ESSnet W orkshop on Q - PowerPoint Presentation

Slide1

ESSnet Workshop on Quality of Multisource Statistics

Example of Slovenia: use of administrative data sources in statistical surveys

Barbara Kutin Slatnar

,

Environmental

Statistics

Division

-

Agriculture

,

Forestry

,

Fishing

and

Hunting

Statistics

Department

E:

barbara.kutin@gov.si

,

www.stat.si

Introduction

Simple

and

…

Complex

and

…

2

ContentsWhy we combine different sources in a statistical

survey?Quality elements of administrative data sourcesInfluence of useing multisource on survey implementing

flow

Conclusions

3

Why we combine different

sources in a statistical survey?4

?

 Example

Primary data collection (one statistical questionary)

Multisource statistical survey

Administartive data fully replased primary data collection

Respondents burden

↑

↓

↓

NonResponse

↑

↓

↓

Costs

?

?

↓

Knowledge (methodology, IT, other)

↓

↑

↑

People involved in survey implementation

↓

↑

↓

Complexity of data flow and data editing↓↑↓Time needed for survey implementation↑↑↓

Different

aspects have different weights !

Example: Comparison of costs between simple and multisource

statistical survey by method of data collection of primary data source

5

2) Quality elements of administrative data sources 6

Administrative sources, which can be used directly, have all the most important characteristics definitions harmonized with statistical (also

reference

dates

or

periods

)

Data from administrative source are available

in time Used administrative data sources contain a single unique identification of unitsSystem of data control

for key variables in administrative

source is sufficientEuropean statistical legislation that allows the use of high-quality, available administrative data sources

Possible access to the documentation of administrative data sourceForm of administartive data which allows easy use of information

7The

Statistical Act (National

level

)

which

provides the use of

administrative

data sources for official statistics It is good to know

in advance: use of administrative data sources is a procces-

lenghty start and daily cooperation.Statistical Office has experienced professionals in the field of methodology of agricultural statistics as well as in the IT field

cooperation between Statistical Office and

owner of adm

. data source is legally

formalized (signed agreement

and tehnical protocols

)

in the

process is

also involved

management of

the

Statistical

Office (

t

his is important especially in the beginning of the process

and on strategic level)participation of the relevant representatives of different ministries (owners of administrative registers) in the statistical advisory committees – if such committees existThe Statistical Office may occasionally organizes joint workshops (participants are all stakeholders: statisticiants, holders of administrative sources, sometimes also data users)What helps to achieve a certain quality targets of administrative data (timeliness, harmonized definitions, available documentation, sufficient data control, ...):

3) Influence

of useing multisource on survey implementing

flow

8

4)

Conclusions

:

What

we

should be aware?

- Guaranteed positive effects of used administrative data:

- Speeding up the collection of data - S

ubstantive harmonization of certain information by the various institutions providing data to users -

lower non-response

- In lot of cases, in

multisource surveys we

will achieve only p

artial relief of reporting units - Always

have backup

scenario for case

of ‚disappeared or changed‘ administartive

data source (statistics should be

undependent

)

- M

ore

complex

work on organization and supervision of the individual steps of the survey: requires specific steps to coordinate work in the broadest working invironment: - a different organizational approach of implementation of specific survey - in all stages of implementation we can expect much more methodological work - grater burden on the stage of technical integration and adoption of data from different data sources

10

It is very complex, but when

you

implement

thoughtful

process everything runs

smoothly.