This is the old United Nations University website. Visit the new site at http://unu.edu
Abstract
Objectives
Data source
Results
Discussion and conclusions
Acknowledgements
References
Ernesto Pollitt and Nina Triana
The authors are affiliated with the Section of Pediatric Nutrition, Department of Pediatrics, in the School of Medicine at the University of California, Davis, California, USA.
This paper documents the stability, predictive power, and sensitivity of mental and motor development scales and pre-school cognitive tests in the context of economically impoverished populations in low-income countries. Stability and predictive power comprise forecasting; stability includes repeated measures using the same test, whereas predictive power includes different tests. Sensitivity is the track record of the test in discriminating among groups of children exposed to different nutritional interventions. Psychometric data from three longitudinal studies of the assessment of the impact of early supplementary feeding on child development were used. Two studies were conducted in West Java, Indonesia, and the third study was carried out in El Oriente, Guatemala. Repeated measures allowed for the calculation of stability and predictive correlation coefficients. The mental development scales administered up to about 18 months had modest stability but no predictive power. This trend changed during the second year of life as the strength of the stability and predictive power increased. The pre-school tests were good predictors of a child’s enrollment and school achievement. The findings on infant scales and pre-school tests agree with what has been reported in other populations in industrialized countries. There is no reason to believe that the developmental risk of the subjects that were assessed in each of the three target studies strengthens the psychometric attributes that were evaluated.
This paper presents information on whether early childhood developmental and cognitive evaluations forecast a child’s later cognitive and educational competence in the context of economically impoverished populations of low-income countries. In particular, the focus of the paper is on the stability, predictive power, and sensitivity of developmental and cognitive tests used in three different longitudinal studies of the effects of early supplementary feeding on mental and motor development. The information presented should be useful to professionals involved in the evaluation of early childhood development programmes targeted to the prevention or amelioration of developmental delays observed among children with a history of poor dietary intake, frequent infections, and limited educational opportunities. Two of these three studies were done in West Java, Indonesia [1,2]. The remaining study was carried out in El Oriente, Guatemala [3]. The respective research designs included repeated measurements under controlled conditions. Accordingly, the data allowed for quantitative estimates of the forecasting power of the developmental scales administered during the first 24 to 36 months of life and of cognitive tests administered during the pre-school period (36-84 months). The data also shed light on which tests were sensitive to nutritional interventions.
Infant development scales: mental
Infant development scales: motor
Test administered during pre-school period
Tests administered during school period
Stability and predictive power comprise forecasting, stability includes repeated measures using the same test, whereas predictive power includes different tests. Sensitivity is the track record of the test in discriminating among groups of children exposed to different nutritional interventions.
Pearson correlations were used to test stability and predictive power. The correlations were classified according to the size of the coefficients as follows: modest correlations were 0.30 or less, moderate correlations ranged from 0.31 to 0.60, and high correlations were 0.61 or more.
The three studies consisted of nutritional interventions targeted to children who were classified as nutritionally at risk in rural low-income populations. Two studies were conducted in six tea plantations in Pangalengan, 50 km south of Bandung, the capital of West Java [1,2]. The remaining study was done in Guatemala [3]. The first Indonesian study [1] tested the developmental impact of an energy-rich supplement given to infants and pre-schoolers for three months. The impact was assessed at the beginning and at the end of the treatment and again eight years later [4]. The second Indonesian study [2] began in 1993 and tested the effects of an energy and micronutrient supplement on the growth and development of two cohorts of children, 12 and 18 months of age. All measurements were taken every 2 months for 12 months. The study in Guatemala assessed the effects on mental development and cognition of infants, toddlers, and pre-school children of a high-energy, high-protein supplement and a low-energy supplement given to pregnant and lactating women and their offspring up to about seven years of age [5]. A follow-up study of the same subjects and their performance on psychoeducational tasks was carried out about 10 years after supplementation ceased [6]. The tests used in the three studies are listed in table 1.
In general, several developmental scales have been used to study the effects of different forms of intervention (e.g., nutrition, educational stimulation) on mental and motor development among young children in low-income countries. Among the most popular are the Bayley Scales of Mental and Motor Development, which we have singled out in this paper because the two scales were used in the two Indonesian studies. The Mental Scale was designed to assess sensory perceptual acuities, discriminations, and the ability to respond to these; the early acquisition of object constancy and memory, learning and problem-solving ability; vocalizations and the beginning of verbal communication; and early evidence of the ability to form generalizations and classification. In Indonesia the verbal instructions were translated to Sundanese, and some modifications were made of the pictures that are part of particular subtests. For example, some drawings were modified to portray dark-skinned, dark-haired children rather than Caucasian children, and a few utensils that are not present in Indonesian society were replaced by others. An effort was made to maintain the original instructions and degrees of difficulty in the items that were modified. The infant development scale of mental and motor development used in Guatemala was custom tailored for a study on supplementary feeding [7].
The assessment of motor development during the first two to three years of life generally includes measurements of several aspects of motor behaviour, such as gross (e.g., walking, stair climbing) and fine (e.g., finger coordination) motor skills, and motor organization and coordination. None of the studies under review assessed motor development beyond 30 months of age. However, the analysis of the motor development scales includes a section on the power of predicting cognitive performance during the school-age period.
The Guatemala study included a battery of tests that assessed different aspects of cognition, such as embedded figures (assesses the ability to distinguish a figure from among a meaningful visual array), memory for digits (child must recall a sequence of numbers read by the tester), memory for sentences (child was asked to repeat meaningful sentences after the examiner read a substantive paragraph), persistence at working at a puzzle, and verbal inferences (a partial sentence was given to the child, who was asked to complete the idea by supplying the missing words).
TABLE 1. Developmental scales and cognitive tests used in studies on early supplementary feeding and development in Indonesia and Guatemala
|
Age |
Indonesia-1 [1,4] |
Indonesia-11 [2] |
Guatemala [5-7] |
|
6-30 mo |
Bayley Scales of Mental and Motor Development |
Bayley Scales of Mental and Motor Development supplementary
feeding |
Infant development scale constructed locally for purposes of
study on |
|
36-84 mo |
Peabody Picture Vocabulary Test |
|
Battery of 10 to 22 tests of specific cognitive functions
administered yearly beginning at 36 - 84 mo. For data reduction the respective
scores were factor analyzed. A general and a memory factor emerged and was used
for statistical analysis |
|
School age |
Arithmetic test developed from school curriculum |
|
Psychoeducational test battery including tests of literacy,
reading comprehension, numeracy, general knowledge, and Raven Progressive
Matrices |
A Sundanese adaptation of the Peabody Picture Vocabulary Test (PPVT) [8] was used for the assessment of vocabulary/development among pre-school and school-age children in the first Indonesian study [1,4]. The test is considered an indicator of achievement to the extent it measures vocabulary acquisition. The PPVT can be used with children of pre-school and school age.
Arithmetic test. Besides the PPVT, the follow-up study of the first Indonesian intervention included the construction of an arithmetic test that was based on the school curriculum and appropriate for the different ages of the children included in the sample. The internal validity of the test was confirmed by the expected improvement of the test scores for each grade in school.
Psychoeducational tests. The follow-up study in Guatemala included a psychoeducational test battery that was composed of the Raven’s Progressive Matrices (i.e., visual perceptual organization) and tests of complex intellectual aptitudes, abilities, and achievements that are influenced by experience, education, and culture. For example, there were two standardized tests of reading and vocabulary and a knowledge test that was developed locally.
Infant development scales: mental
Infant development scales: motor
Pre-school assessments (36 to 84 months)
Prediction of performance at school age
Stability
The second study in Indonesia provides the most robust data on the stability of the Bayley mental scale. Two cohorts of children (12 and 18 months of age) were followed for a period of 12 months, during which the Bayley Scale of Mental Development was administered every two months. Figures 1 and 2 present stability coefficients for the 12- and 18-month cohorts, respectively. Each figure includes two curves. The first was based on the test-retest correlations with the interim period held constant at 2 months (e.g., testing at 12 and 14 months or at 20 and 22 months). The second curve was based on correlations between assessments having interim periods of different duration (2, 4, 6, 8, and 10 months). Clearly, the length of the interim period was closely related to the strength of the correlation: the longer the interval, the lower the correlation. For example, in the case of the 18-month cohort, the Pearson correlation between the scores obtained at 18 and 20 months was 0.70, and the correlation between the scores obtained at 18 and 24 months was 0.52. Likewise, for the 12-month cohort, the correlation between the scores obtained at 12 and 14 months was 0.53, and the correlation between the scores obtained at 12 and 18 months was 0.22. Furthermore, independently of the length of the interim period, the coefficients were larger for the 18-month than for the 12-month cohort. For this younger group, the correlations between scores obtained at 2-month intervals fell into the moderate correlation category (0.31-0.60). For the 18-month cohort, the correlations between scores obtained at 2-month intervals fell into either the moderate or the high categories (except for one correlation).
FIG. 1. Stability coefficients (Bayley Mental) between scores: 12-month cohort
FIG. 2. Stability coefficients (Bayley Mental) between scores: 18-month cohort
The sizes of the stability coefficient reported for Indonesia are quite similar to those of the stability coefficients obtained with mental development scales among infants and toddlers in the United States, where the Bayley Scale of Mental Development was standardized (table 2).
The infant development scale (IDS) used in Guatemala was administered at 6,15, and 24 months of age. The stability correlations were either modest or not statistically significant. For the mental scale, the highest coefficient was 0.27, which included the scores at 15 and 24 months. For the motor scale, the highest coefficient was 0.29, for these two same ages. All coefficients were consistently lower than the stability coefficients in the Indonesian study [7].
In summary, the age of the subject and the length of time between measurements have strong and distinctive effects on the stability of a. mental development scale. If the period between tests was 4 months or more and the children were 18 months of age or younger, there was little resemblance or association between the first and the second mental score. As children approached 24 to 30 months of age, the same repeated measurements had a moderate to high association. The influence of age on stability was independent of the size of the interim period (up to 10 months) between the first and second assessment points.
Prediction of performance during the pre-school period
The predictive power of the IDS was also assessed. The mental scores from the IDS were correlated with the scores from a battery of cognitive tests administered to the same children every 12 months from 36 to 84 months of age. Each of these tests yielded scores that can be used to test inter-individual or inter-group differences.
To reduce the data, the scores from the several cognitive tests in the battery were submitted to a factor analysis. Two main cognitive factors emerged: a general (including verbal and perceptual organization tests) factor and a memory factor. Most correlations between the mental developmental scores of the IDS and the factor scores either were not statistically significant or showed very modest predictive power. This was particularly clear in those correlations that involved the mental development scores obtained at 6 and 15 months. The only moderate correlations were those that involved the IDS scores at 24 months with either the general (0.51) or the memory (0.33) factor scores at 84 months.
A previous analysis of the power of the IDS developed in Guatemala to predict pre-school test performance had used a cognitive composite score for the entire preschool battery. The cognitive composite at each age was a percentile score of the average of standardized scores for all cognitive tests in the battery. Consistently, the respective coefficients of correlation between the IDS mental and motor scores, on the one hand, and the composite scores at 36, 48, 60, 72, and 84 months were smaller than those found between the IDS scores and the general and memory factor generated from factor analysis [7].
The predictive power of the scales used in Indonesia and Guatemala is similar to the predictive power of the same or similar developmental scales used in the United States (table 3). In the United States several studies that dealt with the predictive power of early development scales showed that the median correlation between the mental development scores obtained at some point between 7 and 12 months and an IQ obtained sometime between 5 to 7 years of age was about 0.20 [9, 10]. This coefficient was close to the coefficients obtained in Guatemala between the assessments at 6 and 15 months and the general score at 72 months (0.18). Furthermore, in the United States the median correlation between the mental development score obtained sometime between 19 and 30 months and the IQ score obtained sometime between 5 and 7 years of age was 0.39, which was very close to that observed in Guatemala between the mental development score at 24 months and the general score at 84 months (0.35).
TABLE 2. Stability coefficients of various infant development scales among children in the United States
|
Age (mo) |
Age (mo) |
|
|
13-18 |
19-24 |
|
|
7-12 |
0.46 |
0.31 |
|
13-18 |
|
0.47 |
TABLE 3. Median correlation across several studies conducted in the United States between infant test scores and childhood IQ
|
Age at childhood test (yr) |
Age at infant test (mo) |
||||
|
1-6 |
7-12 |
13-18 |
19-30 |
Median |
|
|
8-18 |
0.06 |
0.25 |
0.32 |
0.49 |
0.28 |
|
5-7 |
0.09 |
0.20 |
0.34 |
0.39 |
0.25 |
|
3-4 |
0.21 |
0.32 |
0.50 |
0.59 |
0.40 |
|
Median |
0.12 |
0.26 |
0.39 |
0.49 |
|
Prediction of performance during the school-age period
The data from the first study in Indonesia allowed us to assess the power that early development scales have to predict cognition at school age. The results showed that the mental development score of the Bayley Scale of Mental Development administered at 20 months or before had no statistical power to predict a child’s score at ages 8 to 12 years on either the Peabody Picture Vocabulary Score (r= .09) or an arithmetic test (r= .10). This was true even when the mental development scale score used for prediction was the average of the scores derived from two separate evaluations administered three months apart. Likewise, even if the children were classified according to age (6-12 months and 13-20 months) at baseline, the respective scores of these two age groups did not predict the scores on the verbal or arithmetic test,
Table 4 shows the predictive power of the mental scores from the infant development scale administered in Guatemala when the subjects were 15 months old and the scores obtained in different psychoeducational tests at 18 years of age. Briefly, none of the coefficients involving the mental score at 15 months were greater than 0.10. Data from the United States (table 3) show that the median correlation between a mental development score obtained sometime between 13 and 18 months and an IQ score obtained sometime between 8 and 18 years of age was 0.32 [9].
Sensitivity
The infant scales of mental development administered up to about 18 months did not discriminate between infants who had and had not received a nutritional supplement. After 18 months these scales were in fact sensitive to early supplementary feeding [11].
Stability
As shown in figures 3 and 4, the psychomotor development index (PDI) derived from the administration of the Bayley Scale of Motor Development to children between the ages of 12 to 30 months was moderately stable. In general, the stability coefficients of this scale were larger than those observed for the mental scale.
TABLE 4. Predictive power of mental and motor development scores obtained at 15 months and scores obtained in psychoeducational test at 18 years in rural Guatemala (n> 170)
|
Infant development scale |
Type of test |
|||
|
Vocabulary |
Raven Literacy |
Maximum matrices |
Grade |
|
|
Mental |
0.04 |
0.01 |
0.10 |
0.10 |
|
Motor |
0.18 |
0.16 |
0.03 |
0.14 |
Prediction of performance during the school period
As shown in table 4, the motor scores from the IDS constructed in Guatemala had low but significant power to predict performance in several psychoeducational tests administered at about 18 years of age [12]. In fact, the respective coefficients of correlation were larger than those obtained with the mental IDS score. It is noteworthy that other investigators in Canada have reported similar findings regarding infant development scores [13].
Stability
The pre-school battery constructed for the Guatemala study was used to assess the stability and predictive power of pre-school cognitive tests. As noted, two main cognitive factors (general and memory) emerged from the factor analysis of the pre-school test scores.
Table 5 presents the stability coefficients for the general score; these coefficients ranged from 0.29 (correlation between 48 and 72 months) to 0.63 (60 and 84 months). The coefficient obtained between the scores at 48 and 72 months was much lower than the other correlations. This large difference suggests that the low value should be treated as non-reliable. Within rows, the highest coefficients were consistently found in the correlations that included the scores at 84 months.
FIG. 3. Stability coefficients (Bayley Motor) between scores: 18-month cohort
FIG. 4. Stability coefficients (Bayley Motor) between scores: 12-month cohort
In the case of the memory factor score (table 6) the coefficients ranged from 0.26 (48 and 72 months) to 0.66 (60 and 72 months). There are no striking differences in the stability of the memory and the general factor scores. An exception is that the coefficients that involved the scores at 36 months tend to be larger in the general than in the memory factor. In general, the effect of the size of the interim period between the first and the second evaluation observed in the analysis of the stability of the Bayley Scale is also observed in the correlations of the pre-school battery. The size of the coefficients declines as the duration between tests increases. However, note that the interim period between the first and second assessment of mental and motor development was measured in months, whereas the interim period between tests with the pre-school battery was measured in years.
Table 7 presents the stability coefficients for the Binet Scale, which has been one of the most popular tests in different parts of the world for the assessment of general intelligence. These coefficients were obtained from a longitudinal study conducted in the United States [14] that included IQ testing from 3 to 12 years of age. Clearly, these coefficients are much larger than those reported in tables 3 and 4. Thus, although it may be argued that the coefficients reported in tables 5 and 6 suggest that the cognitive test battery administered in Guatemala was moderately stable, it is also true that pre-school tests of cognition could be more stable than what was suggested by those particular stability coefficients.
The general factor score obtained at 48, 72, and 84 months of age predicted whether the children were or were not to enroll in school (fig. 5). In addition, the same factor score at 72 and 84 months predicted whether a child would pass from one grade to the next in primary school and his or her achievement in school. This predictive power was confirmed even after the effects of the social and economic background factors were controlled for [15].
Sensitivity
The general factor score and the memory factor score derived from the pre-school battery at 48 and 60 months discriminated between the effects of the high-protein, high-energy supplement and those of the low-energy supplement given to the subjects of the study in Guatemala [5]. Furthermore, such differential effects were more evident among those children who fell at the lower end of the social and economic distribution within the villages. The factor scores at 36, 72, and 84 months did not discriminate between groups. It is important to point out that other studies have shown that intelligence scale batteries such as the Griffith Scale were also sensitive to the effects of supplementary feeding during several pre-school ages [see, for example, 16].
TABLE 5. Stability coefficients of general factor score generated from battery of pre-school cognitive tests: Guatemala
|
Age (mo) |
Age (mo) |
|||
|
48 |
60 |
72 |
84 |
|
|
36 |
0.40 |
0.39 |
0.45 |
0.62 |
|
48 |
|
0.42 |
0.29 |
0.42 |
|
60 |
|
|
0.57 |
0.63 |
TABLE 6. Stability coefficients of memory factor score generated from battery of pre-school cognitive tests: Guatemala
|
Age (mo) |
Age (mo) |
|||
|
48 |
60 |
72 |
84 |
|
|
36 |
0.42 |
0.39 |
0.34 |
0.38 |
|
48 |
|
0.46 |
0.26 |
0.18 |
|
60 |
|
|
0.66 |
0.57 |
TABLE 7. Stability coefficients of Intelligence Test (Binet Scale) from 36 to 84 months among children in the United States
|
Age (mo) |
Age (mo) |
|||
|
48 |
60 |
72 |
84 |
|
|
36 |
0.83 |
0.72 |
0.73 |
0.64 |
|
48 |
|
0.80 |
0.85 |
0.70 |
|
60 |
|
|
0.87 |
0.83 |
Infant development scales that yield a mental score (e.g., the Mental Development Index in the Bayley Scale) are often used as tests of intelligence under the assumption that the psychological constructs that these scales assess are the same as or similar to the constructs assessed by intelligence tests that yield an IQ score (e.g., Stanford-Binet, Wechsler Intelligence Scale for Children) and that are administered to pre-school (36 months and later) or school-age children. This assumption has been seriously weakened by the consistency of findings from different studies showing that the mental development scores, particularly those obtained before 18 months of age, have little if any power to predict later IQ [9,17]. Several investigators, however, have documented that the predictive power of the early mental development assessments (particularly during the first year of life) is strengthened among develop-mentally delayed children [18-20].
The psychometric data reported here show that among nutritionally at-risk children, the Bayley Scale of Mental Development and the Infant Development Scale administered before 18 months also had very poor predictive validity. On the basis of the knowledge that the mental development scale has higher predictive power among developmentally at-risk children, it seemed reasonable at first to assume that the scales used in Indonesia and Guatemala should have had comparatively higher predictive power. Although this assumption is justified because the children who were studied were developmentally at risk because of their continuous exposure to nutritional deficiencies, poor health, and limited educational opportunities, the data showed that the assumption was wrong. As has been shown in the United States, the predictive power of the scores obtained after 18 months modestly forecast the cognitive test performance of pre-school children. However, the power to predict achievement or performance in school was zero.
What is said above should not be used as a basis to conclude that early developmental assessments will not discriminate between infants and toddlers who have or have not been exposed to early interventions (e.g., supplementary feeding, educational stimulation). The scales may well be sensitive to inter-group differences in the breadth of their behavioural repertoire, but any group advantages in mean scores do not reflect advantages in intelligence.
On the basis of the above considerations, it is recommended not to use infant scales that allegedly tap mental development constructs during the first 18 months of life in the evaluation of early child development programmes, carried out with the purpose of forecasting whether the programmes foster cognition and educational competence. If such scales are used for evaluation of programmes, they should be used beginning at about 24 months of life.
The recommendations submitted above should not be extended to Version II of the Bayley Scales of Mental and Motor Development. This is a new scale, and there are no data yet available from longitudinal studies to draw any conclusions regarding stability and predictive power among children in low-income countries. New developmental scales are likely to include test items that assess particular infant cognitive skills that are allegedly related to later intelligence, such as visual attention. Evidence gathered during the last decade has shown that responses to novel visual stimuli and habituation assessed during the first 12 months of life will predict later IQ [17].
The motor development scores had modest power to predict performance in the psycho-educational tests administered at about 18 years of age, but the mental scores had no power to do so. It has been proposed that if undernutrition delays the development of early motor actions (creeping, walking) that lead to developmentally meaningful behaviours (e.g., exploration of the environment), then early motor development scores of infants who are nutritionally at risk should be correlated with their cognitive test scores in later childhood [12]. Our finding on the modest predictive power of the IDS supports this assumption. There are other longitudinal studies on early supplementary feeding and child development that also gathered data on motor and cognitive development that would allow us to test this proposition further. However, to our knowledge such information has not been published. The data reported from Guatemala do not constitute an endorsement for the use of motor scales to assess the impact of early childhood development programmes on cognition.
The difference between the moderate stability scores of the pre-school battery of cognitive tests used in Guatemala and the high stability of the Binet Scale administered in the United States suggests that there is much room for improvement in the development of pre-school tests for the evaluation of early childhood development programmes. This statement is also validated by data that were collected in Cali, Colombia, in a study on the effects of a multifocal (health, nutrition, and education) intervention on cognitive development from 42 to 88 months of life among urban undernourished children. The battery of tests included several tasks, many of which were borrowed from tests of intelligence (e.g., Wechsler test for pre-school children) that have been standardized and have demonstrated construct validity in the United States. In Cali the stability coefficients (one-year interim period) ranged from 0.48 to 0.86. These coefficients are closer to those reported for the Binet Scale than to those observed in Guatemala. The difference suggests that the room for improvement in the predictive power of pre-school tests is also found in the context of populations where malnutrition is endemic.
The programmatic conclusion from this paper is that pre-school tests of cognitive test performance have the power to predict school enrollment and achievement and are helpful instruments to evaluate the success of early childhood development programmes in fostering educational competence. Theoretically, it is justified to claim that the age of the child at the time of the evaluation has a strong moderating effect on the size of the relationship between early mental and cognitive assessments and later achievement and competence.
This work was supported in part by a grant from the Nestle Foundation, Lausanne, Switzerland.
1. Hussaini MA, Karyadi L, Husaini YK, Karyadi D, Pollitt E. Developmental effects of short-term supplementary feeding in nutritionally-at-risk Indonesian infants. Am J Clin Nutr 1991;54:799-804.
2. Pollitt E. Early supplementary feeding, motor development, activity and cognition. Report submitted to the Nestle Foundation. Lausanne, Switzerland: Nestle Foundation, 1998.
3. Martorell R, Habicht JP, Rivera J. History and design of the INCAP longitudinal study (1966-1977) and its follow-up (1988-89). J Nutr 1995;125:102S-1041S.
4. Pollitt E, Watkins WE, Hussaini MA. Three-month nutritional supplementation in Indonesian infants and toddlers benefits memory function 8 y later. Am J Clin Nutr 1997;66:1357-63.
5. Engle P, Gorman KS, Martorell R, Pollitt E. The Oriente Study: infant and preschool psychological development. Food Nutr Bull 1993;14:201-14
6. Pollitt E, Gorman KS, Engle PL, Rivera JA, Martorell R, Rivera J. Early supplementary feeding and cognition. Monographs of the Society for Research in Child Development 1993;58.
7. Lasky RE, Klein RE, Yarbrough C, Kallio KD. The predictive validity of infant assessments in rural Guatemala. Child Dev 1981;52:847-56.
8. Dunn LM, Dunn LM. Manual for Forms L and M: Peabody Picture Vocabulary Test - Revised. Circle Pines, Mich, USA: American Guidance Services, 1981.
9. McCall RB. A conceptual approach to early mental development. In: Lewis M, ed. Origins of intelligence. 2nd ed. New York: Plenum Press, 1983:255-301.
10. McCall RB. The development of intellectual functioning in infancy and the prediction of later IQ. In: Osofsky JD, ed. Handbook of infant development. New York: Wiley, 1979:704-41.
11. Pollitt E, Oh SY. Early supplementary feeding, child development, and health policy. Food Nutr Bull 1994;15: 208-14.
12. Pollitt E, Gorman KS. Long-term developmental implications of motor maturation and physical activity in infancy in a nutritionally at risk population. In: Schürch B, Scrimshaw NS, eds. Activity, energy expenditure and energy requirements of infants and children. Lausanne, Switzerland: International Dietary Energy Consultancy Group, 1990:279-96.
13. Siegel LS. Infant motor cognitive and language behaviors as predictors of achievement at school age. Adv Inf Res 1992:7:227-37.
14. Sontag LW, Baker CT, Nelson VL. Mental growth and personality development: a longitudinal study. Monographs of the Society for Research in Child Development. 1958:23, no. 68.
15. Gorman K, Pollitt E. Determinants of school performance in Guatemala: family background characteristics and early abilities. Int J Behav Dev 1993,16:75-91.
16. Grantham-McGregor S, Powell CM, Walker SP, Himes J. Nutritional supplementation, psychosocial stimulation and mental development of stunted children. Lancet 1991;338:1-5.
17. Colombo J. Infant cognition: predicting later intellectual functioning. Newberry Park, Calif, USA: Sage, 1993.
18. Werner EE, Honzik MP, Smith RS. Prediction of intelligence and achievement at 10 years of age from 20 months pediatric and psychologic examination. Child Dev 1968:39:1063-75.
19. Kopp C, McCall RB. Predicting mental development for normal, at risk, and handicapped infants. In: Baltes PB, Brimm G, eds. Life-span development and behavior. New York: Academic Press, 1982:35-63.
20. Brooks-Gunn J, Lewis M. The prediction of mental functioning in young handicapped children. In: Vietz PM. Vaughan HG eds. Early identification of infants with developmental disabilities. Philadelphia, Pa, USA: Grune & Stratton, 1988:331-55.
