfully aware of weaknesses and vulnerabilities that they may or may not have feared before the assessment. Any of these changes in perceptions may affect the caregivers’ view of themselves and of their role as caregivers. Also, infants often change during the assessment process, as their caregivers become more involved in the alliance with the clinician, and they experience, at least temporarily, another adult’s concern and interest in their family. Emphasizing the potentially therapeutic value of an assessment underscores that the synthesis process is not simply wrapping up the assessment and conveying information, but is also a time to explore with the caregivers the meaning of the process for them and their infant.

Besides the vast amount of qualitative data obtained by careful observation during testing, most developmental assessment procedures provide several methods of quantifying results. As described later in this chapter, most developmental tests have been standardized and normed on a sample of children selected to represent the performance of typically developing children. It is by comparing an individual infant’s performance against this set of norms that most test scores are derived. These scores are then used to convey something about the infant’s level of developmental skills acquisition, relative to other presumably typical infants.

Of the several different types of scores available, standard scores and age equivalents are the most commonly used. Standard scores are the most robust scores obtainable and represent the infant’s performance in relation to other infants of about the same chronological age. Common forms of standard scores include Z-scores, deviation scores, and T-scores, the latter two being linear transformations of the former. Most tests provide some guidance on how to describe an infant’s standard score by placing the scores in descriptive bands (average, below average, mildly delayed). Often, in order to increase the interpretability of these scores, they are converted to percentile ranks. Table 4.2.1.2 presents one common method for descriptively banding standard scores, although many other methods also are used.

Age equivalent scores represent an estimate of the chronological age (usually expressed in months) at which the typically developing infant would demonstrate the skills observed in the infant being assessed. This type of score is often appealing to many caregivers and others with little or no training in psychometric tests, since the interpretation is seemingly straightforward. However, age equivalent scores are notoriously easy to misinterpret and may lead to erroneous conclusions. First, age equivalent scores on infant developmental tests tend to be highly unstable, with the infant’s performance on only one or two items largely affecting the age equivalent. Second, age equivalent scores may imply too much about an infant’s development, especially when their skills are highly scattered. Because of the way in which age equivalent scores are computed, the score may greatly over- or underestimate an infant’s developmental level. Due to the way in which tests are scored, infants who are very consistent in their ability to do tasks typical for their age often receive an age equivalent score much higher than their chronological age, even though they are not yet developmentally ready to do the things associated with that age equivalent. Conversely, infants who, for whatever reason, struggle with some less developmentally mature tasks likely will receive an age equivalent score much below their chronological age, even if they are able to do many things typical of their chronological age or older. In cases such as these, age equivalents are best expressed as a range, if expressed at all. Third, developmental delay expressed in age equivalents does not adequately address the frequency or severity of the delays. Clearly, a 6-month delay in development is different for a 12-month-old versus a 30-month-old. Standard scores, with associated percentile ranks, better express this delay in terms that are less dependent on the chronological age of the infant. Finally, the type of data used to compute age equivalents is too weak statistically to support the calculation of confidence intervals that are useful in placing the score within the context of appropriate bands of error. This lack of confidence intervals further exacerbates the preceding limitations. Of course, there are times when age equivalents may be the best or only option
available for expressing an infant’s performance (when the extent of developmental delay is so great that standard scores cannot be easily calculated and percentile ranks below the first percentile are obtained).