Page 177 - Special Topic Session (STS) - Volume 1

STS423 Ylva A.R. et al.
management software on the Swedish market. In the beginning, both
companies were interested in the proposal, and discussions on the
data transfer procedure were initiated. Several possibilities were
discussed, but can be summarised by two main options: 1) a transfer
function in the farm management software, where the farmer actively
initiates the process of exporting data of a set of known variables into
Statistics Sweden’s new web system for agricultural statistics, or
2) the software providers allow Statistics Sweden access to their
databases with all data from their clients (subject to the client’s
consent). There was consensus for the first option for a variety of
reasons. Firstly, the farm management software clients (the farmers)
were most likely not willing to give consent for unconditional
secondary use of all their data (as in option 2). Secondly, option 2
would involve data quality issues, as farmers use the software for
different purposes: from purely planning, a certain degree of follow-
up, to a complete accounting of last year’s result in terms of physical
input and output, and dates for different measures. Hence, in option
2, neither the state nor the quality of the data would be known to
Statistics Sweden. At this stage of the discussions, one of the software
providers informed that they wanted to postpone possible
cooperation. Hence the next step, which was designing an interface
for exporting data in the farm management software, was only
implemented by one of the software providers. At the same time, the
tool for importing data was designed and incorporated into the new
web system for agricultural statistics at Statistics Sweden. XML was
settled on as the format for the data transfer.

2.3 Coordinating surveys
Since 2006, data collection for either the Use of fertilisers and animal
manure survey or the Cultivation measures in agriculture survey had been
carried out every year. The review included examining whether these two
surveys could be merged into one. Issues that had to be considered were
sampling frame, sampling method, number of strata, length of
new/merged questionnaire, time series constraints, and response burden.
In addition, the new web system for agricultural statistics was designed for
coordinated administration and data collection of all ongoing surveys in
the system, independently of agricultural domain, i.e., crop production or
agri-environmental statistics (pesticides, fertilisers, manure and cultivation
measures).

166 | I S I W S C 2 0 1 9