Cognitive development and home environment of rural Paraguayan infants and toddlers participating in Pastoral del Nino, an early child development program.

Abstract. Participants included 106 infants and toddlers living in rural Paraguay and their primary caregiver. Children ranged in age from birth to 24 months and belonged to two distinct groups, including 46 children who had never participated in Pastoral del Nino, an early child development program, and 60 children who had participated in Pastoral for at least half the child's life. This article describes a study comparing the cognitive development and caregiving environment of rural Paraguayan infants and toddlers, from birth to 24 months, who were participating in Pastoral del Nino with that of children who were not participating in Pastoral programs. Cognitive (BSID-II) scores differed between the two groups, with Pastoral infants and toddlers scoring significantly higher at 0-4 months and 20-24 months. IT-HOME scores were significantly higher for Pastoral children at 0-4 months, 5-9 months, 10-14 months, and 15-19 months. Overall, best predictors for BSID-II scores included health, nutrition, and education variables, while best predictors for IT-HOME scores included caregiver education, lingualism, and community participation.

**********

More than 200 million children under the age of 5 in developing countries (Grantham-McGregor et al., 2007) fail to reach their full developmental potential because of capability deprivation. Capability deprivation, a condition extensively discussed by economist Amartya Sen (1999), has two primary indicators, including childhood stunting or "height-for-age less than -2 SD of reference values" (Walker et al., 2007, p. 146) and extreme poverty, which is defined as having an income of "less than US $1 per day" (Grantham-McGregor et al., 2007). Capability deprivation includes inadequacies in childhood nutrition, health care, education, equitable treatment, and healthy socialization and usually results in poor cognitive stimulation and development (e.g., Austin et al., 2006; Paxson & Schady, 2005; Schady, 2006; Walker et al., 2007) and delayed motor and social development (e.g., Paxson & Schady, 2005; Schady, 2006).

Several early child development programs (ECDPs) across Latin America have demonstrated successful strategies for reducing capability deprivation and promoting better development for infants, toddlers, and children at risk under 5 (Engle et al., 2007; Schady, 2006). In a comprehensive review, Engle et al. noted specific gains for children's cognitive development for participants in an ECDP in Jamaica; cognitive and psychosocial gains for ECDP participants in Colombia and Bolivia; and gains in later school performance for ECPD participants in Argentina. Not surprisingly, larger positive effects are found for children enrolled earlier in life and for a longer period (e.g., in Bolivia and Colombia). Financially, investment in ECDPs saves countries money later on (e.g., Barnett, 1997). When a cost-benefit ratio was estimated for the PIDI program in Bolivia, "the benefit of a 5% increase in cognitive scores and a 2% increase in height translated into a benefit of between $1.8 and 3.66 per dollar of project cost" (Engle et al., 2007, p. 237).

Similiar child development results and favorable cost-benefit ratios have been reported for samples in the developed world when young, at-risk children attend quality early childhood programs (e.g., Barnett, 1995). As with the developing world samples mentioned, cognitive and psychosocial gains are evident, as are later positive effects on school performance. Overall, children in the United States who attend quality ECDPs have lower rates of enrollment in special education services, lower rates of grade retention, lower rates of arrests as teens, and higher high school graduation rates than those not enrolled as children (Reynolds, Temple, Robertson, & Mann, 2001). The cost-benefit ratio for U.S. programs has been estimated at $7.16 per dollar for the High/Scope Perry Preschool Program (Barnett, 1996), $7.14 per dollar for the Chicago Child-Parent Center Preschool Program, and $3.78 for the Abecedarian Project (Reynolds, Magnuson, & Ou, 2006).

Not all ECDPs achieve the same degree of success. "Fleet programs," typically available to a large number of children, are usually less effective than "flagship programs" that deliver intensive, specialized, and carefully monitored services to smaller groups of children (McCall & Plemons, 2001). "Fleet programs usually do not provide the same dosage. They tend to have poorer staff-to-child ratios, less parent involvement, and fewer well-trained personnel, largely because they have lower budgets" (McCall & Plemons, 2001, p. 281)

Clearly, an early childhood development program (ECDP) can mitigate or even prevent the effects of capability deprivation (Engle et al., 2007; Schady, 2006), provided it focuses on multiple aspects of child development (Engle et al., 2007; Myers, 1995); occurs early in the child's life (timing); is of sufficient duration and intensity (dosage); involves parents (Engle et al., 2007; Myers, 1995; Ramey & Ramey, 2006) and communities (Myers, 1995); and has a contextual focus (Farran, 2001). Each of these points will now be discussed separately.

Multiple Aspects of Child Development

Some studies have shown a positive impact, even with attention to a single factor, such as nutritional supplementation (e.g., Pollitt, Gorman, Engle, Martorell, & Rivera, 1993). In general, however, positive effects are stronger if several facets of development are emphasized in the program. Effective ECDPs in the developing world typically involve at least several of the following facets of child development: nutritional supplementation and good nutritional practices; breast-feeding; reduction of parasites; information on making oral rehydration salts (ORS); prenatal, perinatal, and postnatal health care, including immunizations for mother and child; cognitive, linguistic, and social stimulation; and parenting.

Timing and Dosage

Extensive evidence asserts that interventions should begin as early as possible, preferably in infancy (e.g., the Abecedarian Project, Ramey & Campbell, 1994) or at least by age 3 (e.g., the Perry Preschool Project, Schweinhart & Weikart, 1983), and should continue at least to school entry. Regarding dosage, programs should be as intense as is developmentally appropriate; those with more intense services tend to have greater effectiveness in reducing capability deprivation (Ramey & Ramey, 2006). For example, the Abecedarian Project, a highly effective ECDP, provided full-day services, five days a week, for 50 weeks. In a comprehensive review of programs located primarily in the United States, Brooks-Gunn, Berlin, and Fuligni (2000) concluded that program dosage "is linked to child outcomes even after controlling for demographic characteristics of the family" (p. 566).

Parent Involvement

Parent involvement is important if it encourages greater parent sensitivity and responsivity (Austin et al., 2006), supports mothers' education (Farran, 2001), and encourages the organization of the caregiving/home environment (Bradley & Caldwell, 1984; Chall, Jacobs, & Baldwin, 1990). Maternal sensitivity is linked with positive child development in developed and developing countries. In South America, studies with Chilean and Colombian populations (Engle et al., 2007) show associations between maternal sensitivity and infant attachment; in Paraguay (Austin et al., 2006), maternal responsivity, as measured by the HOME Inventory, has been linked with better cognitive development for infants and toddlers. The organization of the home environment has long been linked with more favorable child development outcomes across childhood (Bradley & Caldwell, 1984; Chall et al., 1990). Finally, when parent involvement promotes parent education relative to the educational level of other parents in the village, it also correlates significantly with more positive development for children (Farran, 2001).

It appears that parents in developing countries place different meaning on involvement than parents in developed countries, an observation made by Austin (2007), a pediatrician. He was surprised when Cambodian parents brought their children to his clinic for the same kinds of "well but benign" illnesses as did parents in the United States. He found, however that this vigilance had different meaning in Cambodia, where parents were driven by their clear understanding that "a benign-appearing illness can turn into a deadly one much more easily.., and with greater frequency" (p. 269). It is unclear whether developing world parents might attach a similar sense of urgency to participation in an ECDP; if they do, this would surely influence the benefits realized.

Community Involvement

Community involvement, whereby participants design, implement, and evaluate the project themselves, is an important component for success, especially for programs in the developing world. In areas of rapid social change, villagers often believe that traditional wisdom plays an inadequate second to scientific knowledge from "the outside." "Traditional wisdom tends to play a negative and obstructive role representing outmoded ways of thinking that need to be overcome" (Myers, 1995, p. 323). "The crux of the matter is that meaningful community participation should permit participants to make and act on their choices. We are now talking about the participation of participants in the construction of their own future" (p. 315). This is not to say that scientific knowledge has no place in program development, but that community members themselves need to discuss both sources of wisdom and arrive at their own definition of best practices. They do this best when they are involved in all phases of the ECDP.

Contextual Factors

Finally, successful programs need to be mindful of context, particularly with regard to the implications of minority status (Farran, 2001). Contextual factors are often difficult to address, and the context of poverty is an excellent example. Poverty is known to correlate with poor child development and more problematic parenting practices, yet few ECDPs focus on the contextual issues of poverty, such as poor housing, dangerous living conditions, lack of health insurance coverage for children, or problematic parent work schedules. Clearly, "Parenting grows out of the contexts in which parents are functioning," but in Farran's review of selected early childhood programs, "none of the programs ... made any difference to the income, housing conditions, or employment of the parents involved, despite the fact that the families often were chosen because of low incomes" (p. 257).

Pastoral del Nino, an ECDP ongoing in Paraguay since 1995, takes a grass-roots approach to early childhood intervention. Its mission is to support poor children's overall development by addressing multiple areas of capability deprivation, such as health, nutrition, and early education and stimulation. Parents are involved in the intervention; relative to timing and duration, mothers are encouraged to enroll in the program as early as possible in pregnancy Children are served through infancy and early childhood, up to and including 5 years of age. The program is sensitive to context and encourages parents to make changes in housing conditions to lessen the possibility of parasitic infection

On the negative side (Engle et al., 2007), Pastoral is clearly a Fleet program. Found throughout Paraguay and serving at least 10,000 children each year, it is run almost entirely by volunteers who are minimally trained and often minimally supervised. Although general guidelines exist for intensiveness of services, this is dependent on the energy, resources, and inclination of the local volunteers. Parents are taught to intervene indirectly through play rather than directly through focused teaching, an approach that potentially weakens any intervention (Engle et al., 2007; McCall & Plemons, 2001).

This article describes a study comparing the cognitive development and caregiving environment of rural Paraguayan infants and toddlers, from birth to 24 months, who were participating in Pastoral del Nino with that of children who were not participating in Pastoral programs. To our knowledge, this is the first study to examine the child development outcomes of an early intervention program for poor rural Paraguayan children. This study contributes to the literature by identifying those factors in a low-budget, largely volunteer Fleet program that supports child growth and development in rural Paraguay

Child Development in Rural Paraguay, and Goals and Organization of Pastoral del Nino

There is a high rate of grade retention and repetition in Paraguay, especially in its rural schools. Over half the primary-school children repeat at least one grade before completing 6th grade (Austin et al., 2006). Because many of the cognitive skills for school success are developed before 5 years of age, Pastoral is concerned about early childhood stimulation. Pastoral programs stress the importance of a healthy home environment and sensitive, responsive parenting. Parents are encouraged to support child development through play and to extend children's daily experiences with developmentally appropriate conversations. Programs also focus on nutrition, health, and parent empowerment. Parents are encouraged to nurture healthy physical development so as to prevent stunting (e.g., Shrimpton et al., 2001).

Paraguay is a developing country, and its economic and social challenges have been documented in other work (e.g., Austin et al., 2006) and will be presented here only briefly. Paraguay is a bilingual society; its official languages are Spanish and Guarani. In reality, however, Spanish is the language of commerce, a fact that disenfranchises most rural and poor Paraguayans, who are typically fluent only in Guarani (World Development Report, 2006). Economically, Paraguay has suffered from a trade deficit for more than half a decade, with the economic reverberations keenly felt across the country. Nearly one-third of all Paraguayans, including 55 percent of those in rural areas, live below the official poverty level (Pan American Health Organization, 2001).

The Roman Catholic Church, through Pastoral Social (Catholic Social Services), encourages the development of ECDPs in South America. As a result, the children's relief program Pastoral del Nino has been implemented in Paraguay. Three thousand volunteers strong, Pastoral follows the same organizational structure as the Catholic Church, with volunteers on the community, parish, zone, diocesan, and national levels. Pastoral activities are directed by national leaders, also volunteers, who meet several times a year to discuss children's needs and to set the national agenda. National leaders hold trainings on agenda issues for diocesan leaders who, in turn, train zonal leaders. Zonal leaders train parish coordinators, who train community leaders. Each community leader serves 10 to 20 families with children under 5 years old and/or pregnant women. (See Austin, de Aquino, & de Burro, 2007, for a more comprehensive review of Pastoral's structure and activities.)

Leaders meet with their families in the chapel or community center once a month. They conduct training sessions and engage parents in discussion about child development. Community leaders also visit families in their homes and accompany pregnant women to the health post for prenatal checkups. They encourage mothers to have a midwife present during childbirth or to deliver their babies at the health post. The program encourages breastfeeding, as well as timely early childhood vaccinations.

To counteract the dehydrating effects of diarrhea, parents are taught to make oral rehydration salts (ORS) from commonly available foods. The use of ORS can prevent more than 1 million child deaths each year, worldwide (UNICEF, 2004).

To prevent the spread of Chagas disease and other parasites, community leaders discuss the importance of replacing thatched roofs with more impervious materials, such as wood or tin, and installing concrete floors in homes and latrines (see Ferrer et al., 2003; Rojas de Arias, Ferro, Ferreira, & Simancas, 1999). Parasites or helminthic infections are common in Paraguay and across the developing world. Children with large parasite loads may evidence muscular pain, nausea, diarrhea, and fatigue, in addition to lower resistance to other infections. Poor cognitive development and problems with memory, attention, and visual discrimination are often associated with helminthic infection (Levinger, 1996; Sternberg, Grigorenko, & Nokes, 1997).

Because most of the country's official business is conducted in Spanish, rural infants living in impoverished homes likely do not have a birth certificate, and their parents are often disengaged from civic life in other ways. To facilitate participation in civic affairs, parents are encouraged to register their children at birth and to become active participants in at least one other community organization beside Pastoral. They also are encouraged to improve their skills in speaking Spanish.

Research Questions

Given Pastoral's program goals, our research questions included the following: Will Pastoral children score better on health and nutrition measures (height and weight) than non-Pastoral children? Will Pastoral parents evidence greater community participation than non-Pastoral parents? Will Pastoral children receive higher BSID-II and IT-HOME scores than non-Pastoral children? What are the best predictors of BSID-II and IT-HOME scores for Pastoral and non-Pastoral?

Methods

Sample

Participants included 106 infants and toddlers of indigenous descent (Guarani) from birth to 24 months and their parent or guardian. They were recruited from five departamentos (states) in central and southern Paraguay. All participants were living in remote rural areas and were considered at-risk by virtue of extreme poverty, mono-lingualism, low parent education, and limited access to medical and other social services. Participants were Paraguayan nationals speaking Guarani as their first and often only language. They belonged to one of two distinct groups. The first group (n = 46) included infants and toddlers from families who had never participated in Pastoral (NP) or in any other ECD program. The second group (n = 60) included infants and toddlers from families who had been participating in Pastoral del Nino (P) for at least half the child's life, counting from conception to the child's postnatal age in months, weeks, and days. NP and P groups were each recruited from separate communities within the five departamentos to avoid contamination of non-Pastoral participants through knowledge of Pastoral programs. Children with known (frank) developmental delay were not included.

Geographic areas for recruitment of participants were selected based on relative ease of travel to, and within, the area. For the NP sample, families were recruited through the help of religious or secular leaders in the community. For the P sample, Pastoral community volunteers, who were unaware of the reason for the study or the nature and purpose of the dependent measures, helped recruit families who met the study criteria. The sample was non-random and contains an unknown degree of bias. Participation was 100 percent. Each family gave written informed consent according to a protocol approved by Pastoral and the University IRB.

Twenty-four percent of NP children and 33 percent of P children were being breast-fed exclusively or had been breast-fed exclusively for the first six months of life. The mother was the primary caregiver for at least 83 percent of the children regardless of NP or P status (NP = 38/46; P = 51/60). The father was the primary caregiver for one child in each group; another relative was primary caregiver for 7 NP and 8 P children. Regarding marital status, 12 (26 percent) NP and 18 (30 percent) P mothers were single. The remainder were permanently cohabitating (NP = 20, 43 percent; P = 17, 28 percent) or married (NP = 14, 30 percent; P = 25, 42 percent). Regarding childhood mortality, 11 (24 percent) of the NP families and 10 (17 percent) of the P families had lost at least one child to death; 3 P families had had 3 or more children die.

Infants were recruited to fill each of the following age groups: 0-4 months, NP = 10, P = 11; 5-9 months, NP = 10, P = 9; 10-14 months, NP = 8, P = 17; 15-19 months, NP = 8, P = 13; and 20-24 months, NP = 10, P = 10. Table 1 shows the number, mean age, and gender of participants by age category for NP and P.

Instruments

BSID-II

The BSID-II is a standardized measure widely used to assess children's development. With well-established reliability and validity (Bayley, 1993), it includes a Mental Development Index, Motor Development Index, and Infant Behavior Record. Only the Mental Development Index was used for this study. The BSID-II, although not normed on a South American population, has been used extensively in South America, including rural Paraguay (Austin et al., 2006).

Infant-Toddler HOME (IT-HOME)

The IT-HOME is an interview/observation instrument and was used to measure caregiving environment. It includes 45 items divided into six categories: responsivity, acceptance, organization, learning materials, involvement, and variety. The IT-HOME has well-established reliability and validity, as presented in the technical manual (Caldwell & Bradley, 1993). It has been used throughout South America and, specific to this study, with rural Paraguayan populations (for a review, see Austin et al., 2006).

Demographic Questionnaire

A 40-question survey was used to assess basic demographics, health, education, and caregiving practices. The questionnaire, written in Spanish, could be completed as a paper and pencil survey, but in reality, all of the parents preferred to respond to the questions verbally and in Guarani. A native Guarani speaker, also fluent in spoken and written Spanish, conducted parent interviews in Guarani.

Health questions included child vaccination records, whether the primary caregiver understood how to make and use ORS, and composition of the floor and roof in family dwelling. Education questions included highest grade level completed by the mother and father; the primary caregiver's own estimate of his/her facility with speaking and reading Spanish; and the caregiver's practice of trying to teach the child songs or games, tell stories, or play with the child. Family economics questions included monthly income and type of work for father or primary caregiver.

Community participation variables included whether or not the child was registered at birth, and whether or not the caregiver was a member of any community organizations other than Pastoral.

Anthropometric Measures

Anthropometric measures were used to assess nutritional status. Each of the measuring devices was recommended by the Paraguayan Ministerio de Salud and the Centers for Disease Control and Prevention (CDC) for collecting anthropometric data on infants and children under 36 months old. The instruments were easily portable.

As prescribed by the Anthropometric Survey Manual (Catholic Relief Services, 1998), length measurement was taken while the infant was lying on a measuring mat with an attached head board and a moveable foot board. Head circumference was measured using a non-stretchable measuring tape. Using these measurements, the following ratios were calculated: weight-for-length, weight-for-age, length-for-age, and head circumference-for-age. These ratios were compared to a standardized growth chart for Paraguayan children from the Ministerio de Salud, Paraguay; deviations from the standardized means also were recorded.

Procedures

Data were collected by five researchers, who divided the tasks as follows. BSID-II data were collected by Researcher 1 (a child developmental psychologist from the United States), Researcher 2 (an early childhood educator and former Peace Corps Volunteer living permanently in Paraguay), and Researcher 3 (a graduate student in child developmental psychology from the United States). Researchers 1 and 2 had previously used the BSID-II to collect data on an earlier sample of rural Paraguayan infants and toddlers (e.g., Austin et al., 2006). None of those data were used in the present study. Researcher 3 had not previously used the BSID-II but followed standard training procedures to learn to administer the instrument, both in the United States and in rural Paraguay. For 25 percent of the sample randomly selected throughout the study, Researchers 1 and 2 observed Researcher 3's administration of the BSID-II to ensure fidelity of procedures.

Researcher 4, a native Paraguayan fluent in Guarani and Spanish, administered the IT-HOME. Researcher 4 had administered the IT-HOME in a previous study with rural Paraguayans (Austin et al., 2006). Researcher 2 assisted Researcher 4 in administration of the IT-HOME, but not with any children for whom Researcher 2 had administered the BSID-II.

Anthropometric measurements were recorded by Researcher 5, an undergraduate student in pre-medicine from the United States. Researcher 5 was assisted by the other researchers when necessary and by Pastoral and community volunteers. Care was taken to ensure that the researchers remained blind to the child's NP/P status, scores on all other measures, and demographics. A child was not assessed by the same researcher on more than one instrument. Order of assessment (i.e., BSID-II, IT-HOME) and anthropometrics varied randomly.

Assessments took place in the child's home or, if travel conditions precluded that possibility, at a community center or church as close as possible to the family home. When road/travel conditions necessitated assessment at a community center, the IT-HOME was administered to the parent/caregiver as an oral interview. For all IT-HOME assessments-whether in the community center or family home--the child was sitting on the mother's/caregiver's lap or within close range so that caregiver-child interaction variables could be scored accurately. For BSID-II assessments, the child was sitting on the mother's/primary caregiver's lap. For anthropometric assessments, the child was positioned on the measuring equipment by the mother or caregiver, as instructed by Researcher 5, who then read and recorded the calibrations.

Results

Results from the demographic survey will first be discussed. Presentation of results will then be organized around the guiding questions of the study, as follows: Will Pastoral children score better on health and nutrition measures (height and weight) than non-Pastoral children? Will Pastoral parents evidence greater community participation than non-Pastoral parents? Will Pastoral children receive higher BSID-II and IT-HOME scores than non-Pastoral children? What independent measures will best predict BSID-II and IT-HOME scores for Pastoral and non-Pastoral? (See Table 2 for correlations between dependent and independent variables.)

Demographics

Parents' Education Level. There were no significant differences between groups for parents' education level. NP mothers had completed an average of 5.2 (SD = 3.22, range = 0-12) years of schooling and P mothers had completed 5.8 (SD = 3.22, range = 0-14) years. NP fathers had completed an average of 5.3 (SD = 2.9, range = 0-18) years of schooling and P fathers, an average of 5.4 (SD = 3.20, 0-16) years.

Caregiver's Ability To Read Spanish. Data were collected on mothers' self-reported reading ability in Spanish (0 = cannot read, 1 = can read with difficulty, 2 = can read easily). Although more NP mothers reported being unable to read than P mothers, no significant differences were found overall between NP and P for self-reported reading ability in Spanish. Percentages in each level of reading ability were as follows: cannot read Spanish at all, NP = 8 (17 percent), P = 9 (5 percent); reads Spanish with difficulty, NP = 9 (20 percent), P = 14 (23 percent); reads Spanish easily, NP = 29 (63 percent), P = 41 (68 percent).

Languages Spoken. Families did not differ significantly on languages spoken. Thirty-one (76 percent) of NP families and 34 (57 percent) P families spoke Guarani only. The rest spoke Guarani and some Spanish.

Paternal Employment. 42/46 NP mothers and 60/60 P mothers responded to this question. NP and P fathers did not significantly differ on type of employment. Eighteen NP (43 percent) and 34 P (57 percent) fathers had their own farms, while 24 NP and 26 P fathers farmed for someone else.

Family Income. Monthly family income did not differ significantly between the two groups (NP = 59,816 guaranies, or about $11/month; P = 55,518g, or about $10/month) (for exchange rates, see http://coinmill.com/HKD_PYG.html).

Child Registered at Birth. NP and P did not differ by the frequency with which children were registered at birth. For NP, 20/40 (50%) were registered at birth and for P, 29/51 (57%) were registered at birth. Responses were not recorded for the remaining 15 (NP = 6; P = 9) participants.

Question One:

Will Pastoral Children Score Better on Nutrition (Height and Weight) and Health Measures Than Non-Pastoral Children?

Nutrition

Anthropometric Measures. Birth weight, current weight, length, and head circumference, weight category, and length-forage and weight-for-age ratios were used as indications of nutrition. To figure weight category and length-for-age and weight-for-age ratios, standardized length and weight means for Paraguayan children (taken from growth charts published by the Paraguayan Ministry of Health) were used as the index of comparison. The infants were given scores based on their weight and their ratios: 0 = more than two standard deviations below the standardized mean for Paraguayan children, 1 = between one and two standard deviations below the standardized mean, and 2 = less than one standard deviation below the standardized mean. Thus, the higher the score, the more favorable the child's length and weight compared with age. See Tables 3A and 3B for means (SD) for anthropometric measures by NP and P and by age category and gender. Anthropometric data are missing for one P child, who became so agitated that the measurements could not be taken.

Averaging across the age categories, overall birth weight differed between NP and P. At birth, NP infants (Mean = 3.18K, SD = .76) were significantly lighter than P infants (3.54K, SD = .68), t = -2.51, 101df, p = .014. There were no significant differences between NP and P by gender or by age category. Birth weight data was missing for 3 children (1 NP and 2 P), as caregivers could not remember weights.

For weight-for-age and length-for-age ratios, there were no significant differences for either ratio between NP and P by age category or gender. However, within the NP sample on the weight-for-age measure, infants in age categories 1 and 2 (i.e., birth to 10 months) had significantly heavier weight-for-age than NP infants in age category 3 (10 to 14 months) and age category 5 (20 to 24 months), F = 4.711, 4, 45df, p [less than or equal to] .003. There were no significant differences within the NP sample on the length-for-age ratio by age category.

Within the P sample, infants in age categories 1 and 2 significantly differed from infants in age category 3 and age category 5 on both weight-for-age and length-for-age ratios. In both cases where significant effects were found, infants in categories 1 and 2 had healthier weight and length ratios than infants in categories 3 and 5. Tables 4A and 4B gives means for weight and length ratios by NP and P and by age category. Means are also given for gender in order to illustrate trends, but means comparison tests were not run for gender due to sample size.

Health

Vaccinations. More P children (94 percent) were up-to-date on vaccinations than NP children (76 percent), [chi square] = 6.375, 1df, p = .012.

ORS. Due to miscommunication, data were collected for only 59 participants (NP = 27) regarding knowledge of oral rehydration salts (ORS or alimentacion alternativa). Of the group sampled, more P caregivers (97 percent, 31/32) understood and were able to use ORS when necessary than were NP caregivers (11 percent, 3/27), [chi square] = 44.112, 1df, p = .000.

Roof and Floor. P families (50 percent) were more likely than NP families (17 percent) to have cement or tile floors, as opposed to dirt, [chi square] = 12.627, 2df, p = .002. P families were also more likely than NP families to have metal or tile roofs rather than palm, [chi square] = 6.079, 2df p = .048.

Question Two: Are Pastoral Parents More Likely TO Report Planned Educational Activities With Their Child?

In three separate questions, mothers were asked if they had a regular, ongoing practice of trying to teach their child basic concepts, sing with their child, or play with their child. Responses from P mothers differed significantly from NP mothers for setting aside time to teach basic concepts ([chi square] = 6.839, 1df, p = .009), with 65 percent (30) of NP mothers and 87 percent (52) of P mothers indicating that they made ongoing efforts to teach their child. Due to miscommunication, only 56 mothers were asked about frequency of play. For those 56 mothers, responses from P mothers differed from those of NP mothers ([chi square] = 5.951, 2df, p = .051). Thirty percent (9/30) of NP mothers and 27 percent (7/26) of P mothers reported that they never played with their child. Thirty-seven percent (11/30) of NP mothers and 12 percent (3/26) P mothers reported that they played with their child three or four times a week. Thirty-three percent (10/30) of NP mothers and 62 percent (16/26) of P mothers reported that they played with their child daily.

Question Three: Will Pastoral Parents Report Greater Community Participation Than Non-Pastoral Parents?

When asking this question, we reminded the mothers that we were asking about community participation beyond any activities with Pastoral.

There were statistically significant differences between groups on community participation ([chi square] = 58.59, 3df, p = .000). Twenty-one of 46 for NP and 34 of 62 for P reported involvement in a community organization beyond Pastoral.

Question Four: Will Pastoral Children Receive Higher BSID-II and IT-HOME Scores Than Non-Pastoral Children?

Bayley Scales of Infant Development-H (BSID-II)

Overall Mean Scores by Group (NP/P), Gender, and Group x Gender. Results of a 2 (group) X 2 (gender) ANOVA indicated significant main effects for group (F = 9.52, 1,105df, p = .003), with NP children scoring lower than P children. Mean scores for NP = 77.33 (SD = 15.27) and for P = 85.72 (SD = 12.22). Main effects for gender were not significant and the group x gender interaction was not significant. Mean score for NP boys = 79.82 (14.53) and for NP girls = 75.04 (15.88). For P boys, mean score = 84.59 (11.79) and for P girls, it was 86.77 (12.71).

Overall Mean Scores by Age Category. BSID-II scores were also figured for NP and P by age category. Scores differed significantly between NP and P for category 1 (birth to 5 months 15 days; t = -2.355, p = .027) and category 5 (20 months 15 days to 24 months; t = -2.578, p = .021). Tables 4A and 4B give mean scores by gender within age category, but t tests were not run due to small n. All variances were homogeneous for age categories.

IT-HOME

Overall Mean Scores for Group, Gender, and Group x Gender. Results of a 2 (group) X 2 (gender) ANOVA indicated significant main effects for group. P families scored significantly higher than NP families on the total IT-HOME score, (F = 34.58; 1, 105 df, p = .000); Means = NP, 20.70 (4.1); P, 25.82 (4.7). Main effects were not significant for gender. The interaction between gender and group also was not significant.

Overall Mean Scores by Age Category. As indicated in Table 5, IT-HOME scores differed between groups for age categories 1 (0-4 months-old), 3 (10-14 months-old), and 4 (15-19 months-old). IT-HOME scores for P families were significantly higher in each case than IT-HOME scores for NP families. See Tables 4A and 4B.

Question Five: What Are the Significant Predictors of BSID-II and HOME-IT Scores?

Significant Predictors of BSID-II

Full Sample (n =105). In order to find significant predictors of BSID-II scores, variables correlating significantly with BSID-II scores from each category and from the IT-HOME were considered for regression analysis. To avoid multicolinearity, when two independent variables correlated at .60 or above, one of the variables was excluded from further analysis. For the full sample (n = 106) and for NP, weight category and IT-HOME total score were significant predictors (full sample: [R.sup.2] = .22, Adj. = .20; NP: [R.sup.2] = .29, Adj. = .26). For P, IT-HOME total was the best significant predictor ([R.sup.2] = .10, Adj. = .09). See Table 5 for B, Std. Error, Beta, t, and significance levels.

Smaller Sample (n = 56). Different predictors were found for BSID-II scores for the 45 participants (NP = 25) for whom ORS and Play w/Child data were collected. For this smaller sample overall, [R.sup.2] = .45, Adj = .40, with significant predictors of BSID-II including weight category, Play w/Child, and ORS. For NP, significant predictors of BSID-II were Play w/Child and weight category ([R.sup.2] = .43, Adj. [R.sup.2] = .39). For P, significant predictors were Play w/Child, [R.sup.2] = .27, Adj. = .24. See Table 6 for B, Std Error, Beta, t, and significance levels. It is important to note that this smaller sample did not differ from the larger on mean BSID-II and IT-HOME scores or on demographic variables, either overall or by NP/P.

Significant Predictors of IT-HOME Total Score

For the full sample (n = 106), an ongoing program of teaching the child, caregiver self-reported literacy, languages spoken, and participation in community organizations were significant predictors of IT-HOME total scores ([R.sup.2] = .54, Adj. [R.sup.2] = .52). For NP, significant predictors of IT-HOME total included all of the above, except participation in community organizations ([R.sup.2] = .30, Adj. [R.sup.2] = .25). For P, significant predictors of IT-HOME total included caregiver literacy ([R.sup.2] = .15, Adj. [R.sup.2] = .13). See Table 7.

Discussion

This study contributes to the literature in several ways. First, our findings suggest that for some populations in the developing world, participation in a large, inexpensive, volunteer-run Fleet Program can correlate with better child development for young children. Fleet Programs in the United States often do not have strong positive correlations with better child development, but in the developing world, large-scale Fleet Programs may, in fact, be an important first step in addressing children's capability deprivations and raising public awareness.

Second, this study demonstrates how attention to basic contextual factors, including housing and minority status, has the potential to yield positive correlations for child development. Our findings validate Farran's (2001) suggestion that ECD programs should include a focus on the contextual elements of family life, including housing and the implications of minority status. Regarding housing, Pastoral encourages parents to reduce the capability deprivation caused by parasitic infection by upgrading floors and roofs in homes and latrines. In this study, the factor "having upgraded floors" correlated significantly with IT-HOME scores, suggesting that when there is an awareness of the impact of physical environment on child development, there is often a greater awareness of the social and emotional environment as well.

Regarding minority status, in many ways, the Paraguayans in this study are not a minority relative to their rural countrymen. Our participants were poor, but more than half of those in rural Paraguay are poor. On the average, our participants had completed only five years of schooling and were minimally fluent in Spanish, but these characteristics also are similar to the educational achievement and lingualism of other rural Paraguayans. Nonetheless, it is this low educational level and limited lingualism that relegate a majority of rural Paraguayans to minority status relative to the civil society of their country. Pastoral is conscious of the political and legal implications of minority status. Minority status often means alienation from the community; in this regard, Pastoral encourages parents to become involved in civil society through membership in community organizations beyond Pastoral. Increased fluency in Spanish is also encouraged. Indeed, both of these factors--lingualism and community involvement--were predictors of IT-HOME scores, suggesting that children benefit when parents have networks and experiences extensive to the family.

According to Farran (2001), parents' absolute education level is not as important as is their education relative to other members of the village, a conclusion that is well-illustrated by this study. Pastoral parents have the opportunity to become better educated with regard to child health and development, and evidence of this is found in the predictors of BSID-II scores. Similar to other work (e.g., Sternberg, Grigorenko, & Nokes, 1997), the variance in cognitive scores was best predicted by nutrition (birth weight), health (knowledge of ORS), and quality of caregiving environment (IT-HOME), three variables given extensive attention in Pastoral's parent education programming.

Third, the results reinforce Myers' (1995) observation that ECD programs produce the best results when community members are involved in every aspect of the program, from needs assessment to design and implementation. Although Pastoral's program is centralized to the extent that there is a national agenda, the agenda is based on local input and is always broad enough to take into account regional differences. Flexibility is afforded to local volunteers to focus on issues that in another area would not be as pressing. As an example, one of Pastoral's overarching goals is to reduce child and maternal mortality. For most departamentos, this means emphasizing timely vaccinations, hand washing, changes in building materials for homes, etc., but in one departamento, it also meant working with nearby plantation owners to stop using pesticides proven harmful to children and pregnant women.

Because Pastoral volunteers serve in their own communities and neighborhoods, they understand the needs of the community and know how to make solutions culturally appropriate. With some exceptions, they are similar to their peers in education level and income, although they may have better fluency in Spanish than others in the village. Not surprisingly, volunteers are highly respected in the community, in part because of the amount of time they spend in volunteerism while still working to provide for their own livelihood. Pastoral's success also may be due to the meaning that parents place on ECDPs. Similar to Austin's (2007) work in Cambodia, perhaps because parents realize there is a cost in not attending to child development, even when a child appears relatively healthy, they are more interested in supporting Pastoral's initiatives in their own homes and with their own children. This particular issue warrants further study.

Any study carries an unknown degree of bias, particularly comparison studies such as this one. Participant demographics, however, indicate a high degree of similarity between NP and P participants, suggesting that the bias may be minimal. NP participants were similar to P participants on parents' education level, parents' ability to speak Spanish, languages spoken, household income, and paternal employment. Additionally, care was taken to recruit NP participants from the same general area within the same departamento as P participants, although they were purposely not recruited from the same village.

Further, based on similarity of IT-HOME and BSID-II scores, the NP sample appears to represent adequately other rural Paraguayan children who also had not participated in an ECDP. There are similarities between NP cognitive scores and those found in two previous studies of poor, rural Paraguayan children, who, like the NP children in this study, had not participated in Pastoral or any other ECD program. As can be seen in Table 8, BSID-II scores found for NP children in the present study are similar to those found in two other separate samples of rural Paraguayan children, published elsewhere (e.g., Austin et al., 2006). In all, cases scores are within 1 standard deviation of each other (BSID-II SD = 15). Likewise, IT-HOME scores in the non-program sample mentioned above are comparable to IT-HOME scores for the non-program (NP) children in this study. This similarity gives additional evidence that the NP children in this study are likely comparable with other non-program children in rural Paraguay.

Clearly, ECD participation correlates with better cognitive scores in rural Paraguay. ECD participation also appears to minimize the severity of the drop in cognitive scores noted at two years of age in this study, in other studies of poor, rural children in Paraguay (Austin et al., 2006), and in studies conducted in other South American countries (Schady, 2006).

It is notable, however, that although the cognitive decline at 20 to 24 months for P children was not as severe as that for NP children, it still occurred. It is likely that cognitive decline follows the same pattern as physical growth. "Faltering begins en utero or soon after birth, is pronounced in the first 12-18 months, and could continue to around 40 months, after which it levels off. Some catch-up might take place, but most stunted children remain stunted through to adulthood" (Grantham-McGregor et al., 2007, p. 62). If cognitive decline shows the same pattern as physical stunting, we may expect the decline to continue until children are past age 3 before it will level off.

We suggest that the observed cognitive decline might be attributed to several factors embedded in the rural Paraguayan lifestyle and be linked with health, education, community participation, and human development (Elizabeth de Burro, personal communication, June 15, 2003). Regarding health, cognitive decline--even for those participants in an ECD program--may correlate with increased parasitic load. As toddlers become more mobile and are weaned from breast milk, the opportunities to develop parasitic helminths (worms) increase. Parasitic load is known to correlate with depressed cognitive scores, poor memory, inability to sustain attention, and problems with processing visual information (Levinger, 1996).

Regarding education, the decline in cognitive scores may partially be related to limitations in Pastoral's curriculum for older toddlers. Pastoral directors have expressed some concern about quality of programming as children get older, suggesting that present trainings might not be as well suited to the changing cognitive, social, and emotional needs of older toddlers. In particular, there is a need to learn better ways to encourage toddlers' problem-solving abilities (Elizabeth de Burro, personal communication, June 15, 2003). This point is supported by the differences found in IT-HOME scores for NP and P from birth to 24 months. Scores differed significantly between NP and P from birth to 19 months, but the differences were not significant for the 20- to 24-month period. Perhaps the stimulation and problem-solving needs of older toddlers have not been as well communicated to Pastoral parents.

In truth, the decline in cognitive scores at 20 to 24 months probably best reflects the accumulated effects of extreme poverty. Defined as having an income "less than U.S. $1 per day" (Grantham-McGregor et al., 2007), extreme poverty is one of the best indicators of multiple capability deprivations, including physical stunting, poor cognitive development, and lack of development and progress in school. There were some indications of stunting for the oldest infants in both NP and P samples. Severe stunting usually is defined as length (height) at or below 2 standards from the reference mean, while normal development is anything within 1 standard deviation from the mean. Infants in our study were placed in one of three categories, depending on their weight and length, and the categories were coded so that larger numbers indicate more favorable weight and length. Thus, mean scores between 1.00 and 0 (Tables 4A and 4B) indicate that weight and/or height shows some signs of stunting. Stunting usually corresponds with a decrease in cognitive processing, problem-solving, and memory.

We began this article by noting that 200 million children under the age of 5 fail to reach their full developmental potential, due to capability deprivation. According to conservative estimates, 126 million of these children live in extreme poverty, a condition almost always associated with some form of capability deprivation. Even a highly successful ECDP such as Pastoral will not be effective against the national and multinational forces that create the conditions for extreme poverty. In order to encourage better child development, visionary programs such as Pastoral are needed, and so is national and multi-national attention to the systemic causes of extreme poverty.

Authors' note: We are grateful to the many Paraguayan families who graciously welcomed us throughout this project. We thank the selfless volunteers of Pastoral del Nino, Hugo C. Aquino and the dedicated community and religious leaders who worked hard to make this project a reality. Thank you to Mary-Marie Austin, pre-med student, who worked as a volunteer to take children's anthropometric measurements. Our work would have been impossible without the energetic enthusiasm of the Peace Corps volunteers in Paraguay. Thank you. You truly do spread peace. Special thanks to Roxane Pfister for her assistance with data analysis and Teresa Bodrero for her assistance in the preparation of this manuscript. Special thanks also to the Agricultural Experiment Station at Utah State University for its continuing support. Please direct all correspondence to the second author, 2905 Old Main Hill, Utah State University, Logan, Utah 84322-2905.

References

Anthropometric survey manual (1998). Baltimore: Catholic Relief Services. Retrieved June 13, 2006, www.catholicrelief.org/publications/pdf/Heal298_e .pdf#search='Anthropometric%20survey%20manu al%20catholic%20relief%20services

Austin, A. M. B., Blevins-Knabe, B., de Aquino, C. N., de Burro, E. U., Park, K-E, Bayley, B., et al. (2006). Parent socialization, family economic well-being, and toddlers' cognitive development in rural Paraguay. Journal of Research in Childhood Education, 20(4), 255-274.

Austin, A. M. B., de Aquino, C. N., & de Burro, E. U. (2007). Bringing the abundant life to Paraguayan children. Young Children, 62(6), 30-32.

Austin, J. (2007). 88 bikes. Ambulatory Pediatrics, 7, 269-270.

Barnett, W. S. (1995). Long term effects of early childhood programs on cognitive and school outcomes. Future of Children, 5, 25-50.

Barnett, W. S. (1996). Lives in the balance: Age 27 benefit cost analysis of the High/Scope Perry Preschool Program. Ypsilanti, MI: High/Scope Press.

Barnett, W. S. (1997). Costs and financing of early child development programs. In M. E. Young (Ed.), Early child development: Investing in our children's future. New York: Elsevier.

Bayley, N. (1993). Bayley scales of infant development (2nd ed.). San Antonio, TX: Psychological Corporation.

Bradley, R., & Caldwell, B. M. (1984). 174 children: A study of the relationship between home environment and cognitive development during the first 5 years. In A. W. Gottfried (Ed.), Home environment and early cognitive development: Longitudinal research (pp. 5-56). Orlando, FL: Academic Press.

Brooks-Gunn, J., Berlin, L., & Fuligni, A. S. (2000). Early childhood intervention programs: What about the family? In J. P. Shonkoff & S. J. Meisel (Eds.), Handbook of early childhood intervention (2nd ed., pp. 549-589). New York: Cambridge University Press.

Caldwell, B. M., & Bradley, R. (1993). Home observation for measurement of the environment. Little Rock, AR: University of Arkansas at Little Rock.

Centers for Disease Control and Prevention. (n.d.). Overview of the CDC growth charts. Retrieved March 14, 2003, from www.cdc.gov

Chall, J. S., Jacobs, V. A., & Baldwin, L. E. (1990). The reading crisis: Why poor children fall behind. Cambridge, MA: Harvard University Press.

Engle, P. L., Black, M. M., Behrman, J. R., Cabral de Mello, M., Gertler, P. J., et al. (2007). Child development in developing countries 3: Strategies to avoid the loss of developmental potential in more than 200 million children in the developing world. The Lancet, 369, 229-242.

Farran, D. C. (2001). Critical periods and early intervention. In D. B. Bailey, Jr., J. T. Bruer, F. J. Symons, & J. W. Lichtman (Eds.), Critical thinking about critical periods (pp. 233-266). Baltimore: Brookes Publishing.

Ferrer, J. F., Esteban, E., Murua, A., Gutierrez, S., Dube, S. et al. (2003). Association and epidemiologic features of trypanosoma curzi and human T cell lymphotropic A virus type II in inhabitants of the Paraguayan Gran Chaco. American Journal of Tropical Medicine & Hygiene, 68(2), 235-241. Retrieved April 20, 2006, www.ajtmh.org/cgi/content/full/68/2/235

Grantham-McGregor, S., Cheung, Y. B., Cueto, S., Glewwe, P., Richter, L. et al. (2007). Child development in developing countries 1: Developmental potential in the first 5 years for children in developing countries. The Lancet, 369, 60-70.

Levinger, B. (1996). Nutrition, health, and education for all. New York: United Nations Development Programme. Retrieved October 13, 2006, www.edc.org/GLGNHEA/index.html

McCall, R. B., & Plemons, B.W. (2001). The concept of critical periods and their implications for early childhood services. In D. B. Bailey, Jr, J. T. Bruer, F. J. Symons, & J. W. Lichtman (Eds.), Critical thinking about critical periods (pp. 267-287). Baltimore: Brookes Publishing.

Myers, R. (1995). The twelve who survive: Strengthening programmes of early childhood development in the third world. Ypsilanti, MI: High/Scope Press.

New formulation of oral rehydration salts (ORS) with reduced osmolarity, February 2004, Technical Bulletin #9, UNICEF, retrieved August 19, 2007, www.supply.unicef.dk/catalogue/newORSA&A.pdf

Pan American Health Organization. (2001). Country health profile for Paraguay. Retrieved April 14, 2005, from www.paho.org/English/ SHA/prflpar.htm

Paxson, C., & Schady, N. (2005). Cognitive development among young children in Ecuador. World Bank policy research working paper 3605, May 2005, retrieved August 19, 2007, from www.ds.worldbank.org/external/default/WDSContentServer/IW3P/ IB/2005/05/12/000011823_200

Pollitt, E., Gorman, K. S., Engle, P. L., Martorell, R., & Rivera, J. (1993). Early supplementary feeding and cognition. Monographs of the Society for Research in Child Development, 58(7).

Ramey, C. T., & Campbell, F.A. (1994). Poverty, early childhood education, and academic competence: The Abecedarian experiment. In A. C. Huston (Ed.), Children in poverty: Child development and public policy (pp. 190-221). New York: Cambridge University Press.

Ramey, S. L., & Ramey, C. T. (2006). Early educational interventions: Principles of effective and sustained benefits from targeted early education programs. In D. K. Dickinson & S. B. Neuman (Eds.), Handbook of early literacy research (Vol. 2, pp. 445-449). New York: Guilford Press.

Reynolds, A. J., Temple, J. A., Robertson, D. L., & Mann, E. A. (2001). Long-term effects of an early childhood intervention on educational achievement and juvenile arrest: A 15-year follow-up of low-income children in public schools. JAMA, 285, 2339-2346.

Reynolds, A., Magnuson, K., & Ou, S. R. (2006). PK-3 education: Programs and practices that work in children's first decade. FCD Working Paper: Advancing PK-3(6), January 2006. New York: Foundation for Child Development.

Rojas de Arias, A., Ferro, E. A., Ferreira, M. E., & Simancas, L. C. (1999). Chagas disease vector control through different intervention modalities in endemic localities of Paraguay. Bulletin of the World Health Organization, 77(4), 331-339. Retrieved April 20, 2006, from www.who.int/docstore/bulletin/pdf/issue4/bu5782.pdf

Schady, N. (2006). Early childhood development in Latin American and the Caribbean. World Bank Policy Research Working Paper 3869. Retrieved June 1, 2006, from www-wds.worldbank.org/servlet/WDSContent Server/WDSP/ IB/2006/03/17/000016406_20060317163141/ Rendered/PDF/wps3869

Schweinhart, L. J., & Weikart, D. P. (1983). The effects of the Perry Preschool Program on youths through age 15. In Consortium for Longitudinal Studies (Ed.), As the twig is bent ... Lasting effects of preschool programs (pp. 71-101). Hillsdale, NJ: Erlbaum.

Sen, A. (1999). Development as freedom. New York: Anchor Books.

Shrimpton, R., Victora, C. G., de Onis, M., Costa Lima, R., Blossner, M. et al. (2001). Worldwide timing of growth faltering: Implications for nutritional interventions. Pediatrics, 107, 75-81. Retrieved June 1, 2006, www.pediatrics. org/cgi/content/full/107/5/e75

Sternberg, R. J., Grigorenko, E. L., & Nokes, C. (1997). Effects of children's ill health on cognitive development. In M. E. Young (Ed.), Early child development: Investing in our children's future (pp. 85-125). New York: Elsevier.

UNICEF. (2004). New formulation of oral rehydration salts (ORS) with reduced osmolarity. Technical Bulletin No. 9, New York.

Walker, S. P., Wachs, T. D., Meeks Gardner, J., Lozoff, B., Wasserman, G. A., et al. (2007). Child development in developing countries 2: Child development: Risk factors for adverse outcomes in developing countries. The Lancet, 369, 145-157.

World Development Report. (2006). Equity and development--2005. Washington, DC: The World Bank.

Shannon Peairson

Ann M. Berghout Austin

Utah State University

Cyle Nielsen de Aquino

Elizabeth Urbieta de Burro

Pastoral del Nino, Asuncion, Paraguay

Table 1
Sample Characteristics, Including Mean Age (SD) and Gender by Age
Category of Infants in Pastoral (P) and Non Pastoral (NP)

                   N (Boys)                  Mean Age (SD)

Age Category     NP       NP              NP                 P

0-4 months     10 (4)   11 (8)    2.81 mos (1.03)   2.36 mos * (1.03)
5-9            10 (6)   9 (3)     7.37 (1.64)       6.89 (1.76)
10-14          8 (4)    17 (9)    11.13 (1.36)      12.06 (1.39)
15-19          8 (4)    13 (6)    16.38 (1.30)      16.93 (1.49)
20-24          10 (4)   100 (3)   22.05 (1.43)      22.30 (1.42)

* p <.05, all variances homogeneous

Table 2
Correlations Between Independent and Dependent Variables

                            1          2          3         4

1. Pastoral                 --
2. Age/mos                .036         --
3. Gender                -.006       .121        --
4. Ill in past 2 weeks   -.102      -.109       .003       --
5. Serious illness        .071      -.091      -.030      .204 *
6. Vaccines               .250 **    .160      -.065      .104
7. ORS                    .868 **    .172      -.044     -.008
8. Floor                  .336 **    .011       .069      .114
9. Roof                   .004      -.101       .045     -.071
10. Wt. cat.             -.024      -.406 **   -.201 *    .159
11. Length cat.          -.015      -.389 **   -.192 *    .040
12. Garden                .282 *     .071       .088      .088
13. Caregiver lit.        .111      -.076       .037      .178
14. Mother's ed.          .113      -.074      -.012       198 *
15. Father's ed.          .027      -.182       .031      .153
16. Teach                 .260 **    .370 **    .109      .005
17. Play                  .193      -.239       .204      .136
18. Sing                  .050       .130      -.052      .034
19. Father's job         -.133      -.064      -.134      .153
20. Language              .127      -.048       .064      .041
21. Registered            .068       .293 **    .248 *   -.034
22. Family
    organizations         .555 *     .077      -.064      .053
23. MDI                    295 **   -.276 **   -.031      .078
24. HOME Total            .502 **    .180       .062      .009

                            5          6          7          8

1. Pastoral
2. Age/mos
3. Gender
4. Ill in past 2 weeks
5. Serious illness          --
6. Vaccines               .084         --
7. ORS                    .081       .299 *       --
8. Floor                  .110       .190 *     .262 *       --
9. Roof                  -.025       .065      -.028       .315 **
10. Wt. cat.              .371 **    .162       .125       .122
11. Length cat.           .111       .010      -.005       .05
12. Garden                .004       .158        329 *     .213 *
13. Caregiver lit.        .060       .161       .176       .240 *
14. Mother's ed.          .100       .280 **    .160       .370 **
15. Father's ed.          .078       .149       .053       .250 *
16. Teach                 .132       .236 *     .225       .077
17. Play                  .000       .267 *     .279      -.070
18. Sing                 -.054       .006       .259 *     .028
19. Father's job          .049       .049      -.065      -.045
20. Language              .115       .053       .262 *     .294 **
21. Registered           -.002       .16        .204       .098
22. Family
    organizations         .121       .331 **    .605 **    .162
23. MDI                   .047       .000       .437 **    .166
24. HOME Total            .083       .373 **    .589 **    .412 **

                            9          10        11         12

1. Pastoral
2. Age/mos
3. Gender
4. Ill in past 2 weeks
5. Serious illness
6. Vaccines
7. ORS
8. Floor
9. Roof                     --
10. Wt. cat.              .086         --
11. Length cat.           .125       .611 **     --
12. Garden               -.029       .019       .005        --
13. Caregiver lit.        .088       .060       .089      .106
14. Mother's ed.          .241 *     .130       .087      .131
15. Father's ed.          .305 **    .193       .205 *    .076
16. Teach                 .058      -.122      -.093     -.032
17. Play                  .015       .340 **    .204      .060
18. Sing                 -.09       -.052      -.082      .068
19. Father's job          .028       .018      -.067     -.217 *
20. Language              .199 *    -.005       .098      .060
21. Registered            .099       .018      -.131      .137
22. Family
    organizations        -.034       .172       .139      .337 **
23. MDI                  -.028       .257 **    .210 *   -.049
24. HOME Total           -.024       .129       .114      .151

                            13        14         15         16

1. Pastoral
2. Age/mos
3. Gender
4. Ill in past 2 weeks
5. Serious illness
6. Vaccines
7. ORS
8. Floor
9. Roof
10. Wt. cat.
11. Length cat.
12. Garden
13. Caregiver lit.          --
14. Mother's ed.          .615 **     --
15. Father's ed.          .414 **    .59 **      --
16. Teach                 .035       .137     -.038         --
17. Play                  .436 **    .40 **    .396 **    .256
18. Sing                 -.018       .008     -.051       .075
19. Father's job         -.094       .028      .110      -.033
20. Language              .284 **    .41 **    .280 **    .071
21. Registered            .124       .110      .093       .360 **
22. Family
    organizations         .159       .26 **    .004       .298 **
23. MDI                   .223 *     .232 *    .155       .108
24. HOME Total            .399 **    .52 **    .196       .428 **

                            17       18        19        20        21

1. Pastoral
2. Age/mos
3. Gender
4. Ill in past 2 weeks
5. Serious illness
6. Vaccines
7. ORS
8. Floor
9. Roof
10. Wt. cat.
11. Length cat.
12. Garden
13. Caregiver lit.
14. Mother's ed.
15. Father's ed.
16. Teach
17. Play                    --
18. Sing                 -.026       --
19. Father's job         -.064       .042      --
20. Language              .214       .017    .173        --
21. Registered            .293 *    -.048   -.275 **   .048        --
22. Family
    organizations         .270 *     .138   -.128      .007      .180
23. MDI                   .531 **   -.131    .000      .107      .168
24. HOME Total            .508 **    .149   -.048      .376 **   .265 *

                           22        23

1. Pastoral
2. Age/mos
3. Gender
4. Ill in past 2 weeks
5. Serious illness
6. Vaccines
7. ORS
8. Floor
9. Roof
10. Wt. cat.
11. Length cat.
12. Garden
13. Caregiver lit.
14. Mother's ed.
15. Father's ed.
16. Teach
17. Play
18. Sing
19. Father's job
20. Language
21. Registered
22. Family
    organizations          --
23. MDI                  .244 *      --
24. HOME Total           .475 **   .419 **

Note: * p [less than or equal to] .05; ** [less than or equal to] .01.

Table 3A
Means (SD) for Birth Weight and Present Weight- and Length-for-Age
Category by NP and P and by Age Category

                   0- to 4-months-old
                       (n #boys)

               NP (10, 4)       P (11, 8)

Birth
Weight
  Total        3.34 (.93)       3.53 (.80)
  Boys         3.23 (.57)       3.77 (.68)
  Girls        3.42 (1.16)      2.97 (.90)

Present
weight
for age cat-
  egory
  Total        1.90 (.32) (a)   2.00 (.00) (b)
  Boys         2.00 (.00)       2.00
  Girls        1.83 (.41)       2.00

                   5- to 9-months-old
                       (n #boys)

               NP (10, 6)       P (9,3)

Birth
Weight
  Total        3.14 (.70)       3.67 (.49)
  Boys         3.15 (.66)       3.50 (.50)
  Girls        3.13 (.85)       3.75 (.50)

Present
weight
for age cat-
  egory
  Total        2.00 (.00) (a)   2.00 (.00) (b)
  Boys         2.00             2.00
  Girls        2.00             2.00

                    10- to 14-months-old
                         (n #boys)

               NP (8, 4)         P (17, 9)

Birth
Weight
  Total        3.10 (.54)        3.54 (.72)
  Boys         3.25 (.50)        3.49 (.69)
  Girls        2.95 (.61)        3.60 (.79)

Present
weight
for age cat-
  egory
  Total        2.00 (.00) (a)'   1.44 (.73) (b)'
  Boys         2.00 (.00)        1.34 (.74)
  Girls        2.00 (.00)        1.50 (.76)

Note: For NP, values followed by (a) differ from those followed
by (a)', p [less than or equal to] .01.

For P weight-for-age, values followed by (b) differ from
those followed by (b)', p [less than or equal to] .05.

Table 3B
Means (SD) for Birth Weight and Present Weight- and Length-for-age
Category by NP and P and by Age Category

                         15- to 19-months-old (n #boys)

                          NP (8, 4)     P (13, 6)

Birth weight (k)
  Total                   3.16 (.94)    3.30 (.44)
  Boys                    3.35 (.99)    3.48 (.33)
  Girls                   2.96 (.99)    3.14 (.48)
Present weight for age
category
  Total                   1.50 (.76)    1.69 (.48)
  Boys                    1.74 (.50)    1.83 (.41)
  Girls                   1.25 (.96)    1.57 (.53)
Present length for age
category
  Total                   1.0 (9.94)    1.47 (.88)
  Boys                    1.50 (1.00)   1.67 (82)
  Girls                   1.50 (1.00)   1.33 (.58)

                         20- to 24-months-old (n #boys)

                          NP (10, 4)       P (10,3)

Birth weight (k)
  Total                   3.15 (.77)       3.76 (.93)
  Boys                    3.75 (.38)       4.17 (.74)
  Girls                   2.67 (.66)       3.56 (1.01)
Present weight for age
category
  Total                   1.20 (.79) (a)   1.40 (.84) (b)'
  Boys                    1.75 (.50)       1.67 (.58)
  Girls                    .83 (.75)       1.29 (.95)
Present length for age
category
  Total                   1.20 (.78)       1.20 (.92)
  Boys                    1.75 (.50)       1.33 (.58)
  Girls                    .50 (.84)       1.14 (1.07)

Note. For NP, values followed by (a) differ from those followed
by (a), p [less than or equal to] .01.

For P weight-for-age, values followed by (b) differ from those
followed by (b)' , p [less than or equal to] .05.

For P length-for-age, values followed by (c) differ from those
followed by (c)', p [less than or equal to] .05.

Table 4A
Means (SD) for BSID-II and IT-HOME Total by NP and P, Gender, and Age
Category

               0- to 4-months-old
                    (n #boys)

           NP (10, 4)        P (11, 8)
BSID-II
  Total    72.30 (13.9)      87.45 (9.1) **
  Boys     64.75 (9.5)       87.00 (9.5)
  Girls    77.33 (14.7)      88.67 (10.0)
IT HOME
(45 pos-
sible)
  Total    18.00 (2.8) ***   23.43 (3.6) ***
  Boys     19.00 (3.1)       23.78 (4.3)
  Girls    17.17 (2.5)       22.80 (1.6)

             5- to 9-months-old             10- to 14-months-old
                  (n #boys)                     (n #boys)

           NP (10, 6)     P (9,3)         NP (8, 4)     P (17, 9)
BSID-II
  Total    93.20 (5.8)    88.67 (12.6)    84.13 (9.7)   86.82 (12.6)
  Boys     91.67 (6.5)    94.00 (1.0)     91.00 (2.2)   79.89 (12.3)
  Girls    95.50 (4.2.)   86.00 (15.1)    77.25 (9.4)   94.63 (7.7)
IT HOME
(45 pos-
sible)
  Total    21.67 (5.1)    27.3 (5.1) **   22.00 (2.6)   25.53 (4.4) *
  Boys     19.88 (5.8)    25.25 (6.2)     21.25 (3.0)   25.38 (3.4)
  Girls    23.71 (3.3)    28.22 (5.5)     23.00 (2.0)   27.86 (5.3)

* p [less than or equal to] .05, ** p [less than or equal to] .05,
*** p [less than or equal to] .05

Table 4B
Means (SD) for BSID-II and IT-HOME Scores by NP and P, Gender,
and by Age Category

                           15- to 19-months-old
                                 (n #boys)

                         NP (8, 4)        P (13, 6)
BSID-II
  Total                 75.50 (12.5)   86.69 (13.5)
  Boys                  80.50 (5.8)    82.00 (16.4)
  Girls                 70.50 (16.3)   90.71 (9.9)
IT-HOME (45 possible)
  Total                 20.88 (2.4)    28.38 (4.2) ***
  Boys                  20.75 (2.2)    28.50 (4.9)
  Girls                 21.00 (2.8)    28.29 (3.9)

                           20- to 24-months-old
                                 (n #boys)

                         NP (10, 4)       P (10,3)
BSID-II
  Total                 62.50 (12.2)   78.00 (11.8) **
  Boys                  65.25 (15.7)   88.00 (8.0)
  Girls                 60.67 (10.6)   73.71 (10.8)
IT-HOME (45 possible)
  Total                 21.20 (3.89)   24.90 (5.1)
  Boys                  22.25 (2.4)    27.00 (7.6)
  Girls                 20.50 (4.7)    24.00 (4.1)

* p [less than or equal to] .05, ** p [less than or equal to] .05,
*** p [less than or equal to] .05

Table 5
Regression Models (full sample) for BSID-II Overall and for NP
and P Separately

                              Std.
                  Adj         Error
[R.sup.2]         [R.sup.2]   of Est.   Predictors

Overall     .22   .20         12.61     Constant

                                        Weight
                                        Category
                                        HOME
                                        Total
NP          .29   .26         13.13     Constant
                                        Weight
                                        Category
                                        HOME
                                        Total
P           .10   .09         11.67     Constant
                                        HOME
                                        Total

                     Std.
Predictors   B       Error   Beta   t       p

Constant     47.84   6.23           7.33    .000

Weight       5.00    2.13    .21    2.35    .021
Category
HOME         1.09    .25     .39    4.45    .000
Total
Constant     37.04   10.92          3.393   .001
Weight       11.00   3.41    .42    3.23    .002
Category
HOME         1.034   .49     .28    2.12    .040
Total
Constant     64.32   8.45           7.62    .000
HOME         .83     .322    .32    2.58    .013
Total

Table 6
Regression Models (smaller sample) for BSID-II Overall and for
NP and P Separately

                                  Std.
                      Adj         Error
          [R.sup.2]   [R.sup.2]   of Est.   Predictors

Overall   .45         .40         11.5      Constant

                                            ORT
                                            Plays
                                            w/child
                                            Weight
                                            category
NP        .43         .39         11.6      Constant
                                            Plays w/
                                            child
                                            Weight
                                            category
P         .27         .24         13.7      Constant
                                            Plays with
                                            child

                     Std.
Predictors   B       Error   Beta   t       p

Constant     60.38   5.79           10.43   .000

ORT          10.42   3.61    .35    2.89    .006
Plays        5.70    5.79    .33    2.58    .014
w/child
Weight       7.26    3.33    .27    2.18    .035
category
Constant     54.74   6.54
Plays w/     9.99    3.45    .52    2.89    .008
child
Weight       10.07   3.78    .41    2.67    .013
category
Constant     69.89   4.20           16.65   .000
Plays with   9.99    3.45    .52    2.89    .008
child

Table 7
Regression Models for HOME Total Scores for Overall Sample

                        Std
                        Error
            Adj         of                               Std.
[R.sup.2]   [R.sup.2]   Est.    Predictors       B       Error   Beta

.54         .52         3.50    Constant         14.31   1.06

                                Regular pro-     3.67    0.86    .31
                                gram of teach-
                                ing child
                                Caregiver        1.47    0.52    .21
                                literacy
                                Participation    2.08    0.37    .43
                                in organiza-
                                tions
                                Languages        2.59    0.77    .25
                                spoken

Predictors       t       p

Constant         13.44   .000

Regular pro-     4.25    .000
gram of teach-
ing child
Caregiver        2.80    .006
literacy
Participation    5.70    .000
in organiza-
tions
Languages        3.39    .001
spoken