CAPMAS

1
EGYPT
Central egency for public mobilization and
statistics
CAPMAS
www.capmas.gov.eg
2
CAPMAS
EGYPT
CENSUS EVALUATION POST-ENUMERATION SURVEY
3
CAPMAS
Purpose of Census Evaluation
Measure accuracy (quality) of census data
Coverage error census omission and duplication
of households and persons
Content error nonresponse, response errors,
coding errors, etc.
4
CAPMAS
Objectives of Post-Enumeration Survey
Measure both census coverage and content errors.
PES evaluation results useful for improving
future census operation.
PES results assist census data users in
understanding relative magnitude of error.
PES estimates may be used for adjusting the
census population figures and population
projections
5
CAPMAS
PES Methodology
Independent enumeration of household and
population in sample of areas. - De jure or de
facto enumeration, depending on type of
census enumeration.
Two-Way matching of PES and census from for
sample areas.
6
CAPMAS
PES Methodology (continued)
Reconciliation visit following matching.
Dual-system estimation methodology - Based
on two independent sources of population data
census and PES.
7
CAPMAS
Timing of PES
Ideal to conduct PES about 2 to 4 weeks following
the census, to avoid much movement of population
and recall error
8
CAPMAS
Movement of Population between Census and PES
Non-movers persons in same household at time of
census and PES
In-movers persons who moved into a household
between the time of the census and the PES
Out-movers persons who move out of a household
between the census and the PES
9
CAPMAS
Enumeration Procedures for Moving Population
Procedure A identifies composition of household
at time of census
- Enumeration of non-movers and out-movers
- Advantage out-movers easier to match
- Disadvantage more difficult to obtain
information for out-movers
10
CAPMAS
Enumeration Procedures for Moving Population
(Continued)
Procedure B identifies all current household
members at time of PES
- Enumeration of non-movers and in-movers
- Advantage in-movers easier to match
- Disadvantage more difficult to match
in-movers to census records at previous residence
in different area
11
CAPMAS
Enumeration Procedures for Moving Population
(Continued)
Procedure C combination of Procedures A and B
- Enumeration of non-movers, in-movers and
out-movers
- Match non-movers and out-movers to obtain
match rate for movers
- Count number of in-movers
- Apply match rate for out-movers to number of
in-movers
- Advantage out-movers easier to match,
in-movers easier to identify less bias than
Procedure A, less cost than Procedure B
- Procedure C used for Egypt PES
12
CAPMAS
Defining Objectives for PES
Coverage error evaluation
- Establish geographic domains, age groups for
analysis
- Example geographic domains for Egypt PES
Cairo, other large cities (Alexandria, Port-Said,
Suez), region (Upper and Lower Egypt) by other
urban, rural
Content error measurement
- Identify items to be evaluated for content
error most important characteristics subject to
response error
- Generally avoid items which require detailed
coding such as industry and occupation
13
CAPMAS
PES Questionnaire Design
Select items from census from required for
matching household and persons
- Household identification, including address
- Person characteristics name, sex, age,
relationship to head of household, marital status
Information to establish moving status following
census
- Whether person was in household at time of
census
Items included for content error analysis
Shaded areas in PES questionnaire to enter census
responses for matched persons
14
CAPMAS
Sample Design for PES
Stratified area sample
Sampling frame list of census enumeration areas
(EAs) or segments
P-sample households enumerated in PES within
sample segments
E-sample households enumerated by census within
same sample segments
15
CAPMAS
Sample Design for PES (continued)
Stratification geographic domains of analysis
- Cairo, other large cities (Alexandria,
Port-Said, Suez), region (Upper and Lower Egypt)
by other urban, rural
- Geographic ordering of EAs with systematic
provides implicit stratification
Sample size and allocation
- Depends on precision requirements for domains
of analysis
- Minimum number of sample segments for each
domain of analysis ideally, 30 sample segments
16
CAPMAS
Egypt has done (PES) Twiss
The first for the 1968 population census which
carried out by (PSRC) population studies and
research center at CAPMAS
The second for the 2006 population census which
carried out by cabinet (information and
decision support center (IDSC))
The two (PES) done by the same concepts,
procedures, sample design
17
CAPMAS
Design
The design has six domains.
The (PSUs) primary sampling units shiakh has in
urban area, villages in rural area.
each sample PSU divided into segments each
regiment size has 100 households.
all households enumerated within sample segments.
18
CAPMAS
Matching Operation
Two- way matching between household and person
records in P-sample (PES) and E-sample (census)
Matching procedures should be tested and
described very clearly in matching manual.
Matching rules should be designed to minimize net
matching error rate.
19
CAPMAS
Reconciliation Visit
Follow up on persons enumerated in the census but
not the PES, to determine whether they were
correctly enumerated.
Verify cases of possible matches.
Determine final match status.
20
CAPMAS
Dual System Estimation
Based on assumption of independence between
census and PES.
Depends on two-way matching between census and
PES.
Includes estimate of population missing from both
census and PES.
21
CAPMAS
Dual System Estimation Model
22
CAPMAS
Dual System Estimate of Total population
23
www.capmas.gov.eg