Page 178 - Special Topic Session (STS) - Volume 1

STS423 Ylva A.R. et al.
3. Results and Discussion
3.1 Alternative data sources
Several of the variables (Table 1) previously collected directly from
farmers in the surveys on Use of fertilisers and animal manure and
Cultivation measures in agriculture, via telephone interviews only, or
mixed- mode, could instead be extracted from a combination of various
administrative registers – including the LPIS – and from farm management
software (Redner & Andrist Rangel, 2016). The review of the data
collection process also showed that only a limited number of target
characteristics could be excluded from future surveys, due to continued
user needs.

Table 1. Results of the review of target characteristics and variables in the surveys Use of
fertilisers and animal manure and Cultivation measures in agriculture.
Variables available from LPIS and from Variables available from farm
the register of crop areas with management
subsidies for organic farming
Age of temporary grasses and grazings Field identification and size
Age of fallow (set aside) land Crop
Preceding crop (crop rotation) Type of fertiliser applied (mineral/manure/
other organic fertilisers/soil amendments)
Organically cultivated fields (partly Date of application
derived data)
Application rate (kg or tonnes per ha)
Spreading technique for manures

Concerning the use of the Swedish Block Database (LPIS), the final
method that was developed (Andrist Rangel et al., 2016) was implemented
in the survey on Cultivation measures in agriculture 2016 (Statistics
Sweden, 2017), where it was used to collect data on age of fallow.
However, it was only applied to the objects (farms) included in the sample,
and thus the full potential of the register was not utilised in the 2016
survey. Nonetheless, the farmers’ response burden was reduced, which
was the original goal of the new approach. The reason why the whole
register was not used was because other data related to fallow, which had
been collected via the questionnaire, was to be cross- utilised with the
data on age. However, in future use, or in use for other purposes, it would
be preferable to use the whole register. The Swedish Block Database has
very good coverage, as it includes information on more than 99 percent
of the agricultural area (Swedish Board of Agriculture, 2018); this means
that this approach would eliminate the sampling error present in the
current design. The next step will be to investigate the methodological
aspects of a mixed design, i.e., a combination of a sampling design and a
complete enumeration (via registers) in a single survey.

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