In this document you will find general answers to TEN important questions:
1.
Why did we revise our data template and re-release the Wave 5 data?
2. When will
these Wave V.2 datasets be made available?
3. How can you
access the Wave V.2 datasets?
4.
What are the main substantive changes in Wave V.2, compared to the Wave 5.1
datasets?
5. How has the LIS file
structure changed?
6. Is there new documentation?
7.
Which LIS datasets will be “lissified” in the new template?
8.
What should you do if you want to continue to use the Wave 5.1 data, and where
will you find the documentation for those datasets?
9.
What should you do if you need help using the Wave V.2 datasets?
10.
What should you do if you have feedback for our staff?
____________________________________________________
1. Why did we revise our data template and re-release the Wave 5 data?
After an extensive and comprehensive internal evaluation, one that spanned nearly three years, we decided to revise our basic data “lissification” template. By “template”, we refer to the general set of rules and practices that we use when we “lissify” microdata.
Several factors catalyzed this template revision and, ultimately, prompted us to create a second release of our Wave 5 data.
Here are the highlights of what motivated us:
□ After the release of Wave 4 of the LIS data, the decision was made to stop the Luxembourg Employment Study (LES) project. The LES project, which began with Wave 3, acquired and harmonized data from labour force surveys and from other types of surveys with detailed labour force data. The LES project was ended primarily because we realized that the data were used infrequently. Many of our data users told us that they would find it more useful to have labour market data that could be linked to household income data, even if that meant losing volume and detail in the labour force data. That prompted us to shift our efforts towards “harvesting” as much labour market data as possible from the surveys (mostly income surveys) that are contained in the LIS database, and, eventually, to end the separate LES project. In response to our users’ concerns and preferences, we first created the Wave 5.1 “integrated files”, which contain labour market data from within the LIS surveys, data that can be linked to the income data. Wave V.2 extends this strategy by adding even more labour market data, increasing the standardization of these data across datasets, and placing the labour market variables directly within the LIS person-level data files.
□ During the more than two decades that LIS has been operating, the world around us changed, and, in particular, some consequential policy reforms unfolded. Yet our variables and “lissification” rules remained, for the most part, static.
Between 2004 and 2006, our staff assessed the old-age pensions operating in the LIS participating countries, with an eye toward recent changes and reforms. They then revised our rules for classifying old-age pensions to account for changing pension rules and structures, especially the shifting boundaries between private and public provisions.
Likewise, in the past two decades, child and family programs changed extensively in many countries, especially in Europe, as national governments added new forms of family leave, child care benefits, caregiver supports, and so forth. These changes dramatically complicated the classification of these programs. In response, our staff refined and formalized the rules used to “lissify” income (cash, near-cash, and non-cash) from these family programs.
□ Not only did social policy change, so did the contents and structures of our surveys. Our universe of datasets shifted such that more of our income data arrive at LIS measured as net income. Recognizing that shift caused us to re-organize and refine the way that we distinguish net from gross income, including revising our LIS summary variables.
□ As LIS has matured and grown, the use of our microdata has shifted in countless directions. The changes that we have witnessed include an increase in analyses conducted at the person level. That prompted us to revisit the extent to which we can provide disaggregated income data for persons within households. We concluded that we could go further than we have in the past.
□ Finally, at LIS, we have simply learned a huge amount over the years! We have accumulated more than two decades of hard-earned knowledge about several aspects of data harmonization, about everything from classifying income and identifying relationships within households to applying weights and dealing with missing data, to name but a few. We decided that this was the right time to systematically incorporate a number of these lessons.
2. When will these Wave V.2 datasets be made available?
LIS will make our new Wave V.2 datasets available for our users in three stages:
On 28 February, we put online: Australia-1995, Australia-2001, Denmark-1995, Denmark-2000, France-2000, Germany-2000, Mexico-1998, Mexico-2000, Mexico-2002, UK-1999, and US-2000.
On 31 March, we will put online: Austria-2000, Finland-2000, Greece-2000, Ireland-2000, Norway-2000, Poland-1999, Russia-2000, Slovenia-1999, Spain-2000, Switzerland-1998, and Switzerland-2000.
And on 30 April, we will put online: Belgium-2000, Canada-1998, Canada-2000, Estonia-2000, Hungary-1999, Israel-2001, Italy-1998, Italy-2000, Luxembourg-2000, Netherlands-1999, Sweden-2000, and Taiwan-2000.
3. How can you access the Wave V.2 datasets?
To use the Wave V.2 datasets, simply call them up in your programs using an alias built upon the country code, the reference year and a suffix defining the file type (h for household-level and p for person-level). For example, in your program, you will specificy "FR00h", "DK95p", and so on. The job syntax to access datasets as well as the specificities linked to the SAS, Stata, or SPSS programs are detailed at http://www.lisproject.org/dataccess/jobsub.htm .
4. What are the main substantive changes in Wave V.2, compared to the Wave 5.1 datasets?
The main changes are as follows:
□ Most substantially, we added 16 new person‑level labour market variables and substantially updated and revised the existing 9 variables at both the person‑ and household levels. In conjunction with this effort, we introduced a new system of semi-standardization to make cross-country and over-time comparisons substantially more accurate.
□ We re-organized the old-age pension variables, to account for widespread changes in the structure of pensions, especially changes that affect the distinction between public and private pensions.
□ We refined and formalized rules for the placement of family and child benefits, also to account for changes in the structure on benefits available in the LIS participating countries.
□ We added more person-level income variables.
□ We re-organized the expenditure data block in order to correspond to the U.N. Classification of Individual Consumption According to Purpose (COICOP) classification system.
□ We revised the treatment of net and gross data; see the new LIS Summary Variables
□ We added a new variable that allows users to more efficiently identify persons with partners and also those who are parents.
□ We have updated country codes to reflect the 2-digit ISO Country Codes. See Country Identification Number List
□ We ceased the practice of switching heads and spouses in order to always make the household “head” male. That was an old-fashioned habit, adopted many years ago!
□ What did not change? Given that we have made changes to file structure, added new variables on both labour market characteristics and person-level income variables, and changed the classification of some other, we should also point out that certain things are unaffected by these changes. In particular, while the changes to the income definitions affect where certain income items are found in the LIS classification, they do not affect the sum total of all income sources, i.e., the level of disposable income (DPI) for the household.
5. How has the file structure changed?
There are three main changes to the file structure:
□ We eliminated the child files completely. All data on children in a household can now be found directly in the person-level files. Please, note that this change applies to the entire LIS database, including previous waves.
□ We eliminated the “LES integrated files” that existed in Wave 5.1. All labour market data can now be found directly in the person-level files. If you need to complete your ongoing research using those files, they can still be accessed using revised aliases. (Specific alias information is available in the instructions for accessing LIS/LES data”)
□ We introduced a new set of files, known as “shadow files”. These files contain information for interviewed households with zero or missing household weights. (This situation only occurs when the original data are from a panel study.) The addition of these files allows users to keep some or all of these records if they wish to do so. Prior to the creation of these shadow files, these records were not made available to users. If you wish to use the shadow files, you must call them in your programs using an alias built upon the country code, the reference year, and either the suffix “hs” or “ps” depending on the dataset-level you wish to use. For example, you will specify in your programs "DE00hs", and so on. To fully understand the construction and use of the shadow files, we strongly recommend that you also study the LIS policy on missing data.
6. Is there new documentation?
□ On our website, you will find a new, updated LIS Variables list in the “LIS quick reference guide”
□ You will also find a redesigned and updated LIS Variable Definition list incorporating all ever‑existing variables.
□ A document describing the labour market variables and the new system of semi‑standardization has been introduced.
□ You will also find an updated explanation on how to access LIS/LES data that includes the changes related to accessing the different file types in SAS, SPSS, and Stata (http://www.lisproject.org/dataccess/jobsub.htm).
□ The LIS policy on missing data has been updated.
We strongly recommend that you study these five documents carefully before you begin to use the Wave V.2 data.
7. Which LIS datasets will be “lissified” in the new template?
The new template will be used in the following circumstances:
□ We will create a Wave V.2 version of every dataset contained in the original Wave 5.
□ We will use the new template for all of Wave 6. That will improve comparability over time with future waves.
□ We will use the new template whenever we receive, and lissify, a new dataset that corresponds to a period of time prior to Wave 5. In other words, sometimes we receive datasets for earlier years and, when we do, we will lissify them using this new template, regardless of the wave to which they belong.
□ We will use the new template on those occasions when we “re-open” an old (pre-Wave 5) dataset for the purpose of making major revisions.
□ We will re-release Wave 4 in the new template. We will not consider doing this until after Wave 6 is completed.
We encourage you to shift to the Wave V.2 datasets, but you are free to continue to use the Wave 5.1 data for as long as necessary to complete your ongoing research.
To access the Wave 5.1 datasets, you should call them in your SAS, SPSS or Stata programs using an alias built upon the country code, the reference year, and either the suffix “hr” or “pr”, depending on the dataset you wish to use. For example, "MX98hr" will give you access to the Mexico‑1998 household file for Wave 5.1.
The documentation for these datasets can be found at (www.lisproject.org/techdoc/release1.htm)
9. What should you do if you need help using the Wave V.2 datasets?
Contact LIS user support at usersupport@lisproject.org. Our primary staff person in charge of user support, Emilia Niskanen, will answer your questions and/or she will direct them to another LIS staff person, as appropriate.
10. What should you do if you have feedback for our staff?
Please let us know if you have feedback and/or comments on the changes that we have made. We welcome any and all questions, comments, and constructive complaints. Your input is invaluable to us.
-- The LIS Team.