Emotional/behaviour disorders, pre-school

There is considerable evidence demonstrating the importance of early childhood emotional and behavioural disorders for the long-term mental health of children. The original important epidemiological research was carried out by Richman, Stevenson and Graham (1975) who found 7% of three year old children in a London borough to have medium/moderate to severe behavioural problems, with a further 15% showing mild

behaviour problems.  Incidence in boys and girls were similar, Expressive Language Disorder was associated with Behaviour Disorder, and follow-up after one year showed strong continuities for behaviour and language disorders. In addition, behaviour disorders persisted in two thirds of the pre-school sample when they were eight years old (Stevenson, Richman and Graham, 1985). The conclusion of these researchers was that pre-school behaviour disorder is common, handicapping and persistent.  More recent studies have confirmed this conclusion, with evidence of continuities in problems from preschool to middle childhood (Campbell and Ewing, 1990)  and a review by Campbell (1995) arguing that 10-15% of older pre-school children have significant psychiatric disorders.

While these overall figures are of considerable importance both for appreciation of the significance of children’s difficulties and for service planning, there are some groups which might, on a priori grounds, be expected to show rather different patterns of childhood difficulties and yet which have been very little studied. One such group is the children of highly religious cultural and ethnic groups, because of the specific attributes of such communities. These include apparently high levels of social/community cohesion, focus on family life, ideological conservatism and, obviously, high levels of religious observance. In addition, in many such communities very large families are encouraged, and there are consequently often high levels of poverty. Some of these factors (community cohesion, religious observance) are believed to be protective against child (and adult) psychological difficulties; others (such as economic privation) tend to be associated with high morbidity rates (Melzer et al, 1999). Information about these groups could thus be important both in clarifying the factors associated with children’s distress as well as offering much-needed data for the planning of appropriate services for members of religious cultural groups.

In this study, we report data concerning rates of emotional and behavioural disorders amongst preschool children in one such highly religious cultural/ethnic group, the strictly orthodox Jewish community in London. This community lives mainly in ‘enclaves’, in London and other cities. There is strict adherence to Jewish laws regarding diet, prayer, social and sexual relationships, Sabbath and festivals, and other aspects of life. Two salient features dominate the upbringing of children. First, family size is normally very large: averages of 5-7 children per family have been estimated. Second, schooling follows a distinctive pattern: parents wish to give their children a ‘Torah education’, involving single-sex schooling and a very high proportion of time spent in studying religious texts such as the Pentateuch and Talmud. Few of the schools meeting the requirements of strictly orthodox parents receive state or local authority funding, and thus an important consequence of these two features is economic. The financial burdens of providing for large families, including providing unsubsidised education, may be an important risk factor for psychiatric morbidity among adults in this community (Loewenthal et al, 1995). Very little is known about psychological difficulties among children in the strictly orthodox Jewish community since there is reluctance to admit to problems and to seek help, especially outside the community. Fear of stigmatisation is a powerful factor driving the widespread view that ‘s/he will grow out of it’. The current study attempts to begin an understanding of psychiatric morbidity among children in this community. This study is a necessary preliminary to more detailed study of causal factors in the family, school and community, to be undertaken in future work, and is an important preliminary for improved detection and intervention.

Method

This project was set up to estimate the rate of emotional and behavioural disturbance amongst children aged 3 and 4 years living in the strictly orthodox Jewish community in North London.

Sample

The strictly orthodox Jewish community is located in a relatively tightly designated geographical area in North London. The community is characterised by strict adherence to the laws of Judaism, as codified through legal rulings and traditions over many hundred years, and currently interpreted by the rabbinate of the Union of Orthodox Hebrew Congregations and other groups. Children from this community attend nurseries and schools run by the community itself, making school-based studies a viable method for sampling.

Stages of sampling were as follows.

1. All local nursery schools were sent letters explaining the nature of the project and emphasising that it had been discussed with, and approved by, the local rabbinical authorities.

2. Five nursery schools were asked, and all agreed, to participate in the study. These schools served the local strictly orthodox Jewish population, with a particular focus on one area designated as part of a local ‘SureStart’ project.  

3. All 3-4 year old children in these nursery schools whose parents did not ‘opt out’ were included in the study (parents of six children elected to withdraw their children from the study). This resulted in a sample of 262 children (180 girls, 73 boys, 9 gender unknown). 

Procedure

Teachers of all children in the specified age range were asked to complete the questionnaire pack on each child. Where necessary, the research worker met with teachers to explain the study to them. In some classes, where there was more than one teacher, the teacher with most contact with the children was nominated to complete all the questionnaires. Participating teachers received a one off payment so that they could fill out the questionnaires in their own time.

Parents of children in the classes concerned were written to with a description of the study (including confirmation of rabbinical support) and an opt-out letter. They were also sent their own set of questionnaires to complete. The schools themselves, using a coding procedure by means of which the anonymity of participants was ensured, managed all communications with parents. Completed questionnaires were received from parents of 109 children (25 boys, 84 girls; 42% of the sample); five parents requested that their children be removed from the study.

Materials

1. Teachers’ pack

Strengths and Difficulties Questionnaire, teachers’ version  (Goodman, 1997). The SDQ has been used principally with 5-16 year old children, although the original intention was for it to apply from 3 years of age, and some piloting was carried out on that basis. Other measures, such as the Preschool Behaviour Checklist (McGuire and Richman, 1986a) have more commonly been used with younger children. However, because the main focus of our research has been in the older age group of 5-14 year olds, for which the SDQ has been used, we decided to use it with preschoolers so that age group comparisons may be made more easily across our whole sample.

A teachers Background Characteristics sheet designed specifically for this project (see appendix).

2. Parents’ pack

 Strengths and Difficulties questionnaire, parents’ version (Goodman, 1997).

A parents’ Background Characteristics sheet designed specifically for this project (see appendix).

Ethical issues

As with most studies of this kind, confidentiality and anonymity of responses was a significant concern. This was managed by placing responsibility on the schools for coding questionnaires and contacting parents; effectively, this meant that the researchers were given no information on the names of children in the study. This had particular importance as two of the research team are members of the community. 

Approval for the study had been obtained in advance from the Head of the Rabbinical Court (Beth Din) of the Union of Orthodox Hebrew Congregations. Possible ethical issues (for example, how to manage a situation in which a parent or teacher expresses concerns about a child on the questionnaire form) were discussed with him and appropriate procedures agreed. 

Results

 Characteristics of the Sample

1.1 Proportions of boys/girls, 3/4-year-olds, children with 2 parents. 

Of the 244 children for whom data on gender was provided, 170 (69%) were girls and 67 (28%) boys; the remaining 7 questionnaires were unclear.  Sixty one percent of children for whom age data was available (N=236) were 3 year olds, the remaining 39% were 4 year olds. Almost all questionnaires from parents (N=109) revealed children to be living with both parents.

1.2 Proportions of children in owner-occupied homes, with at least one parent employed, with reported family difficulties (e.g. health, finance), and with reported Special Educational Needs 

Forty-two percent of the sample lived in owner-occupied properties, although these were often of poor quality. While nearly 90% were in employment, much of this was part time or casual in nature, so the actual economic characteristics of this group are substantially over-rated by these data. About 54% of the sample had recognisable ‘family difficulties’; in about 8% of families there was a child with special needs, as rated by teachers.

1.3 Mean number of children in family, and mean ordinal position of target child

Even though most of the parents in the study were still of childbearing age, family size was already very large by general community standards –a mean of 5.4 children per family. Because we were studying 3-4 year olds, this meant that most children had several older brothers and sisters (the mean ordinal position of target child in the family was 4.4).

 Rates of Emotional and Behavioural Disorder

A series of analyses was carried out on the SDQ data to explore the rates of emotional and behavioural disorder in the sample. The SDQ is scored on a variety of subscales (emotional, conduct, hyperactivity and peer disorder, and prosocial abilities) which also give a ‘total difficulties’ score; cut-off points are also given for ‘borderline’ and ‘case’ levels of disturbance. Data are presented separately for teachers’ and parents’ ratings.

2.1 Prevalence of psychological disturbance

Table 1 gives rates of disturbance measured in terms of ‘borderline’ or ‘case’ status, as derived from teachers’ and parents’ ratings separately.

INSERT TABLE 1 HERE

Comparisons between teachers’ and parents’ ratings

The correlations between teachers’ and parents’ ratings were all positive, but some were very small - the smallest was r=.031, for the emotional difficulties ratings. Those for hyperactivity, peer relations and prosocial behaviour were significant (p<.001 in all cases), all other correlations, though positive, were non-significant.

INSERT TABLE 2 HERE                                          

It was hard to test agreement between teachers and parents on caseness because of low numbers, but the impression is that levels of agreement were generally rather low. If cases and borderline cases are collapsed together, there is only one contingency table in which there is significant agreement; this is for conduct disorder. (Fisher exact probability =.015.)

INSERT TABLE 3 HERE

Even though this is the one set of ratings where there was significant agreement over caseness, the actual levels of agreement are not impressive: 15 of the 23 parent cases were not cases according to the teachers, and 17 of the 24 teacher cases were not cases according to the parents. It appears that not only are teachers more likely to rate children as having conduct and hyperactive difficulties than are parents, but the two groups of raters also pick out different children as having problems.

4. Predictors of Difficulties

A regression analysis was carried out on the data in order to explore the predictive power of some of the background factors. For this analysis, dependent variables were the (continuous) measures of emotional and behaviour disorders, plus the total (emotional and behaviour) difficulties score, and the prosocial scores. Both teacher and parent measures were used. Thus a total of 12 (6x2) analyses were carried out. Independent variables were: Age (3 or 4), Gender, Living with both parents, Special Educational Needs, Family difficulties and Firstborn or later born. Only teacher information was used for the independent variables, as there were too many missing cases among the parent responses.

Significant (p<.05) results are given in Table 4. 

INSERT TABLE 4 HERE

It can be seen that important predictors of teacher-rated difficulties were children’s SEN status, gender (boys having more likelihood of problems) and not living with both parents (though most of the sample were living with both parents). 

Comparisons by gender

As studies in the general population have suggested that boys are more vulnerable to preschool difficulties than girls, gender differences in the orthodox Jewish data were examined more closely. The only significant result was for the finding that boys were significantly higher than girls on conduct problems, as rated by parents (t (251)=1.86, p<.05). 

6. Comparison with a General Community Sample

There is no comparative data for non-clinical samples of 3-4 year old children using the SDQ teachers’ form. As part of the pilot study for their national survey of children using the SDQ (Goodman et al, 2000) Goodman and his colleagues collected parent SDQ ratings on 23 three year olds and 17 four year olds, and have kindly made their data available to us. Table 5 looks at the proportion of borderline and ‘cases’ of this group in comparison with the parent-rated data from the strictly orthodox Jewish sample.  

TABLE 5 HERE

There are significant differences between our sample and Goodman’s sample on scores on conduct, hyperactivity, peer and total difficulties, measured both as proportion of children who are ‘case plus borderline case’ (chi squared test) and mean scores (t tests). In all instances, children in our sample score lower than in the comparison group.

Discussion

This study provides some preliminary data on the frequency of emotional and behavioural difficulties in a population not previously studied, the strictly orthodox Jewish community of north London. As expected, this community is characterised by considerable family cohesion, with the overwhelming majority of children living with both parents. It is also characterised by very large family sizes and probably by high levels of economic privation, though our measures were not specific enough to provide detailed documentation of this latter feature. Our data show that teachers are more likely than parents to rate these preschool children as having difficulties, especially of the ‘hyperactive’ kind, and that the levels of such difficulties are probably epidemiologically  significant (15% of the sample rated at ‘case’ level for conduct disorder and 14% for hyperactive disorder on the SDQ). The lack of adequate comparative data makes it hard to know how these levels compare with that found in other groups, although the small parent-rated sample studied by Goodman showed significantly higher rates in the general population. There were few relevant predictive factors of difficulties, although children already perceived by their teachers as having ‘special educational needs’ (quite a broad term, left unspecified in our data sheet) had a clearly heightened risk in comparison with other children. This finding leads us to believe that a more focused and detailed study of special educational needs in the strictly orthodox Jewish community would be helpful, possibly linked to an intervention study.

There are some methodological weaknesses in this study, in addition to the lack of teacher-rated comparative data. Our background characteristics form is quite crude, allowing us to make only general probabilistic statements about links with difficulties. Because of the way preschool education is organised in the strictly orthodox community, we had substantially fewer boys in our sample than girls, yet there is evidence that it is boys who are more likely to show problems. Including some small ‘chedarim’ in the study, which we could not do in the time frame of the present study, would have meant that some teachers may have had problems interpreting the language of the questionnaires, thus requiring a Yiddish translation. The large differences between teacher and parent ratings are difficult to interpret: they may be artefactual, but may also reflect real differences in children’s behaviour in distinct settings.

In conclusion, we have provided data indicating that there are high levels of difficulty in young children in the strictly orthodox Jewish community, but that these levels may nevertheless be lower than one might expect given the economic characteristics of that community. This suggests both that protective factors associated with family and social cohesion are offering to support children’s well-being, and that services building upon these factors may significantly enhance children’s developmental experiences.

Acknowledgements

Many thanks are due to Stamford Hill Sure Start for financial support of this project, and to the participating schools (principals, administrators and teachers) and  parents for their time and interest.