Investigation of local standpoints on forest forest fire caused by

human depletion justifiable timberland management: an

Indonesian case study *
Article · January 2004

CITATIONS READS

59 126

3 authors:

Herry Purnomo Gil Mendoza

CIFOR and IPB University of Illinois, Urbana-Champaign
124 PUBLICATIONS 2,216 CITATIONS 18 PUBLICATIONS 2,126 CITATIONS

SEE PROFILE SEE PROFILE

Ravi Prabhu
Consultative Group on International Agricultural Research
93 PUBLICATIONS 3,948 CITATIONS

SEE PROFILE

All content following this page was uploaded by Gil Mendoza on 29 January 2022.

The user has requested enhancement of the downloaded file.

 Journal of Environmental Management 74 (2005) 111–126
www.elsevier.com/locate/jenvman

Investigation of local standpoints on forest fire caused by human

depletion justifiable timberland management: an Indonesian case
study *
Herry Purnomoa,c, Guillermo A. Mendozab,*, Ravi Prabhuc
a
Faculty of Forestry, Bogor Agricultural University, Bogor, Indonesia
b
Department of Natural Resources and Environmental Sciences, 1102 Sth Goodwin Av., W-503 Turner H, University of Illinois, Urbana, IL 61801, USA
c
Center for International Forestry Research, Bogor, Indonesia

Received 4 September 2003; revised 4 July 2004; accepted 26 August 2004

Abstract
Despite widespread acceptance of sustainability as the ultimate goal of forest management, perspectives about its meaning, significance,
and relevant indicators may still differ. This paper examines local perspectives on sustainability, and evaluates their similarities and
differences. A systematic procedure based on criteria of proximity, pre-existing rights, dependency, knowledge of forest management,
forestry spirit, daily activity, and legal rights was used to identify a small group of relevant stakeholders representing different groups,
institutions, and organizations. Using participatory action research (PAR), stakeholders were asked to identify relevant indicators of
sustainable forest management. The indicators identified by each stakeholder were then compared to a consolidated list assembled by field
facilitators with respect to whether relevant indicators are present or not. Based on the resulting presence/absence matrix, a statistical tool
called the simple matching coefficient was used to estimate the similarity measures among the stakeholders’ perspectives. In addition, cluster
analysis was used to classify groups of stakeholders depending on their similarities to each other. Finally, hypotheses relat ed to the
‘closeness’ of perspectives among local communities, non-governmental organizations, a timber company, and government organizations, as
revealed by their selection of indicators, were tested. Results show that: (a) local communities have different perceptions in terms of what
they consider to be important indicators compared to the NGOs, (b) there are significantly different perceptions between the government and
the timber companies, and (c) there are also different perceptions between urban and field-based personnel of the same organization.
q 2004 Elsevier Ltd. All rights reserved.

Keywords: Sustainability; Criteria and indicators; Participatory action research; Cluster analysis

1. Introduction sustainable forest management. C and I are tools which can

be used to collect and organise information in a manner that
Over the last decade, sustainability of forest resources is useful in conceptualising, evaluating, communicating and
and ecosystems has become a worldwide concern. Conse- implementing sustainable forest management (Mendoza
quently, many national and international initiatives have and Prabhu, 2000a,b; Ducey and Larson, 1999; Prabhu
been launched to promote sustainable forest management. et al., 1996). The term ‘sustainable forest management’ has
Among these, the most comprehensive and far reaching has become a widely used term in forest policy, and has also
been the development of criteria and indicators (C and I) for served as a primary guiding principle in community-based
forest management (Ferguson, 1996; Maser, 1994). Despite
*
This study was funded through a collaborative project between the different interpretations about its meaning and the lack of a
University of Illinois and the Center for International Forestry Research universally accepted definition, sustainability has been
(CIFOR). The ideas contained herein are solely of the authors and do not viewed as an ideal condition to strive toward.
necessarily reflect official views of CIFOR.
Along with the concept of sustainability, participatory
* Corresponding author. Tel.: C1 217 333 9347; fax: C1 217 244 3219.
E-mail address: [email protected] (G.A. Mendoza). management has also become a widely accepted

0301-4797/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2004.08.013
112 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

management philosophy, particularly for community-man- representation. To pursue this analysis of local perspectives,
aged resources. Experience has shown that successful a case study was initiated in 1999 involving a community
community-based forest management entails effective forest in Indonesia. Further visits and subsequent inter-
collaborative decision-making, which in turn depends on actions with stakeholders were conducted in 2000 and 2001
understanding and communication (Mendoza and Prabhu, to update information on their perspectives on forest
2001; Varma et al., 2000; Purnomo et al., 2003). sustainability.
Unfortunately, there are many impediments to effective
communication. A major, yet often overlooked, impediment
2.1. Defining stakeholders
is the fact that forest stakeholders may interpret events and
situations quite differently from one another. This diversity
One of the most important aspects of social research is
of perspectives, coupled with the tendency for people to
believe that their own perspectives are the most legitimate, the identification of the relevant group of stakeholders or
can result in divergent problem definitions, misunderstand- participants who will be actively involved in providing
ings, and the eventual breakdown of the decision-making input. This is particularly true in most forest management
process. situations because of the typically large number of interest
Previous studies have shown that it is vitally important to groups, users, organizations, and other institutions
explore how different stakeholders understand or conceptu- involved. In this study, identification of relevant stake-
alise appropriate forest management. For instance, Kearney holders was done using the ‘Who Counts?’ method
et al. (1999) have reported that differences in stakeholders’ developed by Colfer et al. (1999). The method seeks to
conceptualisations or perspectives on appropriate forest identify the most important stakeholders connected to the
management may have contributed to the controversy in the forest based on dimensions of proximity, pre-existing
Pacific Northwest forests of the United States. Results from rights, dependency, knowledge of forest management
their study indicated the existence of a wide range of (indigenous knowledge), forestry spirit (e.g. culture),
concerns among stakeholders including issues related to the daily activity on site (intensity of activity), and legal
process of forest management. Pokorny et al. (2004) also rights. ‘Scores’ are assigned by the researchers or
examined local stakeholders’ participation in the develop- facilitators to each group of stakeholders with respect to
ment of criteria and indicators of sustainable forest the seven dimensions. The simple scoring system is based
management in Brazil. One of their findings was that on a scale that ranges between one and five (1Zhigh, 2Z
differences in the evaluation of indicators stem largely from relatively high, 3Zmedium, 4Zrelatively low, and 5Z
the different stakeholders’ understanding or perspectives low). The scores for each stakeholder are calculated and
about the ‘verifiers’ of each indicator. serve as the basis for deciding whether a stakeholder is
Using statistical and other analytical methods, this paper included as a participant. What the method provides is a
examines the perspectives of local communities and other simple means for ranking stakeholders according to their
stakeholders on sustainable forest management. Based on importance for the forest, or their dependence on it. This
these methods, the paper makes inferences about their ensures that important stakeholders are not left out when
similarity or divergence. Such analyses can be useful in only a sub-set can be included in the study.
exploring common areas of interests and perspectives The cut off point for defining ‘who counts’ is
among different stakeholders. determined using the means of scores on the seven
dimensions and the experience and resources of the users
of the method, with feedback from those whose
2. Methodology and design of study importance is being assessed. Thus, the cut-off point can
vary depending on the context. Like any other method, the
The study presented in this paper focuses on analyzing procedure has some weaknesses, particularly its apparent
local perspectives about forest sustainability using criteria subjectivity. It also has some strengths and desirable
and indicators. Essentially, stakeholders’ views were features that suit the stakeholder analysis required in the
elicited on what indicators they consider to be the most study. For instance, it offers some flexibility in terms of
important for sustainable forest management and this was the set of criteria for stakeholder inclusion that can be
used as a proxy for their perspectives on sustainability. used. It accommodates and makes use of prior knowledge
Determining people’s perspectives or mental models of about the forest and the stakeholders, and it is amenable
sustainability is important because they drive complicated to a participatory process in the identification of
multi-party processes of decision-making related to natural stakeholders. Moreover, while the scoring is subjective,
resources. This situation is often exacerbated by the fact that the process itself is objective and, more importantly, it is
some groups claim to be acting on behalf of others, or at transparent to all stakeholders. Furthermore, each stake-
least in their best interests. Development practitioners, for holder or interest group has a voice in the identification of
instance, often find before them a tangled web of competing the final list of stakeholders; that is, who counts (key
interests, conflicting perceptions and claims for stakeholders) in forest management.
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126 113

Table 1 Table 2
Presence/absence binary representation of stakeholders’ perceptions of 2X2 matrix of frequencies of indicator occurrence for stakeholders A and B
sustainability based on Table 1

2.2. Determination and analysis of stakeholders’ Cell (a) represents the case where both indicators are present (i.e. both have
values 1), cell (d) represents cases where both indicators are absent (i.e.
perceptions
both have values 0), and cell (b) and cell (c) represent cases where one
indicator is present but absent on the other.
In seeking to determine stakeholders’ perspectives on
sustainable forest management, this study focused on
identifying factors or aspects of forest management that indicators 1, 2, 9 and 10 in Table 1, where Stakeholder A
the stakeholders considered were key in achieving scores those indicators as ‘1’ and Stakeholder B scores
sustainable forest management. These factors were as ‘0’.
elicited through carefully designed semi-structured inter-
views within the framework of participatory action 2.2.1. Cluster analysis of perceptions
research or PAR (Selener, 1997). Other common tools One of the objectives of this study is to examine
for visualization of thought processes in workshop whether there is any commonality of perceptions among
situations, such as flip-chart papers, pens and sticky tape local communities about what they consider to be key
were provided to enable the participants to express what indicators of sustainable forest management. One method
they considered important in achieving good forest that can be used to analyze this type of qualitative
management. Each individual stakeholder independently information is cluster analysis, which is briefly described
generated a list of indicators during these meetings at their below. Previous applications of cluster analysis in forestry
sites. All indicators were then presented, debated, and include: analysis of spatial patterns of tree species
discussed among the group of stakeholders. (Plotkin et al., 2002), forest inventory and remote sensing
Based on these stakeholder perceptions, a consolidated (Wulder, 2002).
indicator set was generated and served as basis for As the name implies, cluster analysis is an approach
comparison. Thus, if an indicator (factor) in this consoli- designed to uncover or discover groups or clusters of
dated set could be matched with a local perception, it was relatively homogeneous data. In the context of the study,
scored as a ‘1’. If on the other hand no match was possible to cluster analysis was used to identify groups of perceptions
any perception, it was scored a ‘0’ as illustrated in Table 1. that exhibit sufficient commonality or similarity. In apply-
This process resulted in the development of a matrix of ing cluster analysis, there are two important issues that are
presence/absence data. particularly relevant to the study. First is the measure of
From the presence/absence matrix for indicators related similarity or proximity. This proximity measure serves as
to all stakeholders of the type shown in Table 1 we the basis for identifying groups or clusters, and also in
constructed a 2X2 table for any two stakeholders as determining how close or similar the groups of data are. The
shown in Table 2. Here, for example, is the 2X2 table of second issue is the manner in which various groups of
stakeholder 1 and 2. In Table 2 the cell values (2, 4, 2 and stakeholders are identified or organized. Related to this
2) were obtained by searching the cells in Table 1 for issue are: number of clusters, and how they are formed and
indicators related to stakeholders A and B and scoring grouped together.
them as frequencies of occurrence where an indicator is There are a number of similarity coefficients that can be
present for both A and B (the pair (1,1)) or present for A used in cluster analysis. However, for binary data
but not B (1,0), or vice versa (0,1) or absent altogether (presence/absence or 0/1) such as those contained in
(0,0). Thus, in Table 2 the frequency of 4 in the first cell Table 1, the two most commonly used coefficients are the
of the second column (this is the cell that shows indicators simple matching coefficient and the Jaccard’s coefficient
present for A, but absent for B) is obtained from the (Everitt, 1993). In this study, we used the simple matching
114 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

Fig. 1. Example of a hierarchical structure*.

coefficient. This coefficient was used in part because co- the similarity measure is calculated reflecting the
absences (i.e. both have values of 0) are considered similarity of the pairs or groups of individuals fused
informative just as co-presence (i.e. both have values of together at that stage. An example of a dendrogram
1). This coefficient calculates the ratio of the number of showing varying similarity coefficients of the sample
matches to the total number of matches and non-matches as problem described in Fig. 1 is shown in Fig. 2.
follows: Different agglomerative methods make use of different
fusion procedures. The most common of these are: the
S Z (a C d)/(a C b C c C d) (1)
single linkage, the complete linkage, and the group
S is the similarity coefficient, a represents the case average linkage methods. The differences between the
where both items are present (i.e. both have values of 1), methods arise because of their differences in defining
d represents cases where both items are absent (i.e. both similarity between an individual and a group of
have values of 0), and b and c represent cases where one individuals, or between groups of individuals. At each
item is present but absent on the other (i.e. one is zero and stage in the fusion process, the methods combine or fuse
the other is one, and vice versa) as shown in Table 2. Thus in individuals or groups of individuals, which are most
the example shown in Tables 1 and 2 the matching similar. The fusion process for each method is graphically
coefficient of stakeholder 1 and 2 is equal to 0.4. shown in Fig. 3. In simple linkage, the individual (or
group that is ‘closest’) is fused first. This is opposite to
2.2.2. Hierarchical clustering of perceptions the complete linkage, which fuses the individual (or
Given the similarity measures, the next issue in group), which is furthest or most distant. In the average
clustering is how to organize the data into meaningful
structures, groups, taxonomies, or classifications. One
common way to aggregate the data is to organize them in
a hierarchy of clusters where the process of classification
consists of a series of partitions running from one cluster
containing all individuals, to n clusters (n is number of
individuals to be grouped) each with one individual
(Everitt, 2001). Hierarchical clustering techniques can
either follow the ‘agglomerative’ method, which proceeds
in a series of fusions into groups, or ‘divisive’ methods,
which proceeds by separating n individuals into groups.
This hierarchical clustering process can be illustrated by a
two-dimensional diagram known as a ‘dendrogram’
because it resembles a tree-like structure as shown in
Fig. 1. The fusion process is organized in stages where in
every stage, one individual (pair or group) is fused to
another individual (pair or group). In each fusion stage, Fig. 2. A sample dendrogram based on Fig. 1 data.
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126 115

a reference or consolidated list and the stakeholders’ list.

For this test, the Pearson Chi-square, which is based on
the Phi correlation coefficient for dichotomous variables,
is used. This test statistic is based on the Phi coefficient
of correlation described by Everitt (2001).

3. Results, analysis, and discussion

3.1. The site and identification of stakeholders

The case study used is a Forest Management Unit

Fig. 3. Examples of three inter-cluster linkage measures. (Taken from
(FMU), which is currently managed by the state owned
Everitt, 2001. Cluster analysis).
company called PT. Inhutani II located in East Kalimantan,
Indonesia. The area covers 14,180 ha of what is classified as
limited production forest and 34,120 ha of production
linkage, also known as the un-weighted pair-group forest. The latter contains 23,890 ha of virgin forest,
method, the distance between clusters is the ‘average of
7280 ha of logged over forest, and 2920 ha of shrubs and
the distance between all pairs of individuals that are made degraded forests.
up of one individual from each group’. In this study, the
According to the government’s plan PT. Inhutani II
group average linkage was adopted to determine the can log the forest at a rate of 1106 ha per year or
groups with relatively similar perceptions. This choice approximately 11 blocks of 100 ha each, for 20 years.
was made because the average distance better represents
The silvicultural system implemented is the Indonesian
the distance between two groups of perceptions compared Selective Cutting and Planting System (Tebang Pilih
to the closest or the furthest. The others do not describe Tanam Indonesia or TPTI). Because PT. Inhutani II has
the common perceptions of the stakeholders in the same no secondary processing industry in the vicinity the wood
category very well. is sold as logs or supplied to Inhutani II’s sawmill in
Stagen, South Kalimantan. Commercial species include:
2.2.3. Testing of hypotheses Shorea spp. (Meranti), Dryobalanops spp. (Kapur),
In addition to evaluating similarities of perspectives Dipterocarpus spp.(Keruing), Shorea laevis (Bangkirai),
among local stakeholders and identifying clusters of Palaquium spp. (Nyatoh), Gonistylus spp (Ramin) and
stakeholders with similar views, it is also instructive to Agathis spp. (Agathis).
determine how close the perspectives are between stake- Table 3 describes the list of stakeholders and their
holders. Three general hypotheses will be tested. First is the primary roles. From this initial list, it was decided to
similarity of perspectives between the local communities’ reduce the number of stakeholders by considering only
and local NGOs. This test is useful because of the close those who play direct roles in the study area. Based on
social relationships between these two groups. Moreover, the ‘Who Counts’ matrix (see Table 4) seventeen of the
NGOs often claim to speak on behalf of the communities on nineteen stakeholders identified scored less or equal to
many issues affecting the local communities. Second, it is ‘3’ in the study area. Fig. 4 illustrates how these
also informative to test how close the timber company’s stakeholders are connected and their relationship to the
perspectives are relative to the local government. This is FMU. The coal mining company was not considered
also insightful because theoretically, timber companies are because of inadequate data relating to coal deposits in
expected to follow regulations mandated by the govern- the forest or study area. This resulted in scores of ‘5’
ment. Their interpretation of these regulations would with respect to forest related activities. CIFOR as an
obviously be colored by their views on what constitutes international research organization was also excluded
sustainability. Hence, it will be useful to determine the because it is not directly involved in forest management
commonality of views and perspectives between the or utilization at the site.
government and the timber companies. Finally, it may
also be informative to test the similarity of views between 3.2. Stakeholders’ perceptions
the urban and field-based offices and personnel of the same
institution. To facilitate the process of eliciting stakeholders’
Recall that the data to be used for analyzing perceptions, we started by asking each stakeholder to
perceptions both in hierarchical grouping and the testing specify those indicators that they considered to be
of hypothesis about the similarity of perspectives important. To simplify the process further, the stakeholders
between groups of stakeholders are essentially binary were asked to group the indicators into social, ecology or
data denoting absence or presence of indicators between biophysical aspects, economics or livelihood, and policy
116 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

Table 3
The 19 stakeholders and their primary roles

Stakeholders Type Relevant primary roles

Inhutani II Malinau Sub Unit Timber company field-based office Harvest timber
Inhutani II East Kalimantan Unit Timber company provincial office Develop plans of timber harvesting
Paya Seturan community Local community Use forests for their livelihoods
Long Lake community Local community Use forests for their livelihoods
Metud community Local community Use forests for their livelihoods
Rian community Local community Use forests for their livelihoods
Loreh community Local community Use forests for their livelihoods
Langap community Local community Use forests for their livelihoods
KANWIL (Provincial office of Ministry of Central government Develop policies and standards
Forestry)
Provincial university lecturers Academicians Educate students
Provincial university students Students Study of forests and communities
Coal mining Non-forest company Mining coal
CIFOR International research institute Research
YAP Local NGO Empower local communities
BIOMA Local NGO Empower local communities
PLASMA Local NGO Empower local communities
Sub District Forest Service (RPH, Resor Local government Executing forest policies at sub district level
Pemangkuan Hutan)
Provincial Forest Service (Dinas Kehutanan I) Local government Executing forest policies at provincial level
Sub-District Office (Kecamatan) Local government Executing sub district policies

‘groups’. A team of facilitators who were also familiar with agglomerative hierarchical method combines individuals
the site later consolidated the list of indicators generated by or groups in a stage-wise progression depending on the
each stakeholder into a single list of common indicators. value of the proximity or similarity measure. In the first
Following this, facilitators compared or matched this step, the two observations closest together are fused. In
consolidated list with each stakeholder’s original lists. the next step, either a third observation joins the first two,
Indicators that were present (or absent) on both lists were or two other observations join together into a different
noted accordingly. cluster (Kaufmann and Rousseeuw, 1990). This process is
As can be expected, the stakeholders generated a long list illustrated in Fig. 1. In the simple example, five elements
of indicators. The consolidated list included 35 social (a, b, c, d, e) are analyzed using the agglomerative
indicators, 27 ecological indicators, 24 economic and hierarchical method. In the first stage, elements a and b
production indicators, and 7 policy indicators. To facilitate are clustered (a, b) with a similarity coefficient equal to 1.
the analysis of these perceptions, and to present the At the second stage, elements d and e are grouped (d, e)
information more concisely, the facilitating team further with a similarity coefficient of 0.9. At stage three, element
grouped the related factors into a consolidated list under c is grouped together with the cluster (d, e) forming a
each of the four groups of factors. This resulted in further cluster (c, d, e). The three-element cluster has a similarity
consolidation where the number of indicators in each of the coefficient of 0.6. Finally, the cluster (a, b) is grouped
four groups are as follows: 9 indicators under social factors, with the cluster (c, d) forming a new cluster of all
9 indicators under ecological factors, 13 indicators under elements (a, b, c, d, e) with an overall similarity
economic factors, and 4 indicators under policy factors. coefficient of 0.5. The same process was used for the
Tables A1–A4 in Appendix A contain the results of the different groups of indicators.
comparisons of stakeholders’ perceptions with the con- The final grouping of clusters, or final partition, is the
solidated list under each group. If a stakeholder’s perceived grouping that will identify groups whose members share
indicator matched an indicator in the consolidated list, a common characteristics. There is no single guideline that
number ‘1’ is added; otherwise the indicator is assigned a determines what the final grouping should be. This could
value ‘0’. be based on the desired number of clusters, or
alternatively, this could be determined by the ‘acceptable
3.3. Hierarchical clusters acceptable amount of similarity’. This can be best
illustrated using a ‘dendrogram’ as shown in Fig. 2. As
The results of the group average linkage based on the the name implies, a dendrogram represents a ‘similarity
simple matching similarity coefficients are shown in tree’ where branches symbolize the ‘stages’ by which
Tables 5–8. As described in Section 2.2.2 the elements are grouped or clustered as discussed above.
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126
Table 4
Scoring matrix provided by key group of stakeholders
Criteria Inhutani II Inhutani II Community Kanwil Univer- Univer- Coal CIFOR NGO Local goverment
Malinau East Kali- (central sity sity mining
Paya Long Metud Rian Loreh Langap YAP BIOMA PLASMA Sub Provin- Sub-
Sub Unit mantan govern- lecturers students
Seturan Lake District cial District
Unit ment)
Forest Forest Office
Service Service

Proximity 1 3 1 1 1 1 1 1 4 2 4 1 4 2 3 3 1 1 1
Pre-existing 5 5 1 1 1 1 1 1 5 5 5 5 5 3 3 3 4 5 3
right
Dependency 1 1 1 1 1 1 1 1 3 2 2 5 3 2 3 3 1 2 3
Forest man- 1 1 1 1 1 1 1 1 2 3 1 5 1 1 2 2 3 1 3
agement
Knowledge
Forestry 1 1 1 1 1 1 1 1 2 3 1 5 1 1 1 1 2 2 3
spirit
Daily 1 1 1 1 1 1 1 1 3 3 3 1 3 2 2 2 1 1 3
activity on
site
Legal rights 1 1 5 5 5 5 5 5 1 1 5 1 5 4 4 4 1 1 2
Total 11 13 11 11 11 11 11 11 20 19 21 23 22 15 18 18 12 13 18
Mean 1.6 1.9 1.6 1.6 1.6 1.6 1.6 1.6 2.9 2.7 3.0 3.3 3.1 2.1 2.6 2.6 1.9 1.9 2.6

The cut off value for determining ‘Who Counts’ is determined using the mean. A stakeholder that receives final score of 3 or less is considered to be important and therefore included. This cut off value was determined based on field experiences
in many countries: stakeholders having means less than or equal to the cut off tend to play the important roles in managing natural resources (Colfer et al., 1999).

117
118 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

Fig. 4. Stakeholders located inside and outside the FMU.

This is illustrated in the dendrogram shown in Fig. 2 for perspectives as far as the set of ecological indicators are
the five-element sample described above. The hierarchical concerned.
partitioning or branching of clusters enables the specifica- Similarly, the hierarchical clustering for the economic
tion of ‘acceptable level of similarity among clusters’. indicators is summarized in Table 7. Using the 80%
Specifying an acceptable level of similarity is akin to similarity as a cut-off point, the results show that 82% (14
‘cutting the dendrogram’; that is, drawing a line across the out of 17) of the stakeholders have similar perspectives in
dendrogram to specify the final grouping. As shown in terms of the set of economic indicators. Finally, based on
Fig. 1, the fusion process progresses with decreasing the results shown in Table 8, 14 out of the 17 stakeholders
similarity or increasing dissimilarity. Hence, one can have at least 80% similar views in terms of the set of policy
specify a priori the level of (dis)similarity as a cut off indicators.
point to determine the final partition or grouping. For
example, cutting the dendrogram in Fig. 2 at a dissim-
ilarity level of 0.4 (or similarity level of 0.6) implies that
Table 5
only the cluster (c, d, e) is acceptable. Cluster analysis using group average linkage of social indicators
Similarly, in the cluster analysis shown in Table 5 for
Agglomeration schedule
social indicators, one may specify 80% similarity as
acceptable. Hence, all groups or partitions that exhibit this Stage Cluster combined Coeffi- State cluster first Next
cients appears stage
similarity will be considered. In the group of social
indicators example shown in Table 5, all clusters formed Cluster Cluster Cluster Cluster
1 2 1 2
until stage nine are considered. Hence, in this case one can
conclude that the clusters fused together from the previous 1 11 16 1.000 0 0 4
2 1 14 1.000 0 0 6
stages have similarities higher than 80%. For instance,
3 8 13 1.000 0 0 8
according to Table 5, the pairs or clusters (11, 16), (1, 14), 4 3 11 1.000 0 1 7
(8, 13), and (3, 11) all have similarity coefficient of 100%. 5 7 15 0.889 0 0 12
Table A1 shows that these pairs have in fact similar sets of 6 1 12 0.889 2 0 12
values. Cumulatively, for social indicators, about 75% (12 7 3 9 0.889 4 0 9
8 6 8 0.889 0 3 9
out of 17) of the stakeholders (i.e. stakeholders 1, 3, 6, 7, 8, 9 3 6 0.824 7 8 13
9, 11, 12, 13, 14, 15, and 16) can be assumed to have at least 10 4 17 0.778 0 0 14
80% similar perspectives in terms of the set of social 11 2 10 0.778 0 0 16
indicators. 12 1 7 0.759 6 5 13
The ecology indicators exhibit even more similarity. As 13 1 3 0.711 12 9 14
14 1 4 0.611 13 10 15
shown in Table 6, clusters formed until stage 12 have a 15 1 5 0.579 14 0 16
similarity coefficient of 80%. In fact, 88% (15 out of 17) 16 1 2 0.556 15 11 0
of the stakeholders have at least 80% similarity in
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126 119

Table 6 Table 8
Cluster analysis using group average linkage of ecological indicators Cluster analysis using group average linkage of policy indicators

Agglomeration schedule Agglomeration schedule

Stage Cluster combined Coeffi- State cluster first Next Stage Cluster combined Coeffi- State cluster first Next
cients appears stage cients appears stage
Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster
1 2 1 2 1 2 1 2
1 9 17 1.000 0 0 4 1 16 17 1.000 0 0 2
2 3 13 1.000 0 0 9 2 1 16 1.000 0 1 9
3 11 12 1.000 0 0 10 3 14 15 1.000 0 0 4
4 6 9 1.000 0 1 5 4 10 14 1.000 0 3 6
5 6 8 1.000 4 0 8 5 12 13 1.000 0 0 6
6 5 7 1.000 0 0 7 6 10 12 1.000 4 5 7
7 4 5 1.000 0 6 12 7 10 11 1.000 6 0 13
8 6 16 0.889 5 0 11 8 6 7 1.000 0 0 9
9 3 14 0.889 2 0 13 9 1 6 1.000 2 8 11
10 10 11 0.889 0 3 16 10 4 5 1.000 0 0 11
11 2 6 0.867 0 8 12 11 1 4 1.000 9 10 12
12 2 4 0.852 11 7 13 12 1 3 1.000 11 0 13
13 2 3 0.790 12 9 14 13 1 10 0.750 12 7 14
14 2 15 0.694 13 0 15 14 1 9 0.643 13 0 15
15 1 2 0.641 0 14 16 15 1 8 0.633 14 0 16
16 1 10 0.524 15 10 0 16 1 2 0.563 15 0 0

3.4. Testing of hypotheses

hierarchical methods is that fusions, once made, are
irrevocable, so that when two individuals have been
As pointed out earlier, one of the objectives of the study
joined, they cannot subsequently be separated. Hence, it
was to test three hypotheses reflecting the views of the
is conceivable that a group of indicators may manifest high
NGOs, local communities, timber companies, government
similarity as a group or cluster, but may be statistically
officials, and local and regional organizations. These tests
different when examined in pairs. This is the motivation for
will shed further light on how close the similarities of
testing the hypothesis about the perspectives of
perspectives are between stakeholder groups. Note that
certain groups of stakeholders. Moreover, the tests
cluster analysis examines groups or clusters with their
described below are over the whole set of indicators
associated similarity measures. One weakness of the
making no distinction among the groups (i.e. social,
ecology, ecology, and policy).
Table 7 The first hypothesis, simply put is: Local communities’
Cluster analysis using group average linkage of economic indicators perspectives are close to those of the NGO’s. This
Agglomeration schedule hypothesis was tested because of the close relations
Stage Cluster combined Coeffi- State cluster first Next between NGOs and local communities, and because
cients appears stage NGOs often speak on behalf of the communities against
Cluster Cluster Cluster Cluster the government. The null hypothesis tested is:
1 2 1 2
Ho : plc Z pngo H1 : plc spngo (2)
1 12 17 1.000 0 0 2
2 6 12 1.000 0 1 3 As described in Section 2.2.3, the null hypothesis is
3 6 11 1.000 2 0 5
tested using the Pearson chi-square developed by Everitt
4 4 7 1.000 0 0 9
5 6 16 0.923 3 0 8 (2001). The result of the test is shown in Table 9. From the
6 13 15 0.923 0 0 10 analysis it is very clear that the first hypothesis should be
7 9 14 0.923 0 0 11 rejected in most cases, as the perceptions of NGOs and
8 6 8 0.908 5 0 9 local communities are generally not similar. It is clear from
9 4 6 0.897 4 8 12
the results shown in Table 9 that the null hypothesis should
10 10 13 0.885 0 6 14
11 3 9 0.885 0 7 12 be rejected for most of the pairs of local community and
12 3 4 0.833 11 9 13 NGO comparisons. Hence, in general the local commu-
13 3 5 0.783 12 0 14 nities have different perceptions in terms of what they
14 3 10 0.733 13 10 16 consider to be the important set of indicators of sustainable
15 1 2 0.615 0 0 16
forest management compared to the perceptions of the
16 1 3 0.582 15 14 0
three NGOs.
 120 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

Table 9 Table 10
Test of hypothesis on the similarity of perspectives between local Test of hypothesis on the similarity of perspectives between government
communities and NGOs and timber company

No Local NGO Asymptotic sig- Remark No Govern- Concession Asymptotic signifi- Remark
community nificance of Pear- ment cance of Pearson
son Chi-square Chi-square
1 Paya Seturan YAP 0.008 Accept H0* 1 Provincial Inhutani II East 0.886 Accept
2 Paya Seturan BIOMA Accept H0** Forest Kalimantan Unit H1
0.030
Service
3 Paya Seturan PLASMA 0.082 Accept H1 2 KANWIL Inhutani II East 0.572 Accept
4 Long Lake YAP Accept H1 Kalimantan Unit H1
0.759
3 Sub Dis- Inhutani II Malinau 0.972 Accept
5 Long Lake BIOMA 0.889 Accept H1 trict Office Sub Unit H1
6 Long Lake PLASMA 0.698 Accept H1 4 Forest Sub Inhutani II Malinau 0.972 Accept
District Sub Unit H1
7 Long Loreh YAP 0.791 Accept H1
Office
8 Long Loreh BIOMA 1.000 Accept H1
9 Long Loreh PLASMA 0.178 Accept H1 timber concession directly providing financial and human
10 Metud YAP Accept H1 resources to stimulate local economic development. NGOs
0.111
11 Metud BIOMA Accept H1
on the other hand focused more on fair benefit sharing
0.227
between the communities and the concession as a condition
12 Metud PLASMA 0.823 Accept H1
of good forest management and were concerned about the
13 Rian YAP 0.170 Accept H1
absence of law enforcement in the area. Local communities
14 Rian BIOMA 0.324 Accept H1 also mentioned the importance of recognizing boundaries of
15 Rian PLASMA 0.692 Accept H1 their paddy rice fields and other land uses with that of the
concession area.
*a, 0.01; **a, 0.05.
These differences occur because of the way sustain-
ability is perceived. The three NGOs in the study perceived
Survey Questionnaire Used in the Study to test the variables: sustainability in a strategic way, while the communities
perceive it more practically. The differences may also occur
Symposium because of gaps in formal education among stakeholders.
 Conduct fire safety lecture and drill to all building Most NGO members were graduates of high schools or
 Conduct orientation earthquake and drill universities and frequently work together with researchers,
 Conduct seminars about first aid and rescue university activists and other NGOs at provincial, national or
 Conduct Jr. Kiddie Fire Marshall (School) even international levels. Mean- while, most local
 Conduct and organize by some local or a company community residents had only been to elementary schools.
House to house campaign The local communities were interested in immediate
 Conduct information drive improvements of their daily lives, meanwhile the NGOs
 Giving fire safety tips tended to take a more systemic and long-term perspective.
 Organize pink fire brigade The second hypothesis: Timber concessionaires’ per-
 Conduct house to house fire safety inspection spectives are close to those of the local government because
 Conduct fire safety campaign
Physical hazard
 Well trained fire fighters in the country
 Modernized fire firefighting equipment and
accessories are provided by the government
 Adequate communication system
 Adequate fire hydrant/or source of water
 Fire safety campaign is done regularly
Occupational hazard
 To save lives and property
 Quick response during the occurring of fire
 Adequate or effective salvage operation to all
buildings regularly
 To rescue fire trap victims
 Obstruction of firefighting operation during of fire
occurrence

The local communities perceived good forest manage-

ment as a way to practically improve their daily lives from
the existing situation; for instance they are interested in the
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126 121
the concessionaire is theoretically expected to follow
regulations assigned by the government, was also tested
as follows:

H0 : phph Z pgovt H1 : phph spgovt

(3)

The results of the test are presented in Table 10. The

results show that in general, there are significantly
different perceptions between the government and the
timber company about the set of indicators of sustainable
forest management.
The timber concession perceived sustainability as a
duty to stimulate economic development in the villages,
move towards better and environmentally friendly timber
harvesting, clarify boundaries between their concession
and other land uses, and improve communication
between themselves and communities. Local
development agencies perceived sustainability as a
minimization of conflicts, good quality of food for low
ranking conces- sion employees who are mostly from
local communities, better quality and stable quantities of
water in the river, and research to support sustainable
forest management. To measure sustainability the local
governments proposed observing the diameter of logs in
the log pond. If these diameters continued to decrease
then the inference would be that the forest was not
being well managed.
The third hypothesis, urban and field-based offices of
the same organization have similar perceptions, was tested
using

Table 11
Test of hypothesis on the similarity of perspectives between urban and
field-based units

No Urban institution Field-based Asymptotic signifi- Remark

institution cance of Pearson
Chi-square
1 Provincial Forest Forest Sub 0.386 Accept
Service District H1
Office
2 Inhutani II East Inhutani II 0.599 Accept
Kalimantan Unit Malinau Sub H1
Unit
 122 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

the same method above. The null and alternative hypotheses that the same institution (i.e. Inhutani, forestry offices) at
are: different levels (i.e. field-based and urban based) may
have different views with respect to what they perceive to
H0 : purban Z pfield H1 : purban spfield (4)
be important indicators of sustainable forest management.
The third hypothesis was also rejected based on the test as This suggests that there can be gaps in perception between
shown in Table 11. This means that there are different field-based and urban-based institutions. Part of the
perceptions between urban and field-based personnel of the problem may be caused by inadequate communication
same organization. The urban unit of Inhutani II in East between the different levels or organizational units, or
Kalimantan perceived sustainability as involving the local different types of work, background, and incentives
communities in the decision-making process, improving the among institutional units.
forest management plan, doing research and securing This divergence of opinions highlights the need for better
the forest area. The field unit of Inhutani II in Malinau communication among and between groups of stakeholders.
perceived it as reducing logging impacts and providing area For instance, the seemingly disparate views between local
for conservation. Furthermore, provincial forest service communities and NGO’s raises an important issue with
perceived sustainability as involving local communities in respect to the ability of NGO’s to represent, or speak on
the decision-making process and maintaining the ecosystem, behalf of, local communities. Careful analysis and good
while the sub-district forest service perceived it as more communication between these two groups must be
assistance for local communities from the timber concession exercised to insure that NGOs are adequately represent-
and low conflict between the timber concession and the local ing the views and perspectives of the local community.
communities. Apparently this difference is caused by the Similarly, differing views between the timber company
nature of their work. Urban offices tended to take a more and the government also highlights the need for closer
systemic, process oriented view, which was somewhat coordination between these two groups to insure that
theoretical. Offices closer to the field were more interested forest utilization policies affecting, or affecting by, timber
in concrete actions aimed at ameliorating the existing utilization companies are properly targeted towards the
situation. key indicators commonly perceived as significant by
both groups.
The insights and inferences gained from this study are
consistent with results reported by other authors particularly
4. Conclusions and recommendation Kearney et al. (1999) and Pokorny and Bauch (2003). Using
statistical methods and clustering techniques, this study
Different stakeholders may view sustainable forest confirms what the two previous studies have concluded
management in many different ways. However, in many using other analytical tools—e.g. conceptual content
respects neighboring local communities have been known cognitive map (Kearney et al., 1999). Results from the
to share similar perspectives about sustainable forest three studies suggest that there can be differences in
management. In this study, cluster analysis and hypoth- perspectives among stakeholders. Kearney et al. (1999)
esis testing were used to examine measures of simi- concluded that different stakeholders not only have different
larities of perspectives and to statistically determine how perspectives but that these differences are often conceptu-
close these similarities are between groups of alized in terms of stereotypes. Pokorny et al. (2004) also
stakeholders. concluded that evaluations of criteria and indicators often
Results from the study indicate that local villages in stem from differences in perspectives about verifiers of
the FMU often have similar perceptions on factors or the indicators. The results of the cluster analyses and
indicators of good or sustainable forest management. The the tests of hypothesis on similarity/dissimilarity of
results also show that there are groups or clusters of perspectives conducted in this study conform to these
stakeholders that show high similarities in certain groups findings.
of indicators (e.g. a large percentage of the selected
stakeholders exhibit high similarity in their perceptions of
good ecological indicators). Further analysis of the results Acknowledgements
also shows that while certain groups or clusters show high
similarity of perceptions collectively, many groups This study was conducted with the participation of Yurdi
actually have divergent views when examined more Yasmi and Carol Colfer of the Center for International
closely using hypothesis testing. This suggests that Forestry Research (CIFOR), Bogor, Indonesia.
although collectively there is general agreement among
groups of stakeholders on the set of indicators under each
sub-group (e.g. social, ecological, or economic) on a finer Appendix A
scale groups may in fact hold significantly different views.
For example, one interesting finding of this research is Tables A1–A4.
 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126 123

122
Table A1
Stakeholders’ perceptions on social indicators of sustainable forest management compared to consolidated set

No Consolidated list Inhutani Inhutani Pro- Sub PLASMA YAP Sub BIOMA Uni- Long Metud Rian Paya Long Lan- KAN- Univer-
of indicators II Mali- II East vincial District District versity Lake Com- Com- Seturan Loreh gap WIL sity
nau Sub Kaliman- Foerst Forest Office stu- Com- munity munity Com- Com- lecturer
Unit tan Unit Ser- Service dent munity munity munity
vice
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1 Concession provides a 1 1 0 1 0 0 1 0 0 1 0 1 0 1 1 0 0
wide range of assistance

H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

to local communities/
villages including finan-
cial and human resource
development asisstance
2 Communities are 0 1 1 0 1 1 1 1 1 0 1 1 1 0 1 1 0
involved in forest man-
agement and decision-
making processes
3 The concession person- 0 0 0 0 1 1 0 1 0 1 0 0 1 0 0 0 0
nel respect and are aware
of local values and forest
functions such as hunt-
ing, communities’ land,
forest resources, etc.
4 Good communication 0 1 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0
and relationships
between the concession
and communities
5 Conflict between the 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 1
concession and commu-
nities is minimum
6 The concession should 0 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0
be able to stimulate
economic development
in the village
(i.e by creating job
opportunities)
7 Fair sharing of forest 0 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0
benefits
8 The concession provides 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
good quality of food for
its employees
9 Results of the research 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
are explained to
community

Presence (1) or absence (0).

Table A2
Stakeholders’ perceptions on ecological indicators of sustainable forest management compared to consolidated set

No Consolidated Inhu- Inhu- Provin- Sub PLASMA YAP Sub BIOMA Univer- Long Metud Rian Paya Long Langap KANWIL Univer-
list of tani II tani II cial District District sity Lake Com- Com- Seturan Loreh Com- sity
indicators Mali- East Foerst Forest Office student Com- munity munity Com- munity lecturer
nau Kali- Service Service munity munity
Sub mantan
Unit Unit
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

1 Forest ecosys- 1 1 1 0 0 1 0 1 1 1 1 1 1 1 1 1 1
tem is main-
tained
2 Water quality 1 0 1 0 0 0 0 0 0 1 1 1 1 1 0 0 0
and quantity of
the stream
remain in a
good condition
3 There are gene 1 0 0 0 0 0 0 0 0 1 1 1 0 0 1 0 0
bank areas
4 The concession 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0 0
respects local
use of the
resources and
carries out
replanting
activities
5 Soil erosion is 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
relatively low
6 Various ani- 0 0 0 0 0 0 0 0 0 1 1 1 0 1 0 0 0
mals continue
to exist in the
forest
7 A buffer zone 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
between the
concession and
community
exists
8 Applied 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
research should
be carried out
to support SFM
9 Soils are in a 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0
good condition

123
 124
Table A3
Stakeholders’ perceptions on economic indicators of sustainable forest management compared to consolidated set

No Consolidated list Inhutani Inhutani Provin- Sub Dis- PLASMA YAP Sub Dis- BIOMA Univer- Long Metud Rian Paya Long Langap KANWIL University
of indicators II Mali- II East cial trict trict sity stu- Lake Commu- Commu- Seturan Loreh Commu- lecturer
nau Sub Kali- Foerst Forest Office dent Commu- nity nity Commu- nity
Unit mantan Service Service nity nity
Unit
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1 Harvesting 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0
activity is
carried out

H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

according to
regulations
2 Harvesting is 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
well equipped,
funded and
environmentally
friendly
3 Managament 0 1 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0
plan is compre-
hensive
4 Damage and 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
other off site
impacts from
harvesting
activity should be
minimized
5 Amount of 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
annual allowable
cut and cutting
areas must be
clear
6 Amount of log 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
production match
with pre harvest
inventory
7 Economic returns 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
from the forest
are re-invested in
the forest
8 Communities 0 0 1 0 0 0 0 0 1 1 0 0 0 1 0 0 0
benefit from har-
vesting and their
economy
improves
9 Primary forests 0 0 0 0 0 0 0 0 0 1 0 0 1 0 1 0 0
are maintained
for future needs
10 Communities can 0 0 0 0 0 0 0 0 0 1 0 0 1 1 1 0 0
utilize forest
resources for
their economy
and daily needs
11 Forest tending is 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
carried out
appropriately
12 Big logs in the 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
log pond
13 Regional devel- 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
opment needs
should be con-
sidered by the

H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

concession

Table A4
Stakeholders’ perceptions on policy indicators of sustainable forest management compared to consolidated set

No Consolidated list of indi- Inhu- Inhu- Pro- Sub PLAS- YAP Sub BIOMA Uni- Long Metud Rian Paya Long Lan- KAN- Univer-
cators tani II tani II vincial Dis- MA Dis- versity Lake Com- Com- Setur- Loreh gap WIL sity
Mali- East Foerst trict trict stu- Com- munity munity an Com- Com- lecturer
nau Kali- Ser- Forest Office dent munity munity munity
Sub man- vice Ser-
Unit tan vice
Unit
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1 Boundaries between differ- 0 1 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0
ent land uses exist and are
clear
2 Security of the forest area 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
3 Classification of forest to 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
ensure good management
practice (i.e. protection,
production, etc.)
4 Sound law enforcement 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0

125
126 H. Purnomo et al. / Journal of Environmental Management 74 (2005) 111–126

making, in: Schmoldt, D., Kangas, J., Mendoza, G., Pesonen, M. (Eds.),
References
The Analytic Hierarchy Process in Natural Resources and Environ-
mental Decision Making. Kluwer, Dordrecht (Chapter 8; in press).
Colfer, C.J.P., Brocklesby, M.A., Diaw, C., Etuge, P., Gu¨nter, M., Harwell, E.,
Plotkin, J.B., Chave, J., Ashton, P., 2002. Cluster analysis of spatial
McDougall, C., Porro, N.M., Porro, R., Prabhu, R., Salim, A.,
patterns in Malaysian tree species. The American Naturalist 160 (5),
Sardjono, M.A., Tchikangwa, B., Tiani, A.M., Wadley, R.L., Woelfel, J.,
629–644.
Wollenberg, E., 1999. The BAG (Basic Assessment Guide for Human
Pokorny, B., Bauch, R., 2003. Participatory analysis of heterogeneity, an
Well-Being), C&I Toolbox Series No. 5. CIFOR, Bogor, Indonesia.
approach to consolidate initiatives at community level. Forests, Trees,
Ducey, M.J., Larson, B.C., 1999. A fuzzy approach to the problem of
and Livelihoods 13 (2), 161–176.
sustainability. Forest Ecology and Management 115, 29–40.
Pokorny, B., Prabhu, R., McDougall, C., Bauch, R., 2004. Local
Everitt, B.S., 1993. Cluster Analysis, third ed. Edward Arnold, London.
Everitt, B.S., 2001. Cluster Analysis, fourth ed. Edward Arnold, London. stakeholders’ participation in developing criteria and indicators of
Ferguson, I.S., 1996. Sustainable Forest Management. Oxford, Melbourne, sustainable forest management. Journal of Forestry 102 (1), 35–40.
Australia. Prabhu, R., Colfer, C., Venkateswarlu, P., Tan, L.C., Soekmadi, R.,
Kaufmann, L., Rousseeuw, P.J., 1990. Finding Groups in Data: An Wollenberg, E., 1996. Testing Criteria and Indicators for the
Introduction to Cluster Analysis. Wiley/Interscience, New York. Sustainable Management of Forests. Phase I. Final Report. CIFOR,
Kearney, R.A., Bradley, G., Kaplan, R., Kaplan, S., 1999. Stakeholder Bogor, Indonesia (Special Publication).
perspectives on appropriate forest management in the Pacific North- Purnomo, H., Mendoza, G., Prabhu, R., 2003. A collaborative model for
west. Forest Science 45 (1), 62–73. community-managed resources: a qualitative soft systems approach,
Maser, C., 1994. Sustainable Forestry: Philosophy, Science and and Department of Natural Resources, University of Illinois. Working Paper
Economics. St Lucie Press, Florida. 2003-1.
Mendoza, G.A., Prabhu, R., 2000a. Multiple criteria decision making Selener, D., 1997. Participatory Action Research and Social Change. The
approaches to assessing Forest sustainability using criteria and Cornell Participatory Action Research Network. Cornell University,
indicators: a case study. Forest Ecology and Management 131, 107–126. New York.
Mendoza, G.A., Prabhu, R., 2000b. Multiple criteria analysis for assessing Varma, V.K., Ferguson, I., Wild, I., 2000. Decision support system for
criteria and indicators in sustainable forest management: a case study on sustainable forest management. Forest Ecology and Management 128,
participatory decision making in a Kalimantan forest. Environmental 49–55.
Management 26 (6), 659–673. Wulder, M.A., 2002. Practical Guide to the Use of Selected Multivariate
Mendoza, G.A., Prabhu, R., 2001. Prioritizing criteria and indicators for Statistics http://www.pfc.forestry.ca/profiles/wulder/mvstats/index_e.
sustainable forest management: a case study on participatory decision html
View publication stats