Overview
The
Tree Nursery Model is an example of how a Bayesian Network can be used to assess the impact of policy interventions on improving the performance of an industry. It this case the industry is the tree nursery industry in the Philippines. The model was developed using interview data from over 100 tree nursery managers in Leyte province, the Philippines, covering private, communal and government nurseries. The interviews were structured to elicit information on nursery cultural practices, nursery operation scheme, technical skills of nursery managers, nursery set-up, volume of seedling production, nursery inputs and facilities, germplasm and seedling quality, and constraints encountered in nursery operations.
Figure 1 shows the model developed for the private tree nursery sector. It has been colour coded to highlight its main parts:
(a) germplasm quality (shown in
green),
(b) seedling demand (shown in
dark blue),
(c) seedling quality (shown in
yellow), and
(d) outcomes, which are
tree nursery sustainability and
effectiveness in this case (shown in
red). At the top of the model (shown in
light blue) are a set of
policy interventions that are linked to one or more parts of the model.
Figure 1: Model for private tree nursery sector
Using the Model
Open the Tree Nursery Model by clicking
here. When the model is first opened it displays the prior probabilities of nodes being in their respective states. If you look at the
production budget node, for example, it displays a
10.8% probability of being
satisfactory. This means, that of the tree nurseries interviewed, only
10.8% had a satisfactory budget. Similarly, only
33.2% of nurseries interviewed were deemed to have
high effectiveness (see the
individual nursery effectiveness node).
Now lets look at how the model can be use to assess the potential impact of policy interventions. Say we want to assess the potential impact of improving the
production budget of nurseries. This can be done by selecting the
satisfactory state for
production budget, as shown in
Figure 2 (indicated by the large blue arrow). The large red arrows in
Figure 2 indicate the impact of this policy intervention. For instance, the model now shows that for those nurseries that have a
satisfactory production budget, the percentage of tree nurseries with
high germplasm availability increases, leading to an
increase in the percentage of tree nurseries that are
sustainable and have
high effectiveness.
Figure 2: Tree nursery model used to assess the potential impact of improving the production budget (indicated by blue arrow)
Combinations of policy interventions can also be assessed using the model using scenarios. For example,
Figure 3 shows the expected improvement in the percentage of tree nurseries having
high effectiveness when
satisfactory for production budget,
yes for importance of genetic quality and
expert for technical skills are selected simultaneously.
Figure 3: Tree nursery model used to assess the potential impact of multiple policy interventions (indicted by blue arrows)
By cumulatively turning on policy interventions, a performance improvement graph can be plotted. For example,
Figure 4 shows the improvement in the percentage of high effectiveness nurseries as possible policy interventions are cumulatively implemented for the both the communal and private (individual) nursery sectors.
Figure 4: Probability of high nursery effectiveness for the communal and individual nursery sectors as interventions are cumulatively implemented
