Thomas Burg Marcus Hudec Content 2 Starting Point Canonical Dimensions for a Typology of Statistical Products Template Examples Impact on Quality Reporting Conclusions amp Next ID: 231554
Download Presentation The PPT/PDF document "Typology of products in Official Statist..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Typology of products in Official Statistics
Thomas BurgMarcus Hudec Slide2
Content
2Starting Point
Canonical Dimensions for a Typology of Statistical ProductsTemplate
ExamplesImpact on Quality ReportingConclusions & Next
Steps
© Burg & Hudec
Vienna, June 3rd 2014Slide3
Content
3Starting
PointCanonical Dimensions
for a Typology
of
Statistical Products
Template
Examples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide4
Starting Point
4
Classical Approach: One-dimensional Type of Statistics Primary Statistics – Secondary StatisticsDeviation between official statistics and academic statistics
Eurostat handbook on Quality Reports - Sample Survey
- Census
- Statistical process using administrative sources
- Statistical process involving multiple data sources
- Price or any other
economic
index processes
- Statistical Compilation
Vienna, June 3rd 2014
© Burg & HudecSlide5
Types of Statistics
5
Vienna, June 3rd 2014
© Burg & HudecSlide6
Content
6Starting
PointCanonical Dimensions for a Typology of Statistical Products
TemplateExamples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide7
Canonical Dimensions
7Three dimensional approach
Data CollectionData ProcessingData Presentation ??Each dimension having its own characterization
© Burg & Hudec
Vienna, June 3rd 2014Slide8
Data Collection
8
Data can be collected having in mind two different purposes:1. Subject of Statistic 2. Auxiliary InformationPossible data sources for Statistical ProductsSurvey
RespondentsAdministrative Data
Non-Statistical purpose
Register Data
Maintained by NSI
Existing Data
Collected for other product New Data Sources Big Data“
© Burg & Hudec
Vienna, June 3rd 2014Slide9
Data Processing (I)
9
Simple Aggregation(„Normal processing“) Modell Based Calculations
Accounting
Data
Matching
© Burg & Hudec
Vienna, June 3rd 2014Slide10
Data Processing (II)
10Model Based processing
Can be used for direct calculation but as well at certain product steps aiming to enhance qualityBroad variety but some are typical in official statistics
© Burg & Hudec
Vienna, June 3rd 2014
Weighting of Sampling Schemes
Small Area Estimation
Index Calculations
Forecasting Methods
Index Calculations
Data Validation Techniques
Disaggregation
Statistical Disclosure Control
Flash Estimation
Backcasting
Methods
Imputation TechniquesSlide11
Data Presentation
11© Burg &
HudecVienna, June 3rd 2014Classical Statistical Tables
MapsIndicators
Systems of Accounts
Difficult to assign or rather „not to assign“!Slide12
Content
12Starting
PointCanonical Dimensions
for a Typology
of
Statistical Products
Template
Examples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide13
Template (I)
13© Burg & Hudec
Vienna, June 3rd 2014Slide14
Template (II)
14© Burg & Hudec
Vienna, June 3rd 2014Slide15
Content
15Starting
PointCanonical Dimensions
for a Typology
of
Statistical Products
Template
Examples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide16
EU SILC (I)
16© Burg & Hudec
Vienna, June 3rd 2014Slide17
EU SILC (II)
17© Burg & Hudec
Vienna, June 3rd 2014Slide18
Register Based Labour Market
Statistics (I) 18
© Burg & HudecVienna, June 3rd 2014Slide19
Register Based Labour Market Statistics
(II)19
© Burg & HudecVienna, June 3rd 2014Slide20
Content
20Starting
PointCanonical Dimensions
for a Typology
of
Statistical Products
Template
Examples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide21
Impact on Quality Reporting
21
Set of Metadata relevant for user depends on characteristics of the Statistical ProductAll quality dimensions are concerned but first of allaccuracy is a topic.Certain expectations on quality reporting
© Burg & Hudec
Vienna, June 3rd 2014Slide22
Data Sources
22© Burg & Hudec
Vienna, June 3rd 2014Slide23
Processing
23© Burg & Hudec
Vienna, June 3rd 2014
Simple Aggregation
Model
Based
Processing
Accounting
Data
Matching
Availability
Model DiganositcsMeasurement
ErrorsMatching rates
Completness
of Metadata
Goodness of Fit
Top Down vs. Bottom up
Adequcy
of Units
Description of Methods
Misclassification errors
Homogeneity of underlying concepts
Analysis of sensitivity
Strength of associationSlide24
Data Presentation
24© Burg & Hudec
Vienna, June 3rd 2014Contents of Quality report not dependent on characteristics
AccessibilityClarityTimeliness
Revisions
Restrictions caused by Statistical disclosure controlSlide25
Content
25Starting
PointCanonical Dimensions
for a Typology
of
Statistical Products
Template
Examples
Impact on Quality Reporting
Conclusions
© Burg & Hudec
Vienna, June 3rd 2014Slide26
Main Conclusions
26© Burg & Hudec
Vienna, June 3rd 2014
One dimensional approach of assigning a type of statistics is not sufficient
Canonical dimensions can describe the characteristics of a statistical product
Characterization of product has impact on set of metadata and expectations on quality reportingSlide27
Next Steps
27© Burg & Hudec
Vienna, June 3rd 2014
Sharpening the proposal Completeness, exact definition etc..
Applying the concept to Standard-Documentation of Statistics Austria