In order for LISSY to properly process user requests, a few exceptions to the usual statistical batch program syntax are required

Dataset Calling

LISSY relies on a three-stage built-in alias to access LIS micro-datasets. Here are examples of aliases properly identified by LISSY

  • &AU01h to access the LIS 2001 household Australian dataset (SAS)
  • $se02w to access the LWS 2002 household Swedish dataset (Stata)
  • US00e to access the LES 2000 US labor force dataset (SPSS)

The general rules for calling a data set are as follows:

  1. Start with a statistical package specific heading to point to the alias:
    PACKAGEHEADING
    SAS&
    Stata$
    SPSS 
  2. This is followed by a dataset abbreviation concatenating the two-digit ISO country code with the last two-digits of the reference year. The country and year four-digit abbreviations can be found here.
  3. End the alias with the one- to two-letter abbreviation used to identify the specific type of dataset within each project:
    PROJECT DATASET TYPE LETTER CODE
    LIS Household File
    Person File
    Household Shadow File
    Person Shadow File
    Wave V Household File (release 1)
    Wave V Person File (release 1)    		 h
    p
    hs
    ps
    hr
    pr
    LWS Household File w
    LES True Labour Force File
    Integrated Labour Force File    		 e
    l

Warnings

  • Please note that Stata is case-sensitive. Be certain to refer to Stata aliases in lower case.
  • For SAS users working with the LIS wave V release 1 datasets, a reference to a format catalog is mandatory. To prevent any SAS formatting errors, make use of the OPTION NOFMTERR SAS option.

Data Confidentiality Protection

Use of illegal commands

To preserve the confidentiality of information pertaining to individuals and/or households in the micro-databases, LISSY prohibits the use of the following commands:

  • PACKAGELIST OF COMMAND
    SASprint, %sysexec
    Statalist, set memory, shell
    SPSS&print, list

Depending which job submission method is used, LISSY will either display a dialog box (submission via JSI) or returns the job to the sender (submission via email) along with an error message explaining the violation: 

The job you have submitted did not pass the security check.

List of problems detected :

_____ERROR_FOUND_____In your job you used the command : << print >> that is not allowed
You are not allowed to print unit of records

Addtional LISSY Checks

In addition to checking for illegal commands, LISSY also filters submissions based on the usage of sequences of commands and/or variables that would end up breaching the rules on data confidentiality. For instance, using commands that display frequencies on continuous variables (e.g: income variables) will be detected by LISSY.

LISSY automatically puts the job in a security review area to be manually reviewed by the staff. If the LIS staff modifies the job in order to maintain confidentiality of the data, LISSY will send back the listing along with the following message: 

#----------------------------------------#
# Your job has been reviewed #
#----------------------------------------#

If the listing contained confidential data, it will be discarded. LISSY will then send the following message: 

#----------------------------------------#
#  Your job has been refused  #
#----------------------------------------#

Long Listings

If the size of a listing is larger than a given limit for a statistical package, the job is also automatically put in the security review area to ensure that there is no attempt to identify individual-level micro-data. If the LIS staff modifies the job, LISSY will send back the listing along with the following message: 

#---------------------------------------------#
# Your listing has been reviewed  #
#---------------------------------------------#

If the job listing is exceptionally long, the job is stopped and automatically discarded. In that case, LISSY will send the following notice to the user: 

#----------------------------------------#
#  Your job has been refused  #
#----------------------------------------#

SAS and Stata tend to generate rather long listings. The Luxembourg Income Study recommends the use of the OPTIONS nosource nonotes (in SAS) and nolog, quietly or noisily (in Stata) to shorten LISSY output.

Project Documentation

In spite of the LIS harmonisation, all datasets and variables are not exactly the same across countries and years. For instance, not all variables are available in each and every dataset. If they are available, they may include slightly different content or have different value codes.

The Luxembourg Income Study strongly recommends that users carefully check the project documentation for each dataset used. Detailed documentation for the LIS, LWS and LES projects is available on-line.