Quality in the Swedish Business Database The Quality Survey 2004 Round Table Beijing 2004 Swedish presentation, session 5, 18 th Round Table, Beijing – October 21, 2004
What is quality? The different users of the register have different view of what good quality is….
What is quality? The different users of the register have different view of what good quality is…. and the only thing they have in common is that..
What is quality? The different users of the register have different view of what good quality is…. and the only thing they have in common is that.. non of them are satisfied with the quality
Another fact is that…. The Quality in the Business Register depend in big extent on the quality of the administrative sources
Quality work! To have a good quality, you have to Have good knowledge of user needs
Quality work! To have a good quality, you have to Have good knowledge of user needs Have good cooperation with the administrative sources
Quality work! To have a good quality, you have to Have good knowledge of user needs Have good cooperation with the administrative sources Have a register maintenance with focus on the user needs (Quality)
Quality work! To have a good quality, you have to Have good knowledge of user needs Have good cooperation with the administrative sources Have a register maintenance with focus on the user needs (Quality) You should be able to measure the quality and quality changes
The Quality Concept in Statistics Sweden
The Swedish Business Database Has a threefold purpose –Public Register –Frame for statistics –Part of the Register system
The Quality Concept in Statistics Sweden (1) Contents Statistical target characteristics Comprehensiveness (completeness)
The Quality Concept in Statistics Sweden (2) Contents Accuracy Overall accuracy Sources of inaccuracy Presentation of accuracy measures
The Quality Concept in Statistics Sweden (3) Contents Accuracy Timeliness Frequency Production time Punctuality
The Quality Concept in Statistics Sweden (4) Contents Accuracy Timeliness Comparability and Coherence Comparability over times Comparability between domains Comparability with other statistics
The Quality Concept in Statistics Sweden (5) Contents Accuracy Timeliness Comparability and Coherence Availability Dissemination forms Presentation Documentation Access to micro data Information service
The Quality knowledge 1. No knowledge about the quality component.
The Quality knowledge 1. No knowledge about the quality component. 2. Vague knowledge about the presence of errors
The Quality knowledge 1. No knowledge about the quality component. 2. Vague knowledge about the presence of errors 3. Good descriptions of the processes, which make it possible to have a picture of the reliability
The Quality knowledge 1. No knowledge about the quality component. 2. Vague knowledge about the presence of errors 3. Good descriptions of the processes, which make it possible to have a picture of the reliability 4. Vague quality indicators, which are based on remarks made by users.
The Quality knowledge 1. No knowledge about the quality component. 2. Vague knowledge about the presence of errors 3. Good descriptions of the processes, which make it possible to have a picture of the reliability 4. Vague quality indicators, which are based on remarks made by users. 5. Quantitative indicators on example coverage and corrections made
The Quality knowledge 1. No knowledge about the quality component. 2. Vague knowledge about the presence of errors 3. Good descriptions of the processes, which make it possible to have a picture of the reliability 4. Vague quality indicators, which are based on remarks made by users. 5. Quantitative indicators on example coverage and corrections made 6. Estimated errors by means of systematic observations, evaluations etc.
Construction of a quality survey Purpose: Measure accuracy
Construction of a quality survey Purpose: Measure accuracy Method: Sending out pre-printed information; the enterprises were asked to correct and make complements The information was later on coded at Statistics Sweden
Construction of a quality survey Purpose: Measure accuracy Method: Sending out pre-printed information; the enterprises were asked to correct and make complements The information was later on coded at Statistics Sweden Pragmatic approach
Construction of a quality survey Variables: State of activity (Active or not) Name Postal address Address of location Telephone Fax Number of Employees (Interval) Activity code (SNI2002, which are on 4- digit level corresponding with NACE rev1.1)
Construction of the sample Sample: 1985 Local Units 7 groups of activity (SNI) 3 Size-classes (employees) A and B not included
Construction of the sample Sample: 1985 Local Units 7 groups of activity (SNI) 3 Size-classes (employees) Response rate 81%, which gave Confident Interval on ca +/- 1,5 Percentage units if the answer was 95%
Result of the Quality Survey State of activity (98.7%) Activity code (95% incl. No code 87%) Size-class (94%) Postal Address (94%) Location Address (91% incl. No information 87%) Phone number (81% incl. No number 69%)
Evaluation of the method A cost efficient method
Evaluation of the method A cost efficient method Reliability: High reliability of the result on the “contact” variable (and size-class) It could be questioned if the method was good enough for measure the quality on the activity code
Comparison with other results Service statistics
Comparison with other results Service statistics 63.3, Travel agencies etc, (12,8% vs. 1.8%) 70, excl. 70.3, Real estate, (3.3% vs. 2.8%) 93, Other services, (3.9% vs. 1.0%)
Comparison with other results Service statistics 63.3, Travel agencies etc, (12,8% vs. 1.8%) 70, excl. 70.3, Real estate, (3.3% vs. 2.8%) 93, Other services, (3.9% vs. 1.0%) Occupation Register
Comparison with other results Service statistics 63.3, Travel agencies etc, (12,8% vs. 1.8%) 70, excl. 70.3, Real estate, (3.3% vs. 2.8%) 93, Other services, (3.9% vs. 1.0%) Occupation Register A,B and F, Agriculture fishing and construction, (12% vs. 5.6%)
Future quality work More contacts with both users and the sources Implement of short term and annual quality indicators Making the Business Database quality survey to a annual survey Making an annual publication based on Statistics Sweden’s quality concept
Number of LEGAL Units in the Swedish Business Register
Percent of enterprises and employees in size class of employees
From the VAT-register it can be read
Number of active units in the Swedish Business Register Legal unitsmore than % Limited enterprises 57 % Sole proprietorships 0,2 % public sector 16 % other legal forms Local units more than % Limited enterprises 52 % Sole proprietorships 5 % public sector 15 % other legal forms
Comparability over time (1) Size-classes
Comparability over time (2) Activity code
Comprehensiveness (1) LU with Activity code
Comprehensiveness (2) LU with Phone Number
Comprehensiveness (3) LU with Location Address
Cake index