Internet Option in the 2010 Census of Korea Ki-Jong Woo Commissioner

advertisement
Company
LOGO
Counting the 7 Billion: The Future of Population
and Housing Censuses
Internet Option in the 2010 Census
of Korea
Ki-Jong Woo
Commissioner
Statistics Korea
1
Challenges and Benefits of Using Modern Technologies
Contents
1
Motivators
2
Internet Option for the 2010 Census
3
Internet Uptake Rates
4
Data Quality
5
Lessons Learned
2
1. Motivators
3
(1) More adverse census-taking environment
Rising cost of census-taking
- on average, cost increases by
1.5 times between censuses
$157.2M
$112.1M
$72.5M
$46.8M
A need to reduce the
cost of census-taking
1995
2000
2005
2010
1$ = 1,150 Won
4
Changes in household patterns (2010)
- one-person households
: 23.9%
- dual-earner households
: 37.2%
23.9%
20.0%
- elderly aged 65+
: 11.3%
15.5%
12.7%
164
Making it difficult to
interview households
1995
414
317
222
(In ‘000)
2000
2005
2010
5
Growing awareness of privacy
- Increasing non-response rate
3.6%
More and more people are
reluctant to cooperate with
2.8%
2.2%
1.7%
government
2006
2007
2008
2009
< Economically Active Population Survey >
6
(2) Opportunities for introducing modern technologies
Internet penetration rate:
- 81.6% (May, 2010)
Improved quality of administrative data
80.6%
81.6%
78.4%
- Population register
- Building register
72.2%
70.2%
- New address system, etc.
2002
2004
2006
2008
2010
< Korea Internet Security Agency >
7
Ample experience of conducting electronic surveys
by Statistics Korea
3
(2005)
6
(2006)
25
(2009)
27
(2010)
8
2. Internet Option
for the 2010 Census
9
(1) Outline of the 2010 Census
History
Conducted every five years since 1925.
2010 Population Census (18th )
Census Day
As of 0:00, November 1, 2010
Enumeration
Period
November 1 – 15 (15 days)
Census
Questionnaire
50 items (short form: 19, long form: 31)
(e-Census : Oct. 22-31, 10 days)
10
Child care
Citizenship
Year/Month
of Arrival
Multicultural
Family
Low
Birth Rate &
Aging Society
Number of children
mothers plan to have
Impairment
care
Low Carbon
Green growth
Social
Participation
Digital
Broadcasting
Social activities
IT facilities
Heating facility
Main means of
commuting
Use of
transportation
means
11
Workforce
119,000: 6,000 officials + 113,000 enumerators
Preliminary: Dec. 2010
Results Release
Short form: Jul. 2011 / Long form : Oct. 2011
$157.2 M
Budget
* 1995: $46.8M, 2000: $72.5M, 2005: $112.1M.
Publicity cost
$11.0M
67.5%
Labor cost
$106.1M
7.0%
7.6%
17.9%
Data processing
$11.9M
Census
activity & supplies
$28.1M
12
(2) Timeline for Internet Option Period
Internet
Option
Preparatory
Period
Oct. 19-21
Door-to-door
Visits
Oct. 22-31
Nov. 1-15
13
(3) Work Flow of the Internet Option
Preparation
Address DB, Household list, Housing DB
Public
Campaign
Create Internet Access Code
Pull & Push Strategy
Give Internet Access Code
Intent to
participate
Thank You
Note (e-mail)
Confirm
No
Yes
Fill Out the Form
Yes
Finished?
Automatic editing
No
Electronic management
Re-contact those who
haven’t completed the form
Yes
DB
• Send SMS to encourage
response
14
(4) Internet Access Code
Country
Internet Access Code
Digit
Canada
*Internet Access Code(15-digit)
Australia
*Census Form Number(13-digit)
*E-Census Number(12-digit)
Japan
*Government approval number(4digit)
*Questionnaire ID(11-digit)
*Verification Number(8-digit)
15
(5) Accessing the Online Form
Access the Internet
Begin the e-Census
For foreigners
Enter Internet Access Code
16
(6) Census Questionnaire for the Internet
Cover
Household
Household Members
Housing
17
(7) Input and Automatic Editing
Range and logic checks
Number of edit specifications : 88 for short form, 227 for long form
18
(8) Automated Response Management
Automatically manage the input process and report the status
Shown here
these households
have completed
the form
This household
has not completed
the form
19
(9) Ensure Security
e-Census System
Internet
Router
External
Firewall
Web Firewall
WEB
Firewall
External Network
DB
Firewall
Internal Firewall
EXT
WEB
Internet
WAS
Firewall
WEB
WAS
DB
User Security
Encoding SSL
WEB Sever
WEB Sever WAS Sever
DB Sever
User Security
Network Security
Web Service
Server Security
DB Security
(Level 1)
(Level 2)
(Level 3)
(Level 4)
(Level 5)
PC Security
 PC keyboard
 Web browser
 No capture
 Encoding SSL
Access Control
 Separating internal
and external
network
 Control irregular
traffic
Web Service Security
 Web Firewall
Security OS
Encoding Data
 Controlling accesses
 Encoding personal
to server
information
20
3. Internet Uptake Rate
21
(1) Overall Internet Uptake Rate : 47.9%
(2) By questionnaire and region
100%
Short
100%
Urban
47.7%
49.6%
100%
Long
49.3%
100%
Rural
40.4%
22
(3) Characteristics of Internet Participants
Daily internet response rate
Beginning period 5.4%, Ending Period 5.0%
(Thousand
Households)
1,000
943
Beginning Period
866
900
Ending Period
800
655
700
611
595
600
491
500
400
546
464
580
408
Weekend effect
300
619
329
258
209
172
200
194
146
101
100
115
56
0
10/19 10/20 10/21 10/2210/23 10/24 10/25 10/26 10/27 10/28 10/29 10/30 10/31 11/1 11/2 11/3 11/4 11/5 11/6 11/7
Preparation
Internet option period
Extension
23
Response rate by the time of day
8 pm~10 pm(16.4%), 10 pm~midnight(12.6%), 10 am~noon(12.1%)
(%)
18
16.4
16
14
11.6
11.0
12
12.6
12.0
12.1
10.7
10
7.9
8
6
4
3.6
1.3
2
0.6
0.3
0
0-2
2-4
4-6
6-8
8 - 10
10 - 12
12 - 14 14 - 16
16 - 18
18 - 20
20 - 22 22 - 24
24
Age group (Internet respondents)
Age
59.8%
54.9%
48.2%
35.3%
35.7%
~19
20~29
34.5%
30~39
40~49
50~59
60~
25
Education of household head (Internet respondents)
Education
59.8%
46.3%
33.4%
33.9%
Elementary
Middle School
High School
College+
26
Size and type of household
Size
Type
64.1%
64.4%
62.1%
53.9%
39.2%
38.3%
31.3%
1-person
hhs
2-person
hhs
3-person 4-person 5+person
hhs
hhs
hhs
Married couple Married couple
w/ children
w/o children
27
Type of housing and occupancy
Housing
Occupancy
59.3%
47.9%
51.7%
47.1%
49.8%
47.5%
43.5%
37.3%
36.0%
Detached Apartment Row house Apartment
Others
dwelling
unit in a house
Owned
Tenement
Rent
Rent-free
28
4. Data Quality
29
(1) Coverage Error
Post-enumeration Survey of the 2010 Census
The Internet option shows lower omission and duplication rates
Internet
Interview
Net omission rate
0.06
0.80
Omission rate
0.20
1.24
Duplication rate
0.14
0.44
30
(2) Content Error
Edit and Imputation
The Internet responses show lower error rates
<Error rates of Internet and interview option by major census items>
31
ANCOVA (Analysis of Covariance) analysis
Controlling for age, education, sex, occupancy,
and the type of living quarters,
Internet responses show less errors.
32
5. Lessons Learned
33
(1) Prove to be a cost-effective tool for census
Reduce personnel expenditure: by $14.2M
- 10,000 fewer enumerators: $3.9M saved
- Shorter enumeration period: 19 days
$10.3M saved
16 days
Save data processing cost: by $3.5M
1,774 fewer data capture operators : $3.5M saved
34
(2) Proven to be a convenient tool for surveys
Reduce enumerator’s visits to households
< Reasons for choosing the Internet option >
72.2%
11.7%
10.5%
2.5%
Convenience
Dislike visit by
enumerators
Incentive
Safety
(from add-on questionnaire on Internet option in 2010 Census)
35
Most participants were satisfied with the Internet option
70.9% (5.72M households) of the Internet participants responded.
- Of them, 73.5% responded “Satisfied”
(Unit: %)
Overall Degree of Satisfaction
1-5
Scale
Very
Satisfied
Satisfied
Neither Satisfied
Nor dissatisfied
Dissatisfied
Very
dissatisfied
No
response
Total
4.01
29.9
43.6
20.4
2.3
1.3
2.4
Short
4.06
31.7
44.1
19.1
1.8
1.0
2.3
Long
3.60
16.2
40.4
30.1
6.6
4.0
2.7
36
(3) Prove to obtain more accurate responses
The range and logic check functions of the Internet option
worked effectively to secure accurate response.
# of edit specifications : Long-form 227, Short-form 88
- # of mean error in short-form : 0.08 for Internet response vs1.83 for Interview
- # of mean error in long-form : 1.27 for Internet response vs 3.56 for Interview
Higher coverage according to the PES
Respondents with higher education are more likely to choose the
Internet option
37
(4) Success Factors
As of 2010, the highest level in the world
0.5% in the 2005 Korean Census
Singapore(15.0%, 2000; 38.0%, 2010), Canada(18.5%, 2006),
Norway(9.9%, 2001), Australia(9.0%, 2006; 37%, 2011),
New Zealand(7.0%, 2006), Switzerland(4.2%, 2000)
Seven
months later,
54.4% in the
2011 Census
of Canada
54.4% in the 2011 Census of Canada
47.9%
37.0%
4.2%
7.0%
Switzerland New Zealand
(2006)
(2000)
8.0%
Japan
(2010)
38.0%
9.9%
Norway
(2001)
Australia
(2011)
Singapore
(2010)
Korea
(2010)
Canada
(2011)
38
(4) Success Factors
High Internet penetration rate for households
: 81.6% as of 2010
User-friendly, well-designed questionnaire
→ convenient to respond
Stability of e-Census system
Competitive sprit for higher Internet participation among provinces
39
(4) Success Factors
Strategy for enhancing Internet responses
Pull
Incentives
- Hand out gifts by drawing
: LED TV(20), Notebook (80), Bicycle(240) …
- Give students two hour credits for volunteer work
Publicity Campaign
- Campaign for e-Census
Push
Internet first
Internet
Interview
40
Thank you!
Download