Longitudinal Studies Past, Present and Future

1
Longitudinal StudiesPast, Present and Future

• ESRC Commitment Long Term
• Two strands
• Select National Jewels
• Ad hoc Studies

2
Select (ESRC) National Jewels

• LS
• Cohorts
• BHPS
• These are funded for the data
  and given funds to
  promote analysis

3
Other Major Longitudinal Studies

• There are many
• BUT
• Access often difficult

4
Major Characteristics of all Longitudinal Studies

• Expensive
• Require Long Term Commitment
• Payoff not immediate

5
Pay off not immediate

• BUT
• There is much that can be done
• MILLENIUM COHORT
• There needs to be work to ensure
• maximum value for investment

6
Do we need any more national jewels?

• Institutions
• 46, 58, 70, ?, 01
• Regional Estimates
• Specific Demographic Groups
• Expense
• Use
• Decisions MUST be based on a clear assessment of
  the substantive questions which the data will,
  over time address.

7
National Data Strategy

• Need long term commitment to a core set of data
• Not all longitudinal
• Paul Boyles review
• National Datasets co-ordinator
• Policy Launch June
• Strategy Development over Autumn
• Feed into new ESRC Strategic Framework

8
ESRCs New Framework

• Capacity
• Delivery
• Engagement
• Performance

9
What we do

• The themes
• Economic Performance and development
• Environment and Human Behaviour
• Governance and Citizenship
• Knowledge, Communication and Learning
• Lifecourse, Lifestyles and Health
• Social Stability and Exclusion
• Work and Organisation
• Research resources

10
Fit into RCUK Ten Year Vision

• Some potential areas (NB EXTREMELY ILLUSTRATIVE)
• Economic Performance
• Population
• Individual Behaviour, Lifestyle and Health
• Public Service Delivery
• Education
• Our Environment and Food

11
Lowest Common Denominator

• All benefit from Longitudinal Data

12
Infrastructure for Next Ten Years

• Twenty first century Social Science requires
  twenty first century infrastructure
• Major area will be modelling individual and
  institutional data over time and space and from
  different sources
• Methodological Development
• merging
• disclosure control

13
Analysis

• 1980 Logistical Regression
• 1987 Current Status data
• 1990 Discrete Time Hazards (DTH)
• 1995 Multilevel DTM
• 1997 Multilevel Multinomial DTH
• 2001 Multiprocess Modelling
• 2003 Onwards
• e-social science
• multiple types

14
Final Summary

• LS Central to ESRC
• LS Central to Social Science
• BUT
• We need to be clear about focus and need