Eurostat A0

1
Luxembourg Income Study (LIS) asbl 17, rue des
Pommiers L-2343 Luxembourg City Tél (352) 26
00 30 20 Email kruten_at_lisproject.org Fax
(352) 26 00 30 30 Web www.lisproject.org
Luxembourg Income Study (LIS) asbl Thierry
Kruten Kruten_at_lisproject.org
Luxembourg, 26-27 October 2006
OECD Conference Assessing the feasibility of
micro-data access
2
LISSY - A Remote Access System What for ?

• Needs of individual information in order to test
  the effectiveness of policy and to set new
  policies implies obtain individual and possibly
  sensitive data which have to be gathered, managed
  and disseminated
• A remote access system offers researchers the
  possibility to analyze microdata from a remote
  location by submitting statistical queries.
  microdata are not physically distributed
  remaining under the control of the agency, while
  the users of the system can perform analyses at
  their own place of work
• The challenge
• The trade-off between a user-friendly system and
  a safe system is evidently present for remote
  access systems. Disclosure control limit the
  detail of the output

3
LISSY - A Remote Access System confidentiality,
user-friendliness and feasibility

• Confidentiality
• Depends on the data providers requirements
• Restriction of queries based on the user output.
  Those queries are filtered either automatically
  or manually (Review queue) on a query basis, and
  a query history database has been set up and
  maintained (Log files)
• User-friendliness
• Oriented towards users. Mainly Academic
  researchers.
• LIS users are mainly interested in more advanced
  statistical analyses and like to use specific
  softwares
• Relevant aspect of the system is speed. LISSY
  outputs are returned within a couple of minutes
  and the system is available 24 hours a day, 7
  days a week. (geographical constraint)

4
LISSY - A Remote Access System confidentiality,
user-friendliness and feasibility

• Feasibility
• Most important aspect of the feasibility is the
  security that can be attained (IT Audit)
• A technological answer through the implementation
  of a secure architecture (Firewall, electronic
  authorization via password, restrictions on some
  IT formats such as HTTP encapsulations etc.)
• But also a capacity for the maintenance and
  administration of the system. The automatic or
  manual handling and evaluation of queries
  determines the labour intensiveness of the remote
  access facility. Other tasks such as the
  management of the users database (update, pledge)
  or the technical administration of the system
  (back-up, technical maintenance, numerous
  statistics concerning system usage) are also
  labor intensive
• These aspects are all directly related to
  budgetary constraints which one of the most
  important items of the LIS budget

5
LISSY - A Remote Access System Confidentiality
policy

• Use of the data is restricted to Social Science
  Research purposes only.
• Access is limited to Researchers working for an
  academic, government or non-profit organization
• Users must register with the LIS and sign a
  pledge to obey the rules governing the use of the
  data
• Registered users receive an "user account" and
  "password" that are strictly personal. Therefore,
  listings generated from user requests will be
  returned ONLY to the email address that was
  registered by the user when applying for access
• Under no circumstances shall also a registered
  user make any attempt to locate and list any
  survey information to identify individual
  persons
• No direct access of any kind is permitted to the
  data or the LIS network. Users agree not to
  attempt in any way to copy individual records
  through listing them in program output or in any
  other fashion

6
LISSY - A Remote Access System Implementation
7
LISSY - A Remote Access System How Lissy works

• The operating system consisting of a series of
  software components which work together to
  receive, process and return statistical requests
• The users submit their statistical requests under
  the form of SAS, SPSS or STATA programs to LIS
  via the Internet mailing system
• The email requests contain the syntax created by
  the user for the specific statistical package
  used and a standardized header identifying the
  user
• user (your user id)
• password (your password)
• package (SPSS, SAS or STATA)
• project (database used)

8
LISSY - A Remote Access System How Lissy works

• The heart of the system is the job control
  component (Post Office). It manages the entire
  access mechanism.
• It retrieves the email requests from the mail
  server
• It prepares these requests for processing by
  checking for all security issues like clearly
  identifying a user, checking for the use of
  illegal statistical commands, check for the usage
  of sequences of commands or variables or any
  other combinations not allowed
• It returns any job that breaches security to the
  sender along with an error message explaining the
  violation
• It distributes the requests to the batch
  processor computers
• It returns the statistical results to the proper
  (registered) user email addresses
• It sends suspicious output to the review queue
  for manual review instead of returning results to
  the user
• And finally it maintains critical databases
  needed for the overall operation.

All the components of the system are physically
separated and at no moment is a user in direct
contact with the data
9
Luxembourg Income Study Any questions are
welcome

• When you can measure what you are speaking about
  and express it in numbers you know something
  about it. But when you cannot measure it or
  express it in numbers, you knowledge is of a
  meager and unsatisfactory kind.
• Kelvin sir William ThomsonBritish Mathematician
  et physician (Belfast, 1824 - Netherhall, 1907)

10
Luxembourg Income Study (LIS) A variety of
support services .

• LIS workshops (Work Hard - Play Hard !)
• LIS conducts annual training workshops 10 days
  pre- and post-doctoral workshop designed to
  introduce young scholars to comparative research
  in income distribution, poverty, and labour
  market outcomes using the LIS database
• Courses include a mixture of lectures and
  assistance and direction using the LIS database
  to explore research issues chosen by the
  participants
• Visiting scholars program
• Grants to the LIS data archives of micro-data and
  to the relevant data documentation are offered.
  The LIS staff is available for consultation,
  assistance and possible collaboration
• Direct on-site access will be allowed for
  datasets whose providers have given us their
  consent for such access