Symposium Outline

Most educational effectiveness research has explored the size, consistency and correlates of school or teacher effectiveness in primary or secondary schools. The early years of education, including pre-school education, have received  relatively little attention.  This symposium seeks to redress this neglect by bringing together a range of papers that focus explicitly on the topic of educational effectiveness in the early years. Contributors from a range of countries explore evidence on factors that influence young children’s educational outcomes in the pre-primary or early elementary years using quantitative and mixed methods approaches. Topics covered include: the equity gap for low SES and ethnic minority children (Belgium – Flanders);  evidence of continued pre-school effects on children’s education outcomes in primary school (England); teacher and school effectiveness in pre-primary education and the generalisability of educational effectiveness models in this context and phase (Cyprus); changes in the attainments of children measured at entry to primary school in relation to a series of major policy changes in the pre-primary early years (England) and findings from an evaluation of the implementation of a new curriculum for the Foundation stage of education intended to improve quality and appropriateness of provision and foster better outcomes for children (Wales). 

Using data from an ongoing large longitudinal study, we investigated the evolution in educational gap between socially disadvantaged and ethnic minority children versus a reference group of mainly white middle-class children. Multilevel repeated measures analyses with a categorical classification of pupils according to social-eco­nomical and ethnic-cultural background confirm that with respect to mathematics achievement, home language effects tend to become smaller or even disappear in the first grades of primary school. However, the effects of SES-related background factors appear to be persistent. Although considerable differences were found between schools, no particular school effects were found for the two educational gaps. 

For low SES-pupils as well as for ethnic minority children the educational gap with the mainstream category of white middle-class children has been widely documented. For the present study, data were drawn from a longitudinal study in the Dutch speaking part of Belgium (Flanders), in which about 6,000 children from 200 schools are followed from the start of K3 till the end of grade 6. The present study focuses on the evolutions in mathematics achievement in first grade.

Somewhat contrary to expectations, previous analysis had shown positive coefficients for the effects of minority ethnic-cultural background and non-Dutch home language, when controlling for such factors as gender, number of years behind, prior achievement at the start of first grade, and SES. Since both ethnic-cultural background and home language are associated with each of the other predictors that were accounted for, these positive coefficients could be interpreted as indications of a catching up movement, as far as the educational gap is related to home language or other ethnic minority cultural features. It then appears that, in Flemish primary schools, the educational gap that is observed for ethnic minority children is soon reduced to a (almost) purely SES-related gap, not different from the gap that is found for Flemish low SES children.

However, the positive coefficients could also be the result of statistical artifacts due to multicollinearity. To check the catching up hypothesis, a different set of multilevel analyses were performed, in which the separate (but mutually associated) predictors relating to social and cultural background were replaced by a categorization of pupils into (1) Dutch speaking, white middle class reference group pupils, (2) Dutch speaking low SES children, (3) non-Dutch speaking low SES children, and (4) non-Dutch speaking high SES children. For these multilevel analyses, a repeated measures design was adopted, introducing an additional level “measure­ment occasion” nested within pupils, with three measurement occasions.

Results indeed show a persisting gap between categories (1) and (2), not changing over time. In contrast, the initially larger gap between categories (1) and (3) becomes significantly smaller by the end of first grade, finally leaving almost no difference between categories (2) and (3). Further division of category (3) shows larger gaps for Turkish and Arab or Berber (mainly Moroccan) children, and a faster closing gap for children with Eastern-European roots.

Analysis of factors associated with these gaps and their gradual closing during first grade revealed that for low SES-children about 30% to 40% of the initial gap in math achievement (start of first grade) is associated with lower levels of initial Dutch language skill, which was measured by means of a test involving both listening comprehension skills as well as audile and visual discriminating and synthesizing skills. The highest proportion (40 %) was found for Flemish (= native Dutch speaking) children. For non-Dutch speaking low SES children, the proportion of the initial gap in math achievement explained by lack of Dutch language skill ranged from 30 % (Arabs, Eastern Europeans, and others) to 38% (Turkish). Lower proportions were found for high SES non-Dutch speaking children. The part of the gap in math achievement associated with lower levels of Dutch language skills appears to be quite persistent, as only a relatively small part of it (3 % to 17 %)was caught up by the end of first grade.

On top of what could be explained by a lower initial proficiency in Dutch language, a relatively large proportion of the gap in math achievement appears to be associated with the degree to which the teacher experiences a cultural gap between the school and the child’s home environment. Between 29 % and 42 % of the gap in math achievement could be explained in terms of this “cultural gap”, as far as (different categories of) ethnic minority children are concerned. For the Flemish low SES children, the cultural gap perceived by the teacher still explains more than 5 % of the gap in math achievement.

In contrast to what was found for the part of the gap associated with lower levels of initial proficiency in Dutch language skill, the part of the gap in math achievement associated with this perceived cultural gap, appears to be caught up for a great deal in the course of first grade; for any of the categories of children distinguished in this study about two thirds of this part of the gap in math achievement was overcome by the end of first grade.  

The degree to which the child’s home environment is seen by the teacher as child-supportive explains a much smaller part of the gap in math achievement: 5 % to 7 % for the non-Dutch speaking children and about 10 % for the Flemish low SES children. Remarkably, almost nothing of this part of the gap was caught up in the course of first grade.

Large school and class effects were found. The differences between schools and differences between classes within schools generally tend to become smaller over time. Such differences were also observed with respect to the growth in mathematics achievement within the category of Flemish low SES students and within the distinct categories of low SES ethnic minority children. However, with respect to growth in math skills during first grade, no random effects at school or class level were found, suggesting that the between-school differences in effectiveness with regard to low SES or ethnic minority children are related to general differences in effectiveness, rather than to some sort of differential effectiveness with respect to these categories of pupils. Moreover, only small proportions of these between school and between class differences appear to be associated with the specific choices made by schools or teachers with respect to the implementation of the Flemish government’s equal educational opportunities policy. This latter finding does not imply that this policy is inadequate, it only suggests that it is of lesser importance in which way this policy is being implemented locally.

In this paper, it is considered important to identify factors that explain differences in the effectiveness of schools and teachers in relation to different criteria rather than search for criterion consistency of school/teacher effects. It also is pointed out that although EER has generated evidence of the school and teacher effect on student achievement at both the primary and secondary school level, only few studies on effectiveness in early-years education have been conducted. Thus, the study presented here attempts to identify factors of school and teacher effectiveness in mathematics and language at pre-primary education. Stratified sampling was used to select pre-primary schools (n=76) in Cyprus. All the pupils (n=2812) who attended the last year of pre-primary education of the school sample were chosen. Student skills in emergent literacy and mathematics were measured at the beginning and at the end of school year 2005-2006. Information was collected on two student background factors: sex and socio-economic status. Quality of teaching was measured through independent observers whereas semi-structured interviews with head teachers generate data about school policy on teaching. The effects of variables measuring quality of teaching and school policy on teaching upon language and mathematics achievement are examined. The importance of establishing both generic and differentiated models is supported.

In this paper, it is considered important to identify factors that explain differences in the effectiveness of schools and teachers in relation to different criteria rather than search for criterion consistency of school/teacher effects. It also is pointed out that although EER has generated evidence of the school and teacher effect on student achievement at both the primary and secondary school level, only few studies on effectiveness in early-years education have been conducted. Thus, the study presented here attempts to identify factors of school and teacher effectiveness in mathematics and language at pre-primary education. Stratified sampling was used to select pre-primary schools (n=76) in Cyprus. All the pupils (n=2812) who attended the last year of pre-primary education of the school sample were chosen. Student skills in emergent literacy and mathematics were measured at the beginning and at the end of school year 2005-2006. Information was collected on two student background factors: sex and socio-economic status. Quality of teaching was measured through independent observers whereas semi-structured interviews with head teachers generate data about school policy on teaching. The effects of variables measuring quality of teaching and school policy on teaching upon language and mathematics achievement are examined. The importance of establishing both generic and differentiated models is supported.

The unidimensionality of the teacher and school effect is considered as one of the most fundamental issues in Educational Effectiveness Research (EER). Unidimensionality relates to whether the results of schools and classes (or teachers in the case of primary schools) are the same for different subjects, school years, and groups of students. Therefore, the main question is whether results of schooling are consistent and stable. Consistency concerns the correlation between rank-orderings of schools/teachers based on different criterion variables, while time stability has to do with the extent to which the rank order of schools/teachers on output remains the same regardless of the time-point at which the effect is measured.

Although the number of studies addressing the criterion consistency of school and/or teacher effect has grown considerably, there is a great deal of uncertainty about the theoretical implications of the answer to the question on consistency of results. This uncertainty can be attributed to difficulties in identifying the sources of consistency or change. Fluctuations or changes in results over time may reflect ‘real’ improvement or a decline in school/teacher performance as well as any random variations. Changes in results may be explained by planned or naturally occurring school/teacher improvement or by non-changing school policies and teacher practices in a changing context, or by both. In this paper, it is considered more important to identify factors that explain differences in the effectiveness of schools and teachers over time and in relation to different criteria rather than search for time stability or criterion consistency of school/teacher effects. Studies investigating the criterion consistency and time stability of the effects of certain factors on achievement can help us develop generic theoretical models of EER. At the same time, it is possible to generate data which justify the importance of establishing differentiated models of teacher and school effectiveness.

In this paper, it also is argued that although EER has generated evidence of the school and teacher effect on student achievement at both the primary and secondary school level, only few studies on effectiveness in early-years education have been conducted. In this context, the study presented here does not only attempt to measure school and teacher effectiveness in teaching mathematics and language at pre-primary level. Moreover, the extent to which factors of effectiveness concerned with school policy on teaching and with the quality of teaching are associated with achievement of pre-primary students in different subjects is investigated.

Stratified sampling was used to select pre-primary schools (n=76) in Cyprus. All the pupils (n=2812) who attended the last year of pre-primary education of the school sample were chosen. Student skills in emergent literacy and mathematics were measured at the beginning and at the end of school year 2005-2006. Information was collected on two student background factors: sex and socio-economic status. Quality of teaching was measured through independent observers whereas semi-structured interviews with headteachers generate data about school policy on teaching. Multilevel analyses are currently conducted to identify the effects of variables measuring quality of teaching and school policy on teaching upon language and mathematics achievement. The results of this study will be compared with the results of effectiveness studies on teacher and school effectiveness conducted in primary schools. This comparison may help us identify the importance of testing the generalisability of theoretical models of EER and establishing both generic and differentiated models.

The Effective Provision of Pre-school and Primary School Education Project (EPPE3-11) is a longitudinal study involving over 2500 children followed from age 3 to 11 years. The relative influence of different factors (child, family, home learning environment, pre-school and primary school) on children’s reading and mathematics attainment and progress and on social behavioural development at age 6 and 10 years is explored using multilevel statistical analyses. The results indicate strong continued effects for a positive early home learning environment, as well as a range of other characteristics, including pre-school influences (measured by indicators of both the quality and effectiveness of the pre-school attended). In addition, the overall academic effectiveness of the primary school (measured independently using national data sets to provide value added indicators) has an effect on a range of educational outcomes. Interactions between pre-school and primary school effects reveal that attending a higher quality or more effective pre-school acts as a protective factor for children who go on to attend a less effective primary school whereas for home children (who did not attend pre-school) the effectiveness of the primary school attended is of particular importance for later attainment and social behaviour.

EPPE3-11 is a longitudinal study that is tracking children’s cognitive and social behavioural development from age 3+ to age 11 years and is funded by the Department for Education and Skills (DfES). The study currently involves over 2500 children in over 860 primary schools, who had previously attended a total of 141 pre-school settings (individual centers) drawn from five types of provision (playgroups, integrated settings combining care and education, local authority day nurseries, nursery classes, nursery schools and private day nurseries). The mixed methods research initially focused on exploring the influence of pre-school education using an educational effectiveness design that used quantitative measures of child outcomes at different time points and indicators of pre-school quality combined with in-depth case studies of more effective settings. A home sample (300+children ) was recruited for comparison with those who had attended pre-school at entry to primary school.  Further details of the methodology are provided by Sammons et al (2005) and Siraj-Blatchford et al (forthcoming).

The findings of the first phase of the research explored the impact of pre-school on young children’s cognitive and social behavioural development at entry to primary school and in follow ups to the end of Key Stage 1 (see Sylva et al 2004).  The results indicated that both duration of pre-school attendance (in months) and quality of pre-school attended had a positive impact on pre-school children’s progress and development from age 3 to 5 years. Multilevel statistical models of children’s progress (controlling for initial attainments/social behaviour at entry to the study were used to separate and control for the influence of important child, family and home learning environment characteristics and to provide measures of individual  pre-school effects for the 141 settings across a range of child outcomes.  Further comparisons were made with the home group using contextualised multilevel analyses to explore differences in cognitive attainments and social behaviour at entry to primary school (Sammons et al 2004).  The results of the first phase of the research informed pre-school education policy in England particularly in relation to the expansion of provision and development of Children’s Centres and drew attention to the need to raise the quality of pre-school provision (Sylva et al 2006).

The paper presents the results of new analyses that compare the continued effects of pre-school at age 6 and 10 years and compares these to the earlier findings. Standardised NFER tests of reading and mathematics are used to provide measures of cognitive attainment and teacher completed questionnaires based on an adapted version of the Goodman Strengths and Difficulties questionnaire provide measures of social behavioural development, including  Self regulation and Hyperactivity.  Multilevel models explore the impact of a wide range of child, family and home learning environment characteristics, and Effect Sizes (ES) are used to identify the strength of the net effects of individual predictors on attainment at the two time points and on progress between age 6 and 10 on the various outcomes. Measures of pre-school quality and effectiveness collected in the first phase of the research have a continued influence on later outcomes after control for the impact of other predictors, though this is stronger for mathematics and social behaviour than for reading at age 10 years. In addition, independently collected value added measures of primary school academic effectiveness in English and mathematics , derived from multilevel analyses of data for successive cohorts using national assessment data  sets for all primary schools in England were tested in the models.

The results indicate that the academic effectiveness of the primary school attended makes a significant contribution to attainment levels and progress, for children in the EPPE3-11 sample. Interesting interactions between pre-school effectiveness and quality and primary school academic effectiveness are also found.  Attending a higher quality or more effective pre-school acts as a protective factor for children who go on to attend a less academically effective primary school, whereas for home children (who did not attend pre-school) the academic effectiveness of the primary school attended is of particular importance for later attainment and social behaviour. The quality of the early years home learning environment also remains a strong predictor of better outcomes both during pre-school and throughout primary education and a stronger influence than family SES or income.

The policy implications of the findings will be discussed.

References

Sammons, P, Elliot, K, Sylva, K, Melhuish, E, Siraj-Blatchford, I, & Taggart, B (2004) The impact of pre-school on young children’s cognitive attainments at entry to reception, British Educational Research Journal, Vol 30, No, 5 pp 691-711;

Sammons, P, Siraj-Blatchford, I, Sylva, K, Melhuish, E, Taggart, B & Elliot, K (2005) Investigating the Effects of Pre-school Provision: Using Mixed Methods in the EPPE Research, International Journal of Social Research Methodology, Vol 8, No, 3 pp 207-224.

Siraj Blatchford, I, Sammons, P, Sylva, K, Melhuish, E & Taggart, B (forthcoming) Educational Research and Evidence Based Policy: The Mixed Methods Approach of the EPPE Project, Evaluation and Research in Education, Special Issue Combining Numbers with Narratives.

Sylva, K, Melhuish, E, Sammons, P, Siraj-Blatchford, I, & Taggart, B (2004) The Effective Provision of Pre-School Education Final Report, SSU/FR/2004/01, Nottingham: Department for Education & Skills Publications.

Sylva, K, Siraj-Blatchford, I, Taggart, B, Sammons, P, Melhuish, E, Elliot, K & Totsika, V (2006) Capturing quality in early childhood through environmental rating scales, Early Childhood Research Quarterly, Vol 21, pp76-92.

The Foundation Phase is a Welsh Assembly Government, national reform covering the combined ages of 3-5 Early Years and 5-7 Key Stage 1 provision. In September 2004, the first stage of the pilot commenced in 41 pilot settings across the 22 local authorities in Wales for 3-5 year olds only. In September 2005 the 41 pilot settings continued and the pilot extended to primary schools. The Monitoring and Evaluation of the Effective Implementation of the Foundation Phase (MEEIFP) project is a two-year evaluation commissioned and funded by the Welsh Assembly Government. It adopted a mixed method approach, consisting of a literature review, systematic observations, field notes, semi-structured interviews, questionnaires and informal conversations involving all major stakeholders. Key findings are reported including recommendations about:  curriculum, pedagogy, assessment and transition; best practice in the Foundation Phase; current quality and standards; qualifications, training and ratios; funding & resources and aspects of organisation and management.

The Foundation Phase is a Welsh Assembly Government, national reform covering the combined ages of 3-5 Early Years and 5-7 Key Stage 1 provision. In September 2004, the first stage of the pilot commenced in 41 pilot settings across the 22 local authorities in Wales for 3-5 year olds only. In September 2005 the 41 pilot settings continued and extended the pilot within the primary schools to include Year 1 children and some Year 2 children where mixed classes were operating. The Monitoring and Evaluation of the Effective Implementation of the Foundation Phase (MEEIFP) project is an evaluative study, commissioned and funded by the Welsh Assembly Government and is a two-year evaluation, focused on implementation. Throughout the evaluation we have worked closely with all the pilot settings and consulted widely with all major stakeholders as well as some Welsh specialists.

The Foundation Phase advocates children learning through first hand, experiential activities and play, and places a child’s personal and social development and well-being at the heart of the curriculum, it contains seven areas of learning (AOLs):

·          Personal and Social Development and Well-being;

·          Language, Literacy and Communication Skills;

·          Mathematical Development;

·          Bilingualism and Multicultural Understanding;

·          Knowledge and Understanding of the World;

·          Physical Development;

·          Creative Development.

The Evaluation Aims

To monitor and evaluate the implementation of the Foundation Phase in Wales through scrutinising:

·          The implementation of the Foundation Phase; process, timing and content, support mechanisms.

·          Transition issues for 3-5 year-olds and where possible, beyond this age.

·          How appropriate and effective are the:

o        Quality and standards;

o        Curriculum, its planning and implementation;

o        Assessment strategies and record keeping;

o        Qualifications and training of staff;

o        Ratio of adults to children;

o        Accommodation and resources;

o        Organisation and management;

o        Partnership with parents, and

o        Quality assurance, inspection.

The MEEIFP evaluation has explored six questions:

1. What does research tell us about effective pedagogy and curriculum for children of ages 3-7, drawing on key documents from the Welsh Assembly Government, for the Foundation Phase and from recent research studies?

2. What is the quality of provision provided by the pilot settings, and is it the kind of quality that has been shown to promote children’s learning and attainment and meet the needs of the individual?

3. How do the 7 areas of learning in the Foundation Phase curriculum contribute to children’s all-round development and what are the transition issues?

4. How does implementation differ for the Foundation Phase in the maintained and non-maintained sectors? And, how might any problems identified be overcome?

5. What are the perceptions of local authority partnerships, staff, parents and governors on the impact of the Foundation Phase in the pilot settings and the issues surrounding their implementation e.g. space, adult: child ratios, ‘play’ and the ‘learning continuum’ 3-7?

6. What are the main strengths of the Foundation Phase and the effective practice associated with it, where are the gaps which would make the implementation more successful e.g. in management or identifying training needs?

Methodology and Sample

The evaluation has both qualitative and quantitative components. We adopted a mixed method approach, consisting of a literature review, systematic observations, field notes, semi-structured interviews, questionnaires and informal conversations involving all major stakeholders.

In the course of the evaluation process one of our field researchers visited each of the pilot settings four times, twice during each academic year 2004-2006. During this time we have conducted observations and interviewed practitioners and parents. In addition, both in the first and the second year of the pilot the MEEIFP Team devised, piloted and conducted semi-structured interviews and questionnaires with quantitative and qualitative components to obtain the perceptions and opinions of all those involved in the pilot, including: head teachers, school governing bodies, owners/ managers/ management committees of non-maintained settings, LEA Directors of Education, LEA Early Years / Primary Advisors, Early Years Development and Childcare Partnerships (EYDCP), representatives from the non-maintained sector Associations, the Early Years Education and Safeguards Team (EYEST) at the Department for Education, Lifelong Learning and Skills (DELLS) and representatives from institutions delivering Initial Teacher Education and Training (ITET) and courses in Early Education and Childcare (Further Education). All these data have been analysed in detail to provide the evidential basis for this report. The Team has also provided a Literature Review addressing specifically the nature and coverage of the current provision, the national language policy and the evolution and the development of the Foundation Phase to date, together with robust national and international research evidence on education and care of children aged 3 – 7.

Table 1 Response rate to questionnaires and interviews

Summary of the main findings will include our findings on:

·         Curriculum, Pedagogy, Assessment and Transition

• Best practice in the Foundation Phase


• Current Quality and Standards


• Qualifications, Training and Ratios


• Funding and Resources

·         Organisation and Management

• Setting Level


• LEA / Association Level


• National Level

·         Parents, and

• Key Recommendations

England has seen massive changes in the Early Years over the last few years. There is now an official early childhood curriculum, free nursery education for three-year-olds, the Sure Start programme has started, the Neighbourhood Nurseries programme has been introduced for the most deprived communities and a national network of children's centres was launched in 2003. During this period, the CEM Centre at Durham University has been collecting consistent data from many thousands of children when they start school at the age of four on a range of variables that have been chosen because they good predictors of later success. These include for example vocabulary, concepts about print and simple arithmetic (without any formal notation). The extent to which these measures have changed from 2002 to 2006 inclusively will be examined and the link between the major initiatives and the findings discussed.

Since 1997, the English government has introduced several national and local initiatives intended to enhance the lives of young children in England and to reduce the impact of poverty and social deprivation.  Evidence that these initiatives have been successful is limited since many of the programs are still fairly new, however some evaluations have been completed – see for example Brooks et al 2003, Anning et al 2005 and the National Evaluation of Sure Start website.  This paper seeks to add to the body of evidence being developed and to provide a distinct perspective.

It investigates the skills and development of children at the start their formal schooling from 2002 to 2006 inclusively.  If the government’s initiatives, many of which are aimed at children from birth to 3 years, have enhanced young children’s development, then it might be expected that this would be evident when they start school and, hopefully, thereafter. Further, these improvements might be expected to be seen year on year for success cohorts of a single age group.

Monitoring improvements over time requires reliable assessment data, whose content remains unchanged, collected on a wide scale over an extended period. The statutory early years assessment (the Foundation Stage Profile) cannot be used for this purpose because it does not discriminate amongst the above average children, was not introduced until 2003, is dependent on judgments which may change over time and has unknown psychometric properties. By contrast the Performance Indicators in Primary Schools (PIPS) On-entry Baseline Assessment collects very reliable objective data which has well-established predictive validity and has remained unchanged for several years. (see for example CEM Centre 1999, Tymms et al 2003).

Using data from 124 schools that carried out the same assessment on all of their children (around 6000 per annum) in September every year from 2002 to 2006 inclusively, it has been possible to investigate changes in the scores of children entering school over that period.  Specifically, data were collected on picture vocabulary, concepts about print, letter identification, word recognition, early reading skills, counting, simple sums, digit identification and formal sums. The assessment was administered by teachers using a computer adaptive programme.  Additional information on the pupils’ age, sex, special needs, first language, deprivation level and the number of terms in pre-school was also recorded.

We examined the background measures for changes during the period of investigation and found that the amount of time spent in pre-school increased noticeably – from 3.09 to 3.77 terms on average. The proportion of children whose first language was not English had also increased – from 13.4% to 15.5%. We then recorded changes in the PIPS Baseline Assessment scores of the whole sample before focusing in on sub-groups the largest being those whose children whose first language was English.  Surprisingly we found remarkably constant results over the period of study. For example the picture vocabulary scores stayed within 0.03 standard deviation of the mean throughout the five years and there was no detectable trend. The one measure which had changed was the “Ideas about Maths” measure which deals with concepts such as biggest, smallest, most etc. It increased by about a fifth of a standard deviation between 2002 and 2004 as shown below.

Further findings will be presented in detail but none give any indication of large changes.

The results are discussed in relation to the Early Years initiatives in England and also to in terms of their relevance to the children’s life prospects. Potential criticisms of the negative findings are set out in detail. These include the possibility that the pupil in the schools are not representative of the national population and that the initiatives were not located near the schools. The possibility that the variables studied are not relevant to the initiatives is also considered as is the criticism that the main effect might be expected on the less able and the deprived.

References

Brooks, G., P. Cole, M. Hines, M. Lewis, T. Ohn, A. Pollock, L. Ritchie and C. Vincent (2003). Achievement In Adversity: Rotherham Rawmarsh Sure Start in 2002. Sheffield, University of Sheffield.

Centre, Centre. (1999). Performance Indicators in Primary Schools: Baseline Assessment: Technical Report: CD-ROM Version. Durham, CEM Centre, University of Durham.

National Evaluation of Sure Start website http://www.ness.bbk.ac.uk/

Tymms, P., C. Merrell and B. Henderson (2000). "Baseline Assessment and Progress during the First Three Years at School." Educational Research and Evaluation 6(2): 105 - 129.