Although scientists often make controlled laboratory conditions conform to
the demands of particular statistical analyses, it is also possible to fit multivariate
and time series analyses to complex field conditions. Oceanographers,
meteorologists, geologists, and field ecologists are all accustomed to working
with "experiments" designed by "nature". This differs from survey research in
that researchers may gather detailed data on complex systems over time,
within an experimental framework that embraces variability. Agroforestry
researchers can monitor and test relationships between specific practices,
processes, species, site characteristics, farm types and land user groups. They
can sample and monitor agroforestry systems or trees-in the-landscape
according to maps of existing variation in practice or in vegetation. Such
"inferred" or "insinuated experiments" can explain the economic and ecological
significance of these patterns in the landscape. Researchers may also use
monitoring data to develop models that can simulate potential changes in
land use systems. For example, agroforestry researchers could use farmers'
records and their own field observations to model an existing land use system
and to predict the likely outcome of a prolonged drought or a new settlement
program.