The relationship between data quality and quality profile of the process of territorial data collect

1
The relationship between data quality and quality
profile of the process of territorial data
collection in Italian Consumer Price
survey Federico Polidoro (Istat,
polidoro_at_Istat.it) Rosabel Ricci (Istat,
roricci_at_Istat.it) Anna Maria Sgamba (Istat,
sgamba_at_Istat.it)
2
outline of presentation
subjectobjectiveapproachmain
itemsconclusive remarks
quality in Consumer Price Survey testing a set
of indicators concerning production process and
data quality non parametric approach

• recent survey on Consumer Price territorial data
  collection net
• measuring the quality indicators approach
• main results

3
recent survey on Consumer Price territorial data
collection net
in the Italian Consumer Price survey data are
collected in two different ways
context

• at central level, directly by Istat
• at territorial level, by Municipal Offices of
  Statistics (MOS) in the chief towns participating
  in the survey

central data collection 20
territorial data collection regards 80 of the
basket of products
territorial data collection 80
4
recent survey on Consumer Price territorial data
collection net
between 2004 and 2005 Istat has carried out a
survey on Italian territorial net of price data
collection
analysing the process of territorial data
collection in order to build a system to evaluate
and monitor some aspects of the quality of CPI
statistics
purpose
5
recent survey on Consumer Price territorial data
collection net
Outline of Istat survey on MOS
coverage method survey unit method of data
collection reference period of data survey
date
entire national territory census MOS
participating in the CP survey direct, with
questionnaire (2004) September 16 - November 15
(complete cycle of CP collection data
collection) 2004 December
6
recent survey on Consumer Price territorial data
collection net
data coming from Istat survey on MOS

• technical and human resources available for the
  price data collection
• organization and management of the work process
  of price data collection
• costs of activity

7
recent survey on Consumer Price territorial data
collection net
interviewed MOS 91 answering MOS 86
answering MOS 94.5

• 86 MOS have replied including
• 82 out of 85 MOS participating in the CPI
  calculation in the 2004
• all 30 MOS disseminating autonomously local CPI
  in the 2004
• 5 MOS have not answered including
• 3 participating in the CPI calculation in the 2004

8
the Istat survey on MSO the main results on
technical and human resources
Consumer price data collection municipal net by
demographic width of the chief towns (average and
percentage values)
9
the Istat survey on MSO the main results on
technical and human resources
Consumer price data collection municipal net by
geographical area (average and percentage values)
10
recent survey on Consumer Price territorial data
collection net
on the basis of the results on human resources

• shortage of resources available for the price
  data collection, that is concentrated in specific
  areas
• the human resources available are not dedicated
  exclusively to the data collection in the field,
  neither to the CP domain
• the relevant resort to external resources that
  shows the growing difficulty by MOS in finding
  internal resources adequate to the tasks required

11
recent survey on Consumer Price territorial data
collection net
consequently the data coming from the Istat
survey on MOS can allow
to identify the critical aspects of the process
of territorial data collection where to action to
improve qualityto build a set of basic
indicator to measure the quality profile of the
process of territorial data collectionto stress
the relationship between data quality and quality
profile of the process of territorial data
collection to define an unique system to evaluate
and monitor the quality of CPI
?
?
?
12
measuring the quality indicators approach
build quantitative indicators for comparable
evaluation (data quality / data collection
process)
operative steps

• basic indicators
• standardisation of basic indicators
• aggregation of indicators
• ranking

13
basic indicators
three classes of indicators

• sample coverage (SC)
• data collection infrastructure (DCI)
• ? micro data accuracy (DQ)

14
indicators of sample coverage
so sample outlets tmp total municipal
population srh sample rentals for
housing tmt total municipal tenants epm
amount elementary prices collected in modern
distribution tep total amount of elementary
prices
the higher the indicator is the better the
representative is
15
indicators of data collection infrastructure
pc personal computer in municipal bureau
dedicated to CP survey pe municipal employees
in charge of CP survey ppc palm personal
computer available dc data collectors involved
in collecting prices in the field so sample
outlets hw hours worked by data collectors ms
municipal surface
the higher the indicator is the better the data
collection infrastructure is
16
indicators of micro data accuracy
nbvi elementary items for which information
neither on brand nor on variety are
available tepmq amount of monthly/quarterly
items tnr temporary non responses taep
amount of elementary price spdpm substitutions
with price of the previous month of new
elementary items different from the current
price ts amount of substitutions rep reduced
elementary prices
the higher the indicator is the better the micro
data accuracy is
17
standardisation and aggregation of basic
indicators
standardisation
i (indicator) 1 12 h (chief town)
1.84 variation between 0 and 1 easy ranking
aggregation
18
regional area synthetic indicators
regional level
P resident population in each Province n
chief towns in the region participating in
the survey r (region) 1.20
indices with base Italy 100
19
regional and geographical area synthetic
indicators
geographical area level
P resident population in each Province n
chief towns in the geographical area
participating in the survey a (geographical
area) 1.5
indices with base Italy100
20
ranking
ranking analysis by chief town by regional
area by geographical area
cograduation between ranking Spearman index
21
main results
regional areas ranking
22
main results
geographical areas ranking
23
conclusive remarks
confirmation of empirical evidences coming from
the current activity high territorial
variability for each aggregate indicators
?
?
SC ? 73,1 DCI ? 142,1 DQ ?
206,8 GI ? 76,2
at regional level
24
conclusive remarks
very critical situation for areas A2 and A5 for
area A5 inadequacy of resources available for the
amplitude of sample (vice versa for area A1) good
performance for areas A3 and A4 that show the
highest value for the GI
?
?
?
25
conclusive remarks
at level of geographical area discordance between
indicator of SC and indicator of DCI (?
-0,5) at level of geographical area high
concordance between indicator of DCI and
indicator of DQ (? 0,7) deepening the
indicator approach (advantages and disadvantages)
?
?
?
26
conclusive remarks
relevance of the results of the survey on
territorial net for data collection relevance of
the correct registration of information about
brand and variety relevance of the system of
flags for each event concerning micro
data implementing a current system for monitoring
the production process and micro data quality
?
?
?
?