Evidence check: Grammar schools Contents

3Admissions and testing

Self-selecting intake

36.We were told by the witnesses that children from lower socio-economic backgrounds are less likely to attend academically selective schools. Grammar schools have typically selected an ‘advantaged’ intake, and children from lower socio-economic families are less likely to pass the 11-plus for a variety of reasons. The Minister told us that this, which we might refer to as ‘the socio-economic entry gap’, could be because less well-off pupils are less likely to apply due to a lack of primary school support and/or encouragement to do so; or having parents who would prefer their children did not attend a selective school: “They may think it would be socially difficult for those children”,60 the Minister told us. He emphasised the importance of new or expanded grammar schools working with non-selective primary schools to remedy this problem. Engagement with outreach programmes could “dispel these myths and encourag[e] the parents to apply for [grammar school] places.”61

37.Mr Sibieta said that, in terms of the social mix and socio-economic mix, the evidence was clear: “only 3% of children who attend grammar schools are eligible for free school meals, which compares with about 17% of children in the area as a whole”.62 He described some reasons for this discrepancy. First, less well-off children demonstrate lower levels of academic achievement by age 11—even by age 3 and age 5—than their selectively-schooled counterparts. This attainment gap widens throughout childhood. Professor Vignoles added that historic evidence shows a large degree of misclassification: children demonstrating “levels of achievement that you would expect to find in grammar schools [ … ] were not in grammar schools”.63 This, Professor Vignoles told us, was due to self-selection.

Age of selection

38.Professor Jesson pointed out that the Green Paper suggests there is nothing set in stone about selection at only age 11,64 and the Government has indicated that testing at different ages is part of their policy-scoping. The Minister agreed that 11 “may be too young for some children”,65 which was why the Government is seeking “flexibility in the age of entry into a selective school”,66 at age 14 or 16 for example. Professor Jesson regarded this as “a crucially important element”,67 namely that selection at different ages could help to ensure the envisaged expansion is comprehensive and inclusive.

39.Professor Jesson told us that in Leicestershire the system was designed so that it was open for everyone. In that county children “transferred into a high school at the age of nine. They stayed there until 14. If they wanted to go to a grammar school, they transferred”.68 We asked about existing research into the impact on children of their moving schools mid-studies. Dr Allen spoke about the impact of delaying entry to 14. She said that “we will not see more free school meal children passing a 14-plus than an 11-plus. We will probably see fewer”.69 Dr Allen went on to tell us that the evidence from Cranbrook—a grammar school in Kent—previously selected at 13, but changed to 11 because later selection had led to difficulties with recruitment.70 She emphasised that, at present, 2% per year join the current grammar school sector between the ages of 13 and 14, and much higher in some grammar schools.71 Dr Allen reiterated her overriding concern that “ad hoc entry points [mean that] some families are more motivated than others to enter”.72

Tutor-proof tests

40.The existence of self-selection has led to the desire to reform admissions and the need to develop entrance tests that are ‘tutor-proof’—so that ability to pay for additional, private assistance is not a significant influencer on a child’s ability to pass an exam. Professor Jesson told us that the evidence for the old-style 11-plus tests was “pretty poor”.73 We asked the academic experts about the prospects of a better test being created in the future. Dr Allen urged caution regarding the design of ‘smarter’ tests, noting that it was important to consider what indeed was being measured: future attainment, or current achievement.74 Dr Allen went on to cite examples of recent research that challenges the notion that ‘ability’ is fixed: “everything we are starting to learn about the brain [ … ] suggests that is not right.”75 In particular:

research came out from the University College London Institute of Cognitive Neuroscience. They said that older teenagers and young adults were actually fantastically good at being trained to improve their fundamental maths skills and their reasoning abilities. They were better at improving than younger children, which provides evidence that the brain is malleable.76

41.In light of this evidence, and none provided to convince us to the contrary (that ability is fixed at 11), the notion that we can measure likely future academic achievement must be addressed by the Government. Account must be taken of evolutive ability. It also raises the prospect of students changing schools mid-way through their schooling, giving rise to potential disruption or at least instability during the formative years of a child’s development.77

42.Mr Sibieta identified a further issue with testing, namely that tests which are used for setting are capable of being remedied at a later date. If for example a child underperforms on test-day, they can be re-assessed relatively easily, and moved between sets.78 The same reassurance cannot be found in the school entrance context: mistakes could mean a child being placed in a school which is unsuitable for them. It is therefore at least more costly to correct mistakes or one-day underperformances. It is also noteworthy that movement seems to have been considered in a unidirectional manner—where children are moved to grammar schools, not removed from them. This might have implications for the capacity of these new grammar schools.

43.We asked the Government about its intended reforms to the testing process, namely how it could be ensured that advantage does not skew ability to pass in an unfair direction. Dr Leunig told us that “it is possible to level the playing field.”79 For example, the same principle as that which is applied when allowing children born in August a lower pass-mark. In theory, the same could be applied in the context of socio-economic background. “Even without a perfect, tutor-proof test, which [ … ] does not exist”,80 Dr Leunig told us that such adjustments were viable. Dr Leunig made a final important point on this topic, noting that empirical testing of this entrance system should be trialled: “try it and see whether it works better or worse than the predecessor. If it works better you build on it. If it works less well, you withdraw it”.81

44.If, as the Minister suggested, a tutor-proof test is a ‘holy grail’, selection tests should not be the only basis on which admissions to grammar schools are based. The Government has yet to demonstrate how an admissions system could be designed in a manner which would be immune to gaming, or being reduced to the ability to pay.


60 Q86

61 Ibid

62 Q13

63 Q12

64 Q31

65 Q96

66 Ibid

67 Q31

68 Q26

69 Q32

70 Ibid

71 Ibid

72 Ibid

73 Q26

74 Ibid

75 Ibid

76 Ibid

77 This was a point raised during Q32 [Professor Jesson]

78 Q29

79 Q91

80 Ibid

81 Q97




10 February 2017