Text and photo: Kaisa Järvelä

THL's Hypa Data Reveal How Finland Has Fared in the 2000s

The Finnish Social Science Data Archive (FSD) has awarded its third Data Management Award in the Aila Seminar. This year, the award was granted to Pasi Moisio who has overseen the archiving of all the Welfare and Services in Finland (Hypa) data at the FSD.

One era of Finnish research on living conditions is now over: next year, the National Institute for Health and Welfare (THL) will no longer collect the Hypa data that have charted the welfare of the Finnish population and its use of social and health services.

The diverse, carefully executed living conditions survey was collected four times during the 2000s. It has provided us with unique data on the welfare of Finnish people. For instance, decision-makers and researchers have acquired an understanding of the share of the population that has had to seek help from bread queues.

Fortunately, Hypa's legacy continues. The survey questions relating to wellbeing will be incorporated into THL's Health and Well-being for Residents (ATH) study in the future.

'THL has had around ten regular surveys which have now been combined whenever possible', Research Professor Pasi Moisio explains.

Questions concerning social and health services will get their very own survey in the future, thanks to the need for information caused by the social and healthcare reform. A pilot survey has already been created, and the first report on it has been published.

Model data for open science at THL

Hypa's uniqueness derives from the fact that THL decided to archive the data at the FSD as early as at pre-collection stage, long before the current trend towards open access to science.

'Matti Heikkilä, our Deputy Director General at the time, had the firm opinion that the data would be made available for other researchers', Pasi Moisio recalls.

Unfortunately, it soon became clear that the progressive idea would not be easily implemented. Archiving the data seemed to fall through because Hypa's background information had been collected from registers. At the beginning of the 2000s, the Statistics Act was interpreted to prohibit handing over interview data combined with register data.

'Without background variables such as gender, place of residence and income level, the data would have been useless to researchers. For this reason, the entire archiving plan had to be put on hold', Moisio explains.

The first three Hypa datasets ended up locked in Moisio's desk drawer, with Moisio himself doing his best to grant access permits to researchers eager to study the data.

'It did not take long to realise that the arrangement made no sense. Access permits must be granted in a centralised manner, for both judicial and technical reasons.'

The Act came to be interpreted in a different way eventually, and the Hypa data was moved from Moisio's drawer to the data archive. The user statistics of the Aila Data Service reveal that interested users soon found the data: 12 different researchers have downloaded Hypa datasets from Aila for a total of 39 times since May 2014.

Unique data on bread queues and informal care

The Hypa surveys have provided Finland with data on several issues that are central to wellbeing but can easily remain hidden in statistics – for instance, the aforementioned bread queues.

One of the things that questions concerning food aid and running out of money for food in the 2013 Hypa survey revealed was that the percentage of Finnish people who had had to seek help from bread queues within the past year was 2.3%. In addition to this, over one tenth of the respondents reported having been in a situation in which they had run out of money for food.

kuvituskuva

Data Management Award Winner, THL's Research Professor Pasi Moisio also happened to be the first user of FSD's services when operations began in 1999.

'These figures provided a much needed framework for public discussion, making it possible to estimate how large and frequent a phenomenon bread queues are in Finland', Moisio states with satisfaction.

Thanks to the Hypa surveys, we have also gained an understanding of the number of carers in Finland. Two of the surveys included questions charting the number of Finnish people that care for a person close to them with some regularity without the support of informal care allowance.

'One of the main findings was that the number of people caring for a person close to them on a daily basis was nearly twice as large as the number of Finnish people who receive informal care allowance', Moisio discloses.

Since then, the figures have been used frequently in contexts such as legal discussions on reforming informal care allowance.

Besides stimulating discussion, the Hypa data are also valuable because the same questions have been repeated over the waves. This has enabled research on how the wellbeing and attitudes of Finnish people have changed over time.

Moisio notes that time series are particularly important in social sciences and in surveys that concern wellbeing, as the responses and their interpretations are subjective.

'The only absolute measure consists of the answers previously given to the same question.'

Random sampling is the prerequisite of statistical analysis

Moisio also wants to emphasise the high quality of the Hypa survey.

'We had better make some noise about why the findings of extensive surveys based on random sampling are more trustworthy than data collected with online surveys or the like.'

Moisio has observed that surveys based on non-probability sampling have crept alarmingly close to real research in Finland. Even projects conducted by the ministries have featured such surveys.

'The temptation to take a shortcut exists because representative random sampling is expensive and difficult to execute. Nonetheless, inference based on statistical analysis is only possible when answers have been collected by random sampling', Moisio points out.

The Hypa surveys also provide researchers with various special treats, such as a 50 per cent rotating panel. This means that half of the survey respondents have participated in its previous collection wave.

But what additional value do the archiving of and open access to the data offer? Does the reform spirit of the turn of the millennium seem to have been worth it in retrospect?

Surprisingly, Moisio seems slightly bewildered by the question.

'Open access to data is so self-evident due to Finland's current science policy objectives that even considering the issue feels strange', he explains.

Finding a reason for why archiving is self-evidently rational proves to be easier.

All the Hypa data have been archived at the Finnish Social Science Data Archive (FSD) and can be downloaded for research purposes with the permission of the original research team.

HYPA data

The Welfare and Services in Finland data (HYPA) are a survey series collected by the National Institute for Health and Welfare (THL, previously STAKES) that extensively charts Finnish people's use of social and health services.

The data have been collected four times: in 2004, 2006, 2009 and 2013.

Each wave included around 4,500 Finnish respondents. Half of the interviewees in each survey are new, whereas half of them were also interviewed in the previous wave.

The basic data have been collected through telephone interviews. Additionally, individuals aged over 80 years were interviewed face-to-face and separate postal surveys were sent to families with children in 2006 and 2012.

Response rates in Hypa surveys are notably high, over 70 per cent.

Data Management Award

The Finnish Social Science Data Archive (FSD) grants the Data Management Award to either a person, a research team or an organisation as an acknowledgement of excellent data management and attention given to data lifecycle when collecting, organising and archiving data.

The Data Management Award was established in 2014.

It has previously been awarded to Doctor of Arts Michail Galanakis and Doctor of Social Sciences Janne Kivivuori.

Creative Commons -license