Transcript Innovations in Data Collection and Management

Innovations in Data Collection
and Management
Geoff Bascand
February 2009
Overview
• Innovations can help increase efficiency, reduce
respondent load and improve data quality
• This session will discuss:
–
–
–
–
Examples of innovation
Why innovation matters
Challenges in innovation
Opportunities
Modernisation of Statistics Production
• Common vision - reduce time in data collection
processing to provide more resource for analysis
• Key themes:
– The results have been mixed
– Still a real commitment to standardisation
– Less optimistic view on time to achieve standardisation
Q) Is there an opportunity for shared learning on
modernisation approaches?
Proposed Load Limits for businesses
Size of business
Maximum number of
Stats NZ data
collections
Maximum time taken
(hours)
Small
Three
Four
Medium
Four
Six
Large
Seven
Ten
Extra Large
No limit
No limit
Load hotspots by time taken
Hours taken
Large
Medium
Small
15 minutes
3 597
11 535
27 807
30 minutes
264
624
63 751
One hour
1 834
9 242
18 851
Two hours
3 919
6 717
2 111
Three hours
1 983
2 185
1 191
Four hours
1 030
293
2 051
Five hours
456
127
90
Six hours
227
41
37
Seven hours
134
4
16
Eight hours
59
4
1
Nine hours
30
1
10 hours
13
1
1
15 hours
25
3
4
20 hours
6
25 hours
3
1
25+ hours
4
3
2
Number of businesses surveyed and
respondent load
Number of businesses surveyed
Actual 2002-07, projected to 2012
(000)
300
Enterprises surveyed
(Projected)
Total time taken
(Projected)
250
200
150
100
50
0
2002
2003
2004
Source: Statistics NZ provisional time taken dataset
2005
2006
2007
2008
2009
2010
2011
2012
Respondent Load
• Need to ensure willing supply of information
• Common strategies for reducing load:
–
–
–
–
Demonstrating value of information collected
Reduce load on respondents
Make it easier to respond
Identify and manage areas of unreasonable load
Q) Should there be an international Respondent
Load standard?
Use of administrative data & standard
reporting
• Reduce direct survey activity and/or increase
range of statistics
• Innovations:
– register based Censuses
– strong relationships with software providers to enable
dynamic extraction of data to meet statistical needs
• Statistical methods can focus on administrative
data quality and plugging the gaps
Strong and targeted relationships
• Understand your respondents, and develop
customised solutions
• The Navajo nation example, helped influence:
– form design
– mode of data collection
– the value of promotion via third party partners e.g..
community based organisations
Q) Are we rigorous enough when measuring the
effectiveness of innovations?
CRM technologies
• Centralised customer management enables:
– Creation of targeted relationships
– Identification of areas of overlap and load
– Improved efficiency
• Innovations include combining call
management infrastructure and multi modal
collection
Q) Should we investigate developing an
integrated CRM for respondents and data
users?
Web collection
• Web collection can – reduce load and costs as
well as increase data quality
• Several good examples of web collection
• Standardisation and integration needed
between modes
Q) What are the opportunities for web collection
beyond Census?
Operations Research
• Use scarce resources wisely
• Use behavioural psychology and statistical
methods to:
–
–
–
–
Improve user and respondents experiences
Reduce costs in data collection
Data mining techniques
Optimising call times
Some Opportunities and a Challenge
• Opportunities
– Expert working group on standardisation of processes
and technology?
– An international Respondent Load standard?
– Integrated CRM for respondents and data users?
• Challenge
– To use rigour when measuring the effectiveness of
innovations