Page 331 - Invited Paper Session (IPS) - Volume 1
P. 331
IPS155 Laura B.
were made available to the researchers (usually universities) that requested
the dataset with a letter via ordinary mail. The dataset was sent in magnetic
tape archives. Since the late Nineties, with the Banca d’Italia’s web site,
households’ survey dataset have been made freely available on the web site,
and lately also earlier datasets have been included: datasets are now out for
every single survey in full (also including data found in the monographic
sections of the individual surveys) from the 1989 wave on.
For households’ survey data GDPR (as earlier the Italian law) states that
individual data can be disseminated after data minimization. We therefore
anonymize households’ information and don’t include identification variables
(birth date and place, residency, day and hour of interview, etc.) in the datasets
freely available in dedicated Public Use Files (PUF) on the Banca d’Italia web
site. PUF is a common choice for data for which anonymization and the
exclusion of some (indirect) identification variable are enough for
safeguarding confidentiality.
No dataset request form is needed in order to download the PUF file. Many
other institution (as for example the ECB and Banco di Espaňa) ask households’
survey data users to fill in a request form online and data are then sent (or
download is unblocked) after the request is validated by the data provider. We
have deliberated for no request form and the PUF file is free for download
without the identification of the user, also for the consideration that PUF can
easily be transmitted between researchers. Maybe also thanks to this easy
access, households’ survey data are widely used, as witnessed by the
numerous research papers that have been written over the years, and their use
by some universities during exercising classes. Banca d’Italia just makes it clear
to users that data are distributed for research purposes only and their use is
under full and exclusive responsibility of the user. The drawback of this wide
dissemination policy is that we can’t track the users of households’ survey data,
as researchers usually fail to send us back their paper. We are therefore
thinking about adding an online compiling form on the web page, in order to
collect some basic information (the name of the researcher, institution or
university affiliation, title of the research project, etc.) before unblocking the
download of the PUF.
In the light of providing researchers with as much as useful information as
possible, aside the archives for every single survey we deliver additional
datasets with specific data adjustments. The first is an historical database
containing data from 1977 but only including the subsets of variables
considered useful for longitudinal analyses, after having performed a number
of adjustments: as the way questions are posed (and therefore variables
and/or variables options) may differ over time, variables are ex-post
harmonized, so that time-series are continuous and longitudinal analyses can
be easily performed by the researchers, without obliging them to prepare the
320 | I S I W S C 2 0 1 9