Selection and achievement

1
Selection and achievement

• Results from the NCDS

2
The NCDS sample

• Entered secondary schools in 1969
• Few LEAs fully comprehensive so creaming
  occurred
• Adjusting for 11 year achievements aimed to deal
  with this
• Full report by Steedman (1980) unavailable some
  results in Chapter 12 of Fogelman et al.(Growing
  up in Great Britain, Macmillan, 1982)

3
Limitations

• The sparseness of the sampling means no school
  contextual data e.g. achievement mix
• Prior data only at 7 and 11 years so detailed
  trajectories unavailable
• No information on school and parental policies
  and choices

4
The politics

• Commissioned by Callaghan Government it was
  steered by a combination of academics (Chaired by
  Halsey) and civil servants.
• Subject to criticism by supporters of selective
  education Caroline Cox and John Marks
• Final report and steering group summary gave
  support to both sides of the debate

5
Results

• At 11 tests of Maths, Reading, General ability
  social class gender used
• At 16 tests of Maths and Reading used as
  (separate) outcomes
• Regression of 16 year scores on 11 year scores
  with slopes varying across school types
  (Comprehensive, Grammar, secondary modern)

6

Reading at 16 by reading at 11.
Comp
Grammar
Sec Mod
Sec Mod
7
Maths at 16 by Maths at 11

Grammar
Comp
Sec Mod
8
Composite 0 Level score by composite 11 year old
achievement
Grammar
Comp
Sec. Mod.
9
What might we now do differently?

• Use school contextual information very
  difficult with an unclustered sample
• Take account of school entry policies and
  parental choices relevant to creaming
• Fit a multilevel model but needs clustered data
• Fit different pupil level variations as a
  function of prior achievement and contextual
  variables