Understanding the Early Years - Early Childhood Development in South Eastman, Manitoba - October 2003

II. The outcomes for children of South Eastman

A. How the outcomes were measured

This section provides more information about the specific measures of children's outcomes. A child's cognitive skills, behaviour, and physical health and well-being outcomes were measured in two ways, using the Early Development Instrument (EDI) and the National Longitudinal Survey of Children and Youth (NLSCY) Community Study. In South Eastman, all assessments were conducted in French for students attending Francophone schools which comprised about one third of the South Eastman sample.

Five domains of the EDI (teacher report)

Cognitive skills (from the NLSCY - direct assessments of the child)

Vocabulary Skills (Peabody Picture Vocabulary Test, Revised — PPVT-R): assesses a child's receptive or hearing vocabulary. The children hear a word said aloud and are asked to point to one of four pictures that they believe corresponds to the word.

Developmental Level (Who Am I?): is based on copying and writing tasks, which are designed to test children's ability to conceptualize and to reconstruct a geometrical shape and to use symbolic representations, as illustrated by their understanding and use of conventional symbols such as numbers, letters, and words. Children are asked to copy five shapes (such as a circle or a diamond) and to write their names, numbers, letters, words, and a sentence. Because the tasks are not dependent on language, Who Am I? can be used to assess children whose knowledge of English or French is limited.

Number Knowledge Assessment: is designed to test the child's understanding of numbers. Children who do not have this understanding, or who are working in a language different from their mother tongue, often have difficulty mastering basic arithmetic and demonstrating number sense. The Number Knowledge Assessment evaluates children's understanding of quantity (more vs. less), their ability to count objects, their understanding of number sequence, and their ability to do simple arithmetic.

Behaviour outcomes (from NLSCY community study - parent report)

Measuring a child's behaviour is based on a scale administered to the person most knowledgeable about the child, which is usually the mother.⁵ The measurements comprise several questions, each with the same format. For example, the mother is asked how often her child cannot sit still, is restless, or is hyperactive. She answers with one of three possible responses — ''never or not true''; ''sometimes or somewhat true''; and ''often or very true.'' The scale included the following elements:

Positive social behaviour: children who exhibit higher levels of positive social behaviour are more likely to try to help and comfort others. They may offer to help pick up objects that another child has dropped or offer to help a child who is having trouble with a difficult task. They might also invite their peers to join in a game.

Indirect aggression: this element identifies children who, when mad at someone, try to get others to dislike that person; who become friends with another for revenge; who say bad things behind the other's back; who say to others, ''Let's not be with him/her''; or who tell secrets to a third person.

Hyperactivity: hyperactive children cannot sit still; are restless and are easily distracted; have trouble sticking to any activity; fidget; cannot concentrate, cannot pay attention for long; are impulsive; have difficulty waiting their turn in games or groups; or cannot settle to do anything for more than a few moments.

Emotional disorder/anxiety: this element identifies children who seem to be unhappy, sad, or depressed; are too fearful or anxious; are worried; cry a lot; tend to be rather solitary; appear miserable, unhappy, tearful, or distressed; are not as happy as other children; are nervous, high strung, or tense; or have trouble enjoying themselves.

Physical aggression and conduct disorder: these children get into many fights. When another child accidentally hurts them (by bumping into them, for example), they assume that the other child meant to do it, and then react with anger and fighting. Also included are children who kick, bite, or hit other children; who physically attack people; and who threaten people, are cruel, or bully others.

B. What we learned from teachers: results of the Early Development Instrument

The children of South Eastman scored well overall on the five domains, compared with children in the EDI-16 sample (see Table 2.1).⁶ The largest differences were in the areas of Physical Health and Well-being (0.4 points above average), Social Knowledge and Competence (0.9 points above average), Emotional Health and Maturity (0.2 points above average) and Communication Skills and General Knowledge (0.7 points above average). The scores for Language and Cognitive Development did not differ significantly from the EDI-16 sample average.

Figure 2.1 displays box plots describing the distribution of EDI scores for South Eastman compared with the EDI-16 sample. The box plots show the median and percentiles for the distribution of EDI scores for each group (See Inset 3). The median is the mid-point at which 50% of the cases fall above and 50% of the cases fall below. Percentiles refer to the percentages of cases with values falling above and below the number. Ideally, a community would want to have a high median score, with relatively short blocks above and below the median. For South Eastman, Figure 2.1 shows that the median scores for the EDI domains are above those of the EDI-16 sample in all domains except Language and Cognitive Development. The range of scores is indicated by the length of the boxes.⁷ The interquartile range of children in South Eastman is in most cases similar to that of the EDI-16 sample. The exception is Social Knowledge and Competence, which has a wider interquartile range, and a wider range overall. These results generally confirm those presented in Table 2.1.

The EDI-16 was also used to establish a ''low score'' threshold for each EDI domain. The low-score threshold scores were set to the tenth percentile, which means that 10% of all children scored below this score for each domain. Thus, if a community had typical results, we would expect 10% of its children to score below the same threshold scores for each domain. In South Eastman, the percentage of children with very low scores on the EDI was close to 10% (ranging from 6.7% to 10.9%). None of these differences are statistically significant.

The data collected as part of the Understanding the Early Years study included information about where each child resided. Therefore, an analysis was conducted that would give some indication of how children's scores on the EDI were distributed geographically. To achieve this, we determined the average score within each enumeration area, for each domain of the EDI. We then ''smoothed'' the average scores for each enumeration area.

Figures 2.3 through 2.7 display the geographic distribution of the EDI scores for each of the domains. For each map, the yellow and orange areas represent scores that are just below (orange) or just above (yellow) the median score of the full EDI-16 sample. Scores that are somewhat below the EDI-16 median are indicated in light red, and very low scores (which are comparable to the bottom 17% of the population) are shown in dark red. Similarly, relatively high scores are represented in light green, while very high scores (which are comparable to the top 17% of the population) are shown in dark green. . Although the distributions vary by domain, most of the EAs scoring above the median, shown as yellow to green areas, are in the northern and western side of the region. The southern area of the region tends to have average scores below the median, coloured orange to dark red.

The map in Figure 2.3 shows a range of scores from well above to well below the EDI-16 median score (8.8) on Physical Health and Well-being. However, there is a contiguous block of EAs in the westernmost and northwestern sections of the community where the scores are above to well above average. The EAs in the central southern section of the community tend to have lower scores than those in the western and northwestern sections.

Figure 2.4 also shows considerable variation among the EAs in their average scores, which range from a few very low scoring EAs to a few very high scoring EAs, with several just above or below the EDI-16 median (7.9) for Social Competence. In this domain also, the central southern section tends to have relatively low scores.

Figure 2.5 shows that scores for Emotional Health and Maturity for most EAs are close to the EDI-16 median (8.2). There are a few areas where there are concentrations of low-scoring EAs, and only a few EAs with average scores well above the median. The latter are located in the northwestern section of the community.

Figure 2.6 shows that many EAs in South Eastman exhibit scores below the EDI median (8.8) for Language and Cognitive Development. Some of the EAs in the central southern region have very low scores in this domain. The average scores for EAs in the western side of the community tended to be close to the EDI median.

Figure 2.7 indicates a fairly even distribution of scores on the Communication Skills and General Knowledge test, both spatially and in terms of the range of scores above and below the EDI median (7.5). There is one cluster of relatively low scores in the western central region of the community, and several pockets in the westernmost EAs attaining relatively high scores for this domain.

Overall, the maps for South Eastman show some consistency in the pattern for the EAs related to SES and the five EDI domains. In general, the EAs in the western sections of the region received higher scores on the five tests than those in the central and eastern EAs. The map in Figure 1.1 shows that the EAs in the central and eastern areas are predominantly low and very low SES, while those in the western section tend to be low to low-middle SES. Thus, although the maps suggest a moderate relationship, they also indicate that socio-economic background is not a definitive predictor of EDI outcomes. Therefore, other factors that influence children's development should be considered. These outcomes may be more fully explained when additional family and community factors are taken into consideration.

C. What we learned from parents, guardians, and the children: NLSCY community study results

In this section, we discuss the results of the National Longitudinal Survey of Children and Youth Community Study, which measures children's cognitive skills, positive social behaviour, and behaviour problems.

Table 2.2 displays the means and standard deviations of scores on the Developmental Assessment (Who Am I?), on the Positive Behaviour Scale, and on the Receptive Language (PPVT-R) Test for South Eastman. Figure 2.8 displays their distributions.

The average score for South Eastman on the Developmental Assessment (Who Am I?) is 99.0, and on the Positive Behaviour Scale is 99.7, neither of which differ significantly from the 2001-02 UEY average of 100. In addition, the standard deviation on both these tests was close to 15, indicating that the spread of scores on these tests is also similar to those in the 2001-02 UEY sample or the NSLCY sample. However, on the Receptive Language Test, children in South Eastman received an average score of 88.7. This is markedly below the standard Canadian average of 100. The standard deviation for this test in South Eastman was 16.8. Thus, there is cause for concern in this community regarding Receptive Language and vocabulary.

Figure 2.8 shows the distribution of scores on these tests, compared to the Canadian norms. The distribution of scores for Receptive Language is particularly noteworthy: about 75% of students in South Eastman scored below the Canadian average of 100. Thus, the low average score on this test, 88.7, is not attributable to a few children with very low scores; rather it appears to be associated with generally lower scores among the majority of children in the region.

Figure 2.9 shows the prevalence of children with low scores on the Developmental Assessment (Who Am I?), the Positive Behaviour Scale, and the Receptive Language Test. It also shows the percentage of children deemed to have a behaviour problem, based on four measures of behaviour (hyperactivity, emotional disturbance/anxiety, aggression/conduct disorder, and indirect aggression).

For each measure, a score at the 10^th percentile of the 2001-02 UEY sample (for the Developmental Assessment and the Positive Behaviour Scale) or the nationally representative NLSCY sample (for the PPVT-R) was used as the threshold to define a ''low score''. Similarly, children with scores above the 90^th percentile for the NLSCY sample on the behavioural measures were considered to have a behaviour problem. For each measure the prevalence of children in South Eastman with low scores on the three developmental assessments, and the prevalence of behaviour problems, was calculated. This allows one to compare whether the prevalence of children in South Eastman with significant problems in these areas is above or below the national norm of 10%.

The results indicate that the prevalence of children in South Eastman with behaviour problems (hyperactivity, emotional disturbance/anxiety, and aggression/conduct disorder) ranged from significantly below the national norms fixed at 10% (6.3% for indirect aggression) to significantly above the national norms (15.6% for hyperactivity). The prevalence for emotional disturbance (7.2%) and aggression/conduct disorder (12.9%) did not differ significantly from the national norms. On the Developmental Assessment (Who Am I?) and Positive Behaviour Scale, the prevalence of 10.1% and 10.0% respectively did not differ from the national norms. The prevalence of children with low scores Receptive Language PPVT-R was 22.40%, more than twice the national norm. This analysis suggests that the number of South Eastman with behaviour problems approximates national norms, but that there is a serious problem with language skills among children in the region.

The study also included a direct measure of a child's understanding of the system of whole numbers. Scores were classified according to developmental levels:

• Have not reached level 1;
• Reached level 1 (usually attained by 4-year-olds);
• Reached level 2 (usually attained by 6-year-olds);
• Reached level 3 (usually attained by 8-year-olds).

For all of the children who did the assessment across the seven UEY sites, only 1.1% had failed to reach level 1. The majority of children (42.8%) were at level 1, or had made the transition to level 2 (54.2%). Only 1.9% of the UEY children had reached level 3. These results are as expected given that the UEY children were 5 and 6 years old.

In South Eastman, 96.7% of the children sampled had either reached level 1 or had made the transition to level 2. This is nearly identical to the 2001-02 UEY prevalence of 97.0%. Only 1.6% of children in South Eastman had failed to reach level 1.

Overall, the analyses in this section indicate that South Eastman has some marked strengths and weaknesses in early childhood outcomes. Its strengths lie in the areas of children's social knowledge and competence, general knowledge and communication skills, and positive behaviour. On measures of these skills teachers rated children at or above national norms. Its weaknesses are in the areas of language and cognitive development, which pertains to children's mastery of the basics of reading and writing, interest in books, and numerical skills. The parents' ratings of children's behaviour indicated that compared with national norms there are about one-and-a-half times as many South Eastman parents who rated their children as hyperactive. However, on other scores of behaviour, the children in this region fared very well. Some of these results are likely attributable to family background and to various family and community factors. This hypothesis is explored in the next two sections.

⁵ Statistics Canada trained personnel conducted the parent interviews by telephone only in English or French for the NLSCY. Parents without telephones or speaking other languages were not interviewed.

⁶ The EDI sample size, N=289, included valid data only. To be included in the EDI sample for South Eastman, children needed scores on at least 3 out of the 5 EDI domains. This explains why the EDI sample size (N=289) is different from the NLSCY sample size (N=296) for South Eastman.

⁷ The longer the boxes, the greater range of variability in the EDI domain scores. For example, the EDI-16 physical health and well-being domain has a short box which indicates that scores were very similar to one another. In contrast, the communication skills and general knowledge domain has a long box which indicates that, scores varied considerably, ranging from very low to very high scores.