Pros and Cons of Register Research in Finland
Helena Laaksonen
Finnish registers are accurate and reliable, says Senior Researcher Ulla Hämäläinen from the Research Department of the Social Insurance Institution of Finland (KELA). Registers give more reliable and extensive information than surveys on, for example, income and unemployment. However, surveys are also needed, especially for research on expectations and attitudes.
Ulla Hämäläinen and Senior Researcher Pentti Takala are currently involved in a research project studying the use of family leave in Finland. The project seeks to determine whether families behave in the way decision-makers thought families would behave when making decisions on financial incentives. The project combines data derived from the registers of KELA and Statistics Finland. KELA registers give information on fathers who have used family leave whereas employment statistics of Statistics Finland on those who have not.
Keeping track of changes
For longitudinal register research it is essential to keep tract of law reforms and their impact. This is routine for Hämäläinen. The time period under study in their research project starts at 1995 since taxation systems were changed so radically that year that data from previous years are not comparable to the post-1995 data.
- A researcher must be aware of changes in laws and how they will affect, say, the number of benefit claimants.
Another problem comes up when registers allow comparison over decades. The metadata, that is, information on how the data have been generated, may be insufficient.
- Sometimes one must ask oneself whether registers from different decades can be used to study a phenomenon even if they carry the same name. It is worthwhile to keep in mind this unwritten rule: if you get a surprising result, question it, Hämäläinen reminds.
Reliable but expensive
- Finnish registers are reliable and the data are mostly very accurate. Moreover, registers are extensive, and data connected to the same phenomena can be derived from various sources, Ulla Hämäläinen says.
Registers usually contain information displaying what the situation was 2-3 years ago. Although some KELA registers can be used to follow a phenomenon in real time, it is only after data from several different registers are available that the underlying causes and factors affecting the phenomenon can be studied.
- Our most recent registers are from the year 2004, she specifies.
Data protection issues
Hämäläinen regards data protection as a central issue in register research. Researchers are usually given access to part of the data, and cannot therefore identify individuals.
- Statistics Finland is particularly strict in these matter - as it should be, admits Hämäläinen.
Strict requirements can sometimes frustrate researchers. Due to data protection reasons, it is difficult to get access to regional data even though researchers are never interested in individuals. Recently Statistics Finland has sought new solutions to the problem by using encryption systems. This way researchers can access data they want without endangering confidentiality.
Astonishing range of phenomena
Hämäläinen thinks that the main advantage of registers is the great variety of phenomena that they allow to be studied. Registers help us understand what is real and what is not. Research is often portrayed as a straightforward process from a hypothesis through testing to results. Hämäläinen encourages researchers to prepare themselves for surprises:
- Sometimes it is fun to notice having been right and equally often it is fun to notice having got it all wrong.
|