The superior educational attainments of pupils in religious foundation schools in England.

Prais, S.J.

Pupils of religious foundation schools in England show superior educational performance over general (Local Education Authority) schools, the advantage having been estimated in previous studies at learning about a tenth faster for the average pupil by the age of leaving primary schooling (age 11+). In this study access to individual pupils' scores at SAT tests shows that the advantage of religious foundation schools is particularly great for lower-attaining pupils, with only the lowest tenth of those in religious schools attaining the scores of the lowest third in general schools.

Keywords: schooling attainments; religious schools; SATs JEL classification: 121, 128

**********

It has been thought for some time that pupils at schools affiliated to a religious foundation perform better, in broad terms, in national educational tests than pupils at general schools (which, in England, means that great majority of schools sponsored by Local Education Authorities). An evaluation, in quantitative terms, of that advantage would benefit (a) parents choosing schools for their children; (b) educational authorities deciding which kinds of new schools need to be established; (c) research educationists concerned with elucidating aspects of school organisation and of teaching which contribute to success in learning.

School-averages in the US and England

Research interest in this topic was much stimulated a generation ago by a large US sample study of schools and pupils into factors contributing to inequality of educational opportunity--a study carried out by the respected professor of sociology at the University of Chicago, James S. Coleman. This was followed by studies focussing more specifically on differential achievements in religious and private schools showing, for example, that pupils at Catholic schools gained--by school-leaving age--about one 'grade-level' over general schools 'in mathematics, vocabulary and reading comprehension'; in other words, they gained about one year's learning in a total of ten years' schooling--an advantage of about 10 per cent. (1)

A similar proportionate advantage has been noted for pupils in England who are all obliged to take the nationwide compulsory SAT tests, as documented by Dr John Marks (in Burns et al., 2001) in relation to SAT results for 1998. For example, in mathematics at age 11, pupils at religious foundation primary schools in England (CE and RC schools combined) attained scores in Key Stage 2 tests putting them at the equivalent of 5.2 months of schooling ahead of pupils in general LEA schools; tests for English showed a closely similar advantage of 5.8 months (Burn et al., 2001, p. 9). In effect, pupils at religious foundation primary schools had gained the equivalent of about one month's learning for each year of primary schooling over pupils at general schools. At a younger phase of their schooling, at age 7 when they took Key Stage 1 tests, pupils at religious foundation schools were 3.4 months ahead in mathematics and 3.7 months in English; their greater pace of learning thus seems to have been fairly consistent throughout primary schooling. As Dr Marks has also shown, a substantial advantage for religious foundation schools is to be found at the subsequent secondary schooling stage--though once achievement is measured in terms of GCSE and A-level passes at various grades--it is not so easy to translate into equivalent years of schooling gained.

A more detailed study of English secondary school results, based on 'value-added' at GCSE tests in 2000, came to an apparently much more reserved conclusion: 'On the whole, it seems that church schools--whether C of E, RC or 'other Christian'--outperform non-religious schools on some (i.e. not on all: SJP) measures, but only to a very slight degree' (Schagen et al., 2002, p. 34). Even that was qualified because many of the various measures of schooling success were based on numbers of school-subjects passed at GCSE, and religious foundation schools more often required pupils to take Religious Education as an additional GCSE subject. Without going into the merits of including that subject into an aggregate comparative measure of schooling success, it seems that the advantage detected in this way was not substantial.

The real reason for the only modest advantage found in that study, I suspect, was that success was measured in terms of 'value-added'--that is to say, the study attempted to measure the differential gain acquired by pupils at religious secondary schools above the advantage they had already gained from having attended a religious primary school (most attending a religious secondary school have also attended a religious primary school; but not vice versa). Unfortunately no analysis was carried out in that study of the total advantage gained by those pupils who had attended both secondary and primary religious schools (above that gained by pupils who had attended both non-religious secondary and non-religious primary schools; nor of those pupils who had attended a religious primary school followed by a non-religious secondary school). Especially for parents who, in choosing a school, wish their child to benefit from mixing with other children of high attainment, it is the (gross) standards reached that are important, not the 'value-added' in that school.

The above recent findings for England by Marks are consistent with the earlier findings quoted for the US; but, it is important to notice, in one important respect the English results are more convincing since both kinds of school (religious and non-religious) are available in England on a non-fee paying basis, whereas in the US religious schools cannot legally be financially aided out of taxation: one might thus be tempted to attribute some part of the learning-gain of pupils in US Catholic schools to their schools being substantially private, and to their pupils coming from more economically-advantaged homes.

In attempting a judgement on how important, or unimportant, an advantage of six months' learning may be at the end of primary schooling, and what steps might be taken to transfer any teaching techniques (or whatever) to general schools, it is helpful to ask which kind of children benefit most: do all children benefit substantially equally; or do the intellectually most advanced in the age-group benefit more; or, perhaps, the intellectually less advanced? Under the older system of selective secondary schooling in England, it will not have been entirely forgotten that a primary school's reputation often depended on how many of its children it was able to 'get through the 11 + exam' (which gave admission to the upper secondary stream of 'grammar schools'); teaching efforts towards the end of primary schooling were consequently sometimes unduly concentrated on, say, the top tenth of pupils who might be pushed over that hurdle--even if it meant some sacrifice of learning attainments by the rest of the class. It is thus of considerable interest to know at which end of the intellectual attainment range the religious foundation schools show the greatest advantage.

Advantages to low-attaining pupils

To attempt an answer to that question requires individual Pupils' test results, rather than merely--as can be derived from published national statistical returns--average scores for each school as a whole. (2) Fortunately, the requisite detail was made available for this study (in a suitably anonymised form) for primary schools in one London borough, namely, Barking and Dagenham (with which the National Institute has been co-operating on educational research for over a decade); that Borough's educational attainments have now risen to be close to the average of all Inner London boroughs taken together, slightly below those in Outer London (with their greater share of suburban, more prosperous, professional families), and only very slightly below England as a whole--in short, not too unrepresentative of the whole country. (3) There are seven religious foundation primary schools in that Borough with a total of 303 children in their final year of schooling, and 28 Local Authority schools with 2076 children in their final year of schooling.

Table 1 shows the scores attained in mathematics tests (at age 11, Key Stage 2) for each decile of children in each group of schools, together with a conversion of those scores to notional mathematical ages based on National Curriculum concepts (a considerable degree of definitional convention attaches to such conversions; while these conversions must not be treated too precisely, the general import of the figuring may be accepted). Compared with an average advantage at age 11 of about 1.0 years by all pupils in religious foundation schools (0.9 years at the median), we see (final column of table) that the weakest quarter of pupils in religious schools are at an advantage of about 1 1/2 years over those at general LEA schools; while towards the top quarter of the attainment range the advantage of religious schools--though still highly important--is down to about three-quarters of a year. The greater relative advantage of the weakest pupils in religious schools will probably be judged as of greater significance--though 'ceiling effects' in such tests imply that even that implication is not without reservations. (4) Looking at another aspect of that comparison, the average score ('raw marks') attained at the lowest decile of pupils in religious schools (41 per cent) is not attained until the third lowest decile in general schools: on that criterion one might even say, very simply, that general schools have three times as many very low-attainers (scores below 41 per cent) as religious schools--a formidable difference!

A 'long tail of under-achievement' by English as compared with Continental pupils has long been a source of concern to the English schooling system; the tail is so very much shorter in religious foundation schools (only a third as long, on the criterion just mentioned) that a strong incentive is provided for a detailed investigation of causes.

The Institute's previous comparisons of English with Swiss pupils at this age have shown a variability of scores (on a common maths test) of Swiss primary pupils at only about half that of English pupils, based on the standard deviation of scores. Using that same measure of variability on the scores in table 1, the gap between our two types of school is 18.4 at religious schools compared with 23.8 at general schools, ie. 23 per cent less variability in religious foundation than in our general schools. Taken with a large pinch of salt, it can be said that religious foundation schools in England are about halfway on the road towards equality of attainments as are Swiss schools (Prais, 1997, p. 282)! That overall measure, however, does not so immediately convey the great gap in attainments as at the lower third of the attainment range on which the text above has focussed.

Next steps

Several levels of causation can be distinguished for future research. First, are religious foundation schools academically more selective in admitting children, perhaps by gently suggesting to parents that children showing signs of disability would be better helped in a larger neighbouring Local Authority school, where more specialised staff could be available? It can perhaps already be agreed, since we are looking for a gap affecting almost a third of the school intake, that this seems unlikely to form much of the answer; nevertheless, a few children could be affected.

Secondly, it could be that the home-backgrounds of children in religious foundation schools are more educationally and emotionally supportive--for example, having a greater general concern with the ideals of education, fewer single-parent homes--than in general schools. Children, especially from single-parent homes, have more home-background worries and cannot concentrate as well on their schoolwork.

Thirdly, there is the cumulative, infectious, nature of a few disturbed and disturbing children: the ability of the class as a whole to learn together at a good pace can be much impeded--out of all proportion to their number--by having more than a bare minimum of such children. The smaller size of religious foundation schools, at an average of two 'forms of entry' compared with three 'forms of entry' for general schools (based on the schools in our sample local authority, see table 2), may well contribute to a calmer general atmosphere which especially helps children with home difficulties.

Fourthly, the grouping of children of different learning abilities (the degree of 'setting', etc) may be different in the two types of school: this is an aspect of schooling that deserves much more empirical study.

Finally, there is the diplomatically sensitive issue of the quality of teachers. One aspect, at least, could be objectively examined--the average length of service of teachers within a school, or rate of staff turnover. It may well be that religious foundation schools have a significant advantage in this respect, from which much else flows in terms of teaching ability and class-wide attention.

Table 1. Distribution of scores in SAT mathematics test, and of
notional mathematical ages (in years), of pupils at age 11 (KS2 in
2003): 2,400 pupils in 7 religious and 28 general schools in the London
Borough of Barking and Dagenham (a)

 Scores (out of 100)

 Religious General Difference
Percentile schools schools

10 41 22 19
20 49 32 17
30 59 41 18
40 64 48 16
50 69 56 13
60 76 62 14
70 80 68 12
80 87 75 12
90 92 84 8
100 100 99 1
Arithmetic average 67.6 53.9 13.7

 Mathematical ages (years)

 Religious General Difference
Percentile schools schools

10 10.9 9.2 1.7
20 11.3 10.0 1.3
30 11.9 10.7 1.2
40 12.1 11.3 0.8
50 12.6 11.7 0.9
60 13.1 12.1 1.0
70 13.4 12.5 0.9
80 13.8 (b) (b)
90 14.2 13.6 0.6
100 .. .. ..
Arithmetic average 12.6 11.6 1.0

Notes: (a) The tests are marked out of 100 ('scores' or 'raw marks').
These marks are converted by DfES into National Curriculum 'Levels'
such that Level 2 corresponds to marks of 16-18; Level 3 to marks of
19-44; Level 4 (the NC Level 'expected' at this genealogical age
according to National Curriculum principia) to marks of 45-75; and
Level 5 to higher marks. Each National Curriculum Level corresponds
notionally to two years of learning. At each decile of pupils we have
here calculated an average mathematical age based on these conversion
factors. Some pupils did not take the written test (about 3.7 and 2.0
per cent in religious and LA schools respectively); for our purposes
notional marks were imputed, based on the parallel 'Teacher
Assessments' where available. For a very few pupils (about 1.0 and 0.9
per cent) no marks were available on either basis, and such pupils were
omitted from these calculations.

(b) Imputation of a mathematical age is difficult at a score of
precisely 75 per cent since, according to the official discrete
criteria (see previous note), a Level 4 is awarded at just below that
score, and a Level 5 at a score just above it. As a mark of caution to
the reader, a blank has here been shown.

Table 2. Sizes of religious and general schools in Barking
and Dagenham, 2003
 Religious General
 schools schools

No. of schools in Borough 7 28
Total no. of pupils in year 6 303 2076
Average size of school
 (pupils in year 6) 43 74
Average no. of forms
 of entry per school 1.9 3.0

NOTES

(1) See, for example, the report on the original government-sponsored study by Coleman et al. (1966); and his subsequent study with Hoffer and Kilgrove (1984), with a summary conclusion on p. 180). A more recent compendium of his research studies appeared in Coleman (1990), esp. ch. 16-18 (ch. 18 is especially interesting for our concerns here; it is jointly authored with T. Hoffer and A. M. Greeley; the final quotation above is from p. 269). For more recent US references, see Shookrai (1997); results are usually quoted in terms of passing some not-too-precisely defined hurdle, such as 'completing high school', or entering for the Scholastic Aptitude Test, rather than average scores for the whole age-group. Improving the numbers reaching an upper hurdle is not a wholly convincing measure of superior attainments since, as discussed below, that could be achieved at the cost of neglecting lower-attaining pupils.

(2) The study by Schagen et al. was based on school-averages, not on individual pupils. The many graphs they show of the relation between pupils' final attainments and (for example) their initial attainments at KS3 are based entirely on school-averages, and are far from showing the full range of pupils' attainments; nor do the graphs even show the underlying observations for individual schools--only the calculated regression line.

(3) DfES (2004), together with additional table 24 (from their statistics website). This Borough has made considerable educational progress in the past decade; by 2002 it was ahead of England as a whole at KS2 in science, about the same in mathematics, and a little below in English.

(4) The 1.0 years average advantage by religious over non-religious schools in Barking and Dagenham contrasts with the finding of only a five months' average advantage for all England as calculated by Dr Marks. A possible reason is the special mathematics teaching scheme, based on Continental practice, introduced in that Borough in the past decade (that scheme is in many ways more systematic than older schemes, and may fit in particularly well with the general orientation of religious schools). This issue need not be pursued here; it is sufficient for our purposes that the scores in table I have been calculated on strictly the same basis for both types of school, and that the same elements of imputation (see end of footnote (a) to table I) have entered into both sets of calculations as necessary.

REFERENCES

Burn, J., Marks, J., Pilkington, P. and Thompson, P. (2001), Faith in Education, London, Civitas.

Coleman, J.S. (1990), Equality and Achievement in Education Westview.

Coleman, J.S. et al. (1966), Equality of Educational Opportunity, US GPO, reprinted Arno 1979.

Coleman, J.S., Hoffer, T. and Kilgrove, S. (1984), High School Achievement: Public, Catholic and Private Schools Compared, NY, Basic.

DfES (2004), Statistical First Release, GFR 21/2004.

Prais, S.J. (1997), 'Whole-class teaching, school-readiness and pupils' mathematical attainments', Oxford Review of Education, 23, 3 p. 282.

Schagen, S. et al. (2002), The Impact of Specialist and Faith Schools on Performance, National Foundation for Educational Research.

Shookrai, N. H. (1997), 'Why Catholic schools spell success for America's inner-city children', Heritage Foundation Backgrounder, 1128, 30 June.

* Thanks are due to the Education Department of the London Borough of Barking and Dagenham, without whose great help on statistics this note could not have been prepared, and to NIESR for research facilities.