Using the NPD: Some pitfalls and issues

1
Using the NPD Some pitfalls and issues

• Presentation to PLUG
• 13 September 2006
• Andrew Ray
• Schools Analysis and Research Division, DfES

2
Statistical bulletins based on analysis of the
NPD
3
PLASC variables
PLASC

• FSM
• Ethnicity
• SEN
• Looked after children
• EAL
• Not covered here variables linked to postcode

4
FSM is limited as a proxy for social class
FSM
5
Code changes 2002 ? 03
Ethnicity
Data taken from Ray Godfreys Changes in
Ethnicity Codes in the Pupil Level Annual Schools
Census 2002-2003
6
Local Authority variation in special school
provision
SEN
7
Under-reporting of looked after children
Looked after children
Source PLASC quality report on pupils in care
8
Complex relationship with ethnicity
EAL
9
Attainment data
Attainment

• Changing test structures
• Changing qualifications
• Changing point systems
• Ages and Key Stages
• Ceiling effects
• Use of marks data for sublevels

10
Changes to KS3 English levels
Levels and tests
11
Key Stage 2 artificial increase in Bs
Levels and tests
12
Key Stage 1 the move to teacher assessment
Levels and tests
13
Additional qualifications
Quals
14
New GCSE points system introduced in 2004
Points
Grade Old Points New Points
A 8 58
A 7 52
B 6 46
C 5 40
D 4 34
E 3 28
F 2 22
G 1 16
U/X/Q 0 0
Gap between fail and G has increased
New Points (Old Points x 6) (10 x number of
passes at G or above) (but points not awarded for
all qualifications under the old system)
15
Age 15 or KS4
Ages and stages

• End of KS4 reporting allows for more flexible
  rates of learning.
• 98.8 of pupils at end of KS4 are 15, 0.3 were
  14 and 0.9 were 16.

16
Test scores are truncated
Ceilings
17
Marks and levels
Sublevels
18
Sublevels dont directly correspond to marks
Sublevels
19
Coverage of the NPD
The NPD

• Pupils with incomplete records (2 examples)
• Not covered here
• the extent of coverage of the independent sector
• pitfalls in choosing coverage, e.g. when to
  include selective schools or schools that are
  semi-independent (e.g. City Technology Colleges)
• whether or not to impute missing values

20
Pupils with no PLASC or KS4 records
The NPD
Source Trends in Attainment Gaps bulletin
Appendix D
21
Pupils with no PLASC records
The NPD
Source Trends in Attainment Gaps bulletin
Appendix D
22
Issues for PLUG

• Future speakers experts on PLASC or test data?
• Documentation what improvements can be made?
• Further investigation are there useful analyses
  that could illuminate some data issues?
• Can any common standards be agreed, or should
  everything be left to individual analyses?

23
Useful links

• Bulletins
• Trends in Attainment Gaps http//www.dfes.gov.uk/r
  sgateway/DB/SBU/b000665/index.shtml
• Characteristics of Low Attainers
  http//www.dfes.gov.uk/rsgateway/DB/SBU/b000588/in
  dex.shtml
• Variation in pupil progress http//www.dfes.gov.uk
  /rsgateway/DB/SBU/b000481/index.shtml
• Progress by pupil characteristics
  http//www.dfes.gov.uk/rsgateway/DB/SBU/b000402/in
  dex.shtm
• Other key publications
• Results by pupil characteristics 2005
  http//www.dfes.gov.uk/rsgateway/DB/SFR/s000640/in
  dex.shtml
• Ethnicity and education topic note
  http//publications.teachernet.gov.uk/default.aspx
  ?PageFunctionproductdetailsPageModepublications
  ProductIdDFES02082006