Nature conservation in Cross River National Park, south-east Nigeria: promoting collaboration between local people and conservation authorities

Abstract

In most developing countries, there has been a long-standing conflict between nature conservation and local demands for natural resources. This paper reports a study on the preference of local people for different incentives that could help increase local support for nature conservation. It also explores the possibilities for designing a sustainable incentive strategy. Data were obtained from personal interviews conducted with community members around the Okwangwo Division of the Cross River National Park in south-east Nigeria, and were analysed using a multinomial logit model. The results showed that skill development was the most preferred incentive, followed by community forestry. The gathering of non-timber forest products under the supervision of national park officials was least preferred. The local people's preferences were influenced by factors such as education, extraction of bush mangoes, membership of environmental group, farmland size, income from non-timber forest products, ownership of commercial cocoa farms and gender. We conclude that an incentive-based strategy for nature conservation will be more effective if local people are more involved in negotiations regarding the incentive that are being offered. The findings from this study could assist nature conservation managers in designing a more acceptable and effective nature conservation strategy.

Keywords:

• land use conflict
• local participation
• nature conservation
• preferences
• sustainability
• utility

Introduction

Many governments in developing countries face challenges regarding how best to conserve nature. In the past, nature conservation strategies have been dominated by the establishment of government-run conservation projects such as protected areas, which is the ‘classical’ approach to conservation (Blaikie & Jeanrenaud 1997; Adams & Hulme 2001; Adams 2004; Kamuaro 2007). However, these strategies have not been very successful in preventing deforestation and the associated loss of biodiversity (Emerton 1997; Geist & Lambin 2002). This is because nature conservation often conflicts with the local demands for land and forest products (Shyamsunder & Kramer 1996; Newmark & Hough 2000; Salafsky & Wollenberg 2000; Kapoor 2001). Such conflicts have made protected area managers more aware of the importance of incorporating local people's needs in nature conservation strategy (Parry & Campbell 1992; Berkes 2004; Ostrom 2005). Recent nature conservation approaches have focused on local participation and development as means of establishing local support for nature conservation (Wells & Brandon 1992; Narendran et al. 2001; Newton et al. 2006). These approaches include community-based conservation (Western et al. 1994) and integrated conservation with development programmes (Brandon & Wells 1992). Community-based conservation focuses on the involvement of local people in planning and implementing conservation projects to improve accountability and transparency in the management of natural resources. The integrated conservation with development programmes entail providing local communities with incentives (e.g. social and economic development) that can motivate them to support nature conservation (Adams & Hulme 2001). The local participation and development approach allows people who reside in the vicinity of nature conservation areas to participate in the conservation process and link the objectives of conservation with their development needs (Hutton & Leader-Williams 2003; Oldekop et al. 2009).

The limitations of nature conservation projects (e.g. protected areas) regarding meeting local people's needs and conservation have led to the emergence of incentive-based programmes (Newmark & Hough 2000; Abbot et al. 2001; Zimmerman et al. 2001). These programmes combine the strengths of the community-based conservation and integrated conservation with developmental programmes. Incentive-based programmes were developed to help reduce conflicts between the needs of local people and the need to protect the environment. Although some authors (Noss et al. 1999; Soule & Terborgh 1999) have questioned the compatibility of conservation and development, others (Salafsky & Wollenborg 2000; Brown 2002) have argued that a well-designed incentive-based programme can generate positive perceptions of conservation projects and lead to environmental stewardship. Some examples of incentive-based programmes that can achieve economic, social and ecological objectives include tourism development (e.g. park-based employment and tourism revenue sharing), limited harvesting of natural resources and community forestry (Brown 2002; Palmer & Engel 2007). Although meeting the economic needs of local communities in the vicinity of conservation areas while minimizing environmental degradation and biodiversity loss is crucial in Africa (Myers et al. 2000; Mena et al. 2006), the success of participatory approaches is rare (Emerton 1999; Infield & Namara 2001; Salafsky et al. 2001). This may be linked to mistrust between local communities and conservation agencies and the fact that benefits that local people get from a conservation area are often smaller than the losses they incur due to the establishment of nature conservation areas. On the other hand, where local communities are involved in conservation initiatives, their participation is often a component of governmental or non-governmental organization-sponsored project in which the communities have lesser management power (Nelson & Gami 2002). Several factors have been identified that influence the local people's attitudes towards nature conservation and land use. These include education level, gender, ethnicity and years of employment (Mehta & Heinen 2001; McClanahan et al. 2005), membership of social groups and income (Pretty & Ward 2001; Bayard et al. 2007). Other factors such as the distance of the local people's home from the conservation area, household size, occupation and social and economic indicators are also contributory (Wietze 2000; Shrestha & Alavalapati 2006). Some other factors that influence the local people's attitudes are the amount of economic benefits that they get from enterprises linked to conservation area (Walpole & Goodwin 2001), its use and intrinsic values (Bauer 2003), dependency on the forest, wealth, level of land use conflicts and participation in community-based projects (Gillingham & Lee 1999; Gadd 2005; Kideghesho et al. 2007), farmland size, the relationship between local people and conservation officials as well as the years of residency (Newmark et al. 1993). The understanding of all these factors is important in improving the relationship between local people and conservation efforts.

Although the knowledge of the local communities' preferences for incentives that can encourage them to support nature conservation and factors that influence their preferences are important in designing a sustainable conservation strategy (Naughton-Treves et al. 2005), these do not often receive much attention in Nigeria. Designing such a strategy often requires some valuation of the benefits and costs associated with nature conservation. Stated preference methods such as the choice experiments (Boxall et al. 1996; Adamowicz et al. 1998; Blamey et al. 2000) can be applied in such a valuation study. In a choice experiment survey, the respondents are presented with several alternatives and are asked to choose the most preferred alternative. The values of the respondents are inferred from the choices or trade-offs they make. Some published choice experiment studies in Africa that have focused on environmental valuation include local participation in forest management (Brannlund et al. 2009), local preferences for different soil conservation regimes (Bekele & Drake 2003) and farmers' preferences for climate change adaptation strategies (Hassan & Nhemachena 2008). Although incentive-based conservation programmes have been implemented in many developing countries, conflicts between the locals and conservation efforts persist mainly because local people are often not involved in negotiations about the incentives they are to receive (Kellert et al. 2000). The aim of this study was to examine the local people's preferences for different incentives and factors influencing the preferences to help increase local support for nature conservation as well as to improve the current situation. The study involves communities located in the buffer zone of the Okwangwo Division of the Cross River National Park in south-east Nigeria.

Methods

The study site

The Federal Government of Nigeria established the Cross River National Park in 1991 (Nigeria Park Service 2009) to conserve some of the natural rainforests. The park is located in the Cross River State in south-east Nigeria, which harbours most of the rainforest habitat in the country. The Cross River National Park is officially managed as two Divisions, the Oban Division and the Okwangwo Division (OD). The OD – which is the study area – is part of the Boki Local Government Area and covers an area of 1000 km². It borders the Takamanda Forest Reserve in the Republic of Cameroon to the east, which implies that it is also important for regional biodiversity conservation. It is one of the United Nation's biodiversity hotspots (Biodiversity Hotspots 2009). For example, 1545 plant species and more than 1000 species of butterfly have been recorded in the park (Larson 1997). The primate fauna of the OD is diverse and there is confirmed existence of the endangered Cross River gorilla (Gorilla gorilla diehli) (Oates et al. 1990; Eniang 2001). The OD is important for regional watershed protection, protection of biodiversity and the potential for gorilla-based tourism (Caldecott et al. 1990). Along with the creation of the OD, a ‘buffer zone’ which is the study area was also designated by the OD authorities to protect the park. This zone comprises 66 villages (approximately 36,000 people, Ite 2004), whose sources of livelihood were affected by the establishment of the OD (e.g. decrease in agriculture and forest products because of reduced access to land). The economy of the people is largely agrarian, with importance attached to hunting, trapping and gathering of non-timber forest products (Ite 1995; Ezebilo & Mattsson 2010a). Agriculture in this area is characterized by small-scale subsistence mixed-farming system, with livestock production as an integral part. Cash income for household financial requirements is mainly generated from the sale of agricultural products such as cocoa and banana; non-timber forest products such as Gnetum africanum Welw (afang) which is a leafy wild vegetable, and Irvingia gabonensis (bush mango) (Ezebilo 2010). Some residents engage in formal employment such as teaching and nursing while others engage in large-scale cocoa and banana farming and bush mango trade.

Interviews and data collection

Data were collected by means of personal interviews. The questions used for the interview were designed through focus group discussions and pre-test interviews. The focus group comprised 15 people including village heads, youth leaders, heads of women's groups and secondary school teachers whose works were related to environmental education. During the focus group discussions, the group was subdivided into three. Each of the subgroups comprised a village head, two youth leaders, the head of women's group and a teacher who helped note the minutes of the discussion. The moderator of the focus group started the discussion by introducing the subjects and each of the subgroups were asked to state their opinion regarding their perceptions of the OD. The subgroups were asked to discuss the ‘local people versus OD’ conflicts and what could be done to resolve the conflicts. The subgroups were further asked to discuss different strategies that could motivate the local support for nature conservation in the OD. They were asked to design simple questions that can help gain information from the locals regarding what could be done to improve their support for nature conservation in the OD. Questions were drafted based on the summary of the reports from the focus group discussions. The questions were presented to the focus group at the second meeting and members were asked to comment on any difficulty in interpreting the questions and to suggest ways to improve the questions. Some issues were raised; for example, some members of the focus group raised concerns about the inclusion of ‘a no-change scenario’ (i.e. the present situation where the locals are only allowed access to non-timber forest products in the buffer zone of the OD) among the incentive strategies. They were of the opinion that some locals may think that the interview is either sponsored by the national park managers or the government. Because the local people are often suspicious about anything that has to do with the government, they may not be willing to participate in the interviews. As an alternative, the members of the focus group suggested an incentive that was concerned with access to non-timber forest products in the national park under the supervision of the park officials. This led to further focus group discussions and the questions were modified to meet the raised concerns. To validate the questions, 20 people who lived around the OD were randomly selected and interviewed in November 2007. To explore whether other local people have concerns similar to those raised by the members of the focus group concerning the ‘no-change scenario’, the interviewees, during the pre-test, were asked to comment on the ‘no-change scenario’ that was rejected by the focus group. Surprisingly, all of them raised similar concerns as those by the focus group. The questions were further modified to capture the concerns raised by the people involved in the pre-test interviews.

The main interviews were conducted in April and May 2008 with the help of two assistants who were fluent in the local language (Boki) of the study area. The assistants were recruited and trained for two days in interview techniques. Because the interaction between local people and the OD officials had the potential of impacting the local people's' attitudes towards conservation (Newmark et al. 1993), the 66 villages located in the buffer zone of the OD were grouped into three (i.e. 1–5 km, 6–10 km and >10 km radius) based on their geographic location in relation to the OD administrative head office. One village was randomly selected from each of the group. The three selected villages were Butatong, Bukalom and Wula. Of the three villages, Butatong is the closest to the OD administrative office, followed by Bukalom. The interviews were administered in the evenings to include as many categories of the population as possible during the study period. Every other house along the main street of each village was visited. If a house was not occupied, then it was skipped and the next house was visited. The interviewees alternated between the eldest male and the eldest female in each selected household. If the gender of the eldest in a household did not coincide with the interviewees' selection method, then a member of the opposite gender was interviewed and the interviewee's gender was alternated again there onwards. Interviews were conducted in the Boki language. A total of 150 people were interviewed (50 in each village). The interviews began by seeking approval from the village heads of the selected villages, which is in accordance with the tradition of the local people. This was also a way of legitimizing the interviews in the villages. The interviews involved people aged 18–83 years and they were mainly farmers, teachers, nurses, store keepers, tourist guides and students. After explaining the purpose of the interview (i.e. increasing the understanding of improving the local support for nature conservation) and assuring the interviewees about the confidentiality of their responses, they were asked a series of socio-economic questions (education, household size, farmland size, income from non-timber forest products, age, occupation and income from all livelihood activities). The interviewees were asked about the distance of their residence from the OD and whether they belonged to any environmental group. They were asked whether they engaged in the commercial extraction of bush mangoes and whether they owned commercial cocoa farms. Three possible incentives for increasing the local support for nature conservation were presented to the interviewees and they were asked to choose one (as detailed in The choice question).

The choice question

The incentive strategies were developed to describe how to sustain nature conservation as well as livelihoods of local people. The incentive strategies were developed to represent alternative incentive measures that could motivate the local support for conservation in the OD. The potential incentives were described to the interviewees twice, the potential costs and benefits of each incentive were highlighted and they were asked to choose the most preferred incentive. The incentives were as follows:

•	Access to non-timber forest products (NFOR): This incentive involves giving local people access to non-timber forest products in the OD under the supervision of the OD officials. The officials will set a limit regarding when (i.e. seasons) to allow access and the quantity of products that can be extracted. The potential benefit of this incentive to local people is an increase in access to non-timber forest products. Because the extraction of the products will be monitored by experts, it is assumed that the activity will be sustainable (i.e. the risk of overexploitation of non-timber forest product will be reduced). The NFOR incentive is closer to the current incentive strategy practised during the period of this study. The main difference is that the NFOR gives local people access to non-timber forest products in the OD, while the one in practice gives them access to non-timber forest products only in the buffer zone. The NFOR strategy implies an addition to the existing rights.
•	Community forestry (COMFOR): This incentive strategy involves allocating land outside the OD to local people for establishing and managing community forest for their own use. The OD officials will provide local people with advice on sustainable forest management. The COMFOR should give local people more access to forest products and knowledge on sustainable forest management. It is assumed that if local people have access to forest products in the community forest, it would reduce pressure on the resources in the buffer zone. This would help lower the tendency to extend forest products extractive activities to the OD, thus helping achieve the nature conservation goal. The knowledge gained by local people may motivate some of them to plant more trees on their farmlands, thereby helping to increase the forest land area as well as biodiversity in the buffer zone. The COMFOR strategy implies an addition to the existing rights.
•	Skill development (SKILL): This incentive strategy is about linking development with nature conservation such as provision of a vocational training centre to develop local people's skills. The potential benefits from this strategy will include more access to alternate livelihood sources. The SKILL has the potential to give local people access to more livelihood opportunities that depend less on forest resources, which may help increase forest land area and, consequently, biodiversity. The vocational training centre may offer an avenue to enlighten more people about the benefits of nature conservation. The SKILL strategy implies an addition to the existing rights.

The conceptual framework

Although local people have access to non-timber forest products in the buffer zone of the OD and benefit from newly created facilities such as town halls, bridges and classroom blocks (Ezebilo & Mattsson 2010b), conflicts between local people and conservation efforts regarding uses of land persists. It is assumed that if local people are involved in developing an incentive strategy, it will help in identifying incentives that meet their demands, increase the local support for conservation and improve the present situation. The choice experiment can be used to explore the local people's preferences for different incentives. This is based on the Lancasterian consumer theory and random utility theory. The former implies that consumers make choices based on preferences for attributes of the good (Lancaster 1966). Random utility theory utility cannot be observed, but rather that it can be decomposed into a systematic component that depends on a vector of attributes which can be estimated and a random component which cannot be estimated (Hearne & Santos 2005; Greene & Hensher 2009). Choice experiment helps in predicting people's choice by determining the relative importance of various attributes in the choice process (Hanemann & Kanninen 1999). The interviewees were required to choose among a set of incentives; thus, a rational interviewee would choose an incentive that maximizes his or her utility (Deaton & Muellbauer 1980). If the interviewee chooses incentive m, we assume that the utility from that incentive is greater to him or her than the utility from other incentives:

(1)

where U_m is the utility to the interviewee for incentive m, and U_n is the utility to the interviewee for other incentives n. The interviewee's choice can be modelled as maximizing the expected utility from incentive m as:

(2)

where E(U_m ) is the expected utility of incentive m to the interviewee, f_m is a function of the interviewee's attributes (c) and ϵ _m is a disturbance term that is assumed to be independently and identically distributed. Because E(U_m ) is not observable, let A_m be the random variable representing the incentive chosen by the interviewee. It is assumed that the interviewee faces a set of discrete, mutually exclusive choices of incentives and the final choice depends on the interviewee's personal attributes.

Table 1. Variables used in statistical analysis

Variable	Description	Mean	SD	Min	Max
Incentive strategies	Different incentive strategies offered to the interviewees:
	Gathering of non-timber forest products under the supervision of national park officials (NFOR = 0)	0.15
	Community forestry (COMFOR = 1)	0.26
	Skill development (SKILL = 2)	0.59
FARM SIZE	Farmland size in hectares	6.45	6.17	0.50	50.00
DISTANCE	Distance from the interviewee residence to the OD in kilometre	5.46	4.53	0.50	25.00
NONTIMB	Annual income from non-timber forest products in Nigerian Naira	18,217	26,320	5000	250,000
HH_SIZE	Household size (number of persons)	8.59	4.71	2.00	28.00
GENDER	Gender of the interviewee (male = 1, female = 0)	0.55	0.49	0	1
EDUCAT	Interviewee has at least primary school education (yes = 1, no = 0)	0.79	0.41	0	1
MEMBER	Membership of an environmental group (yes = 1, no = 0)	0.22	0.26	0	1
B_MANGO	Extracts bush mangoes for commercial purpose (yes = 1, no = 0)	0.69	0.46	0	1
COCOA	Commercial cocoa farm (yes = 1, no = 0)	0.34	0.48	0	1
Note: 1US $ = 155 Nigerian Naira (NGN), SD = standard deviation.

The econometric model

In this study, three different incentive strategies were presented to the interviewee who was asked to choose the one that he or she preferred most. Although the ordered regression model can be used for analysing the data from this study, the Brant test (Greene & Hensher 2009) revealed that the restrictive assumption (i.e. parallel regression assumption (Harrell 2001)) required for using the ordered model was violated. This implies that the dependent variables are not ordered; thus, the multinomial regression model, which can be used in analysing unordered data, was applied. The Hausman–McFadden test (Hausman & McFadden 1984; Greene 2003) was used to explore the restrictive assumption of the multinomial model known as independence of irrelevant alternatives (IIA). The IIA implies that the ratio of the probabilities of choosing incentive m over other incentives n is independent of the offered choice set (Bjorn & Vuong 1985). To further explore whether multinomial probit is more suitable than the multinomial logit for analysing the data, the Lagrange multiplier test (Greene 2003) was conducted. The hypothesis that the error term is normally distributed was rejected at 1% statistically significant level, which revealed that the assumption required for using multinomial probit was not satisfied. Thus, the multinomial logit model was used in data analysis. Assuming that the error term is independently and identically distributed according to the logistic function, the probability that the interviewee will choose incentive A_m can be modelled (Greene 2003) as:

(3)

where, Pr(.) is the probability that the interviewee prefers the m incentive strategy and β _m represents the parameters to be estimated. Normalization of the alternatives by one of the categories (β _n = 0) yields the multinomial logit model as:

(4)

The multinomial logit model was estimated using the LIMDEP NLOGIT version 4.0.1 econometric software (Econometric Software Inc., New York, NY, USA) and the effects of the interviewees' attributes on preferences for incentives were analysed. The variables that were used in the statistical analyses are presented in Table 1.

Results

All the people selected for the interviews agreed to be interviewed. However, of all the 150 interviewees, only 137 (91%) answered all the questions that were relevant for statistical analysis. Approximately 16% of the interviewees were youths and 84% were adults. According to the Nigerian definition, youth are individuals aged 18–35 years. The interviewees' average annual income from all livelihood activities was 158,554 Nigerian Naira (NGN), i.e. US$ 1023. Approximately 53% of the interviewees were self-employed and engaged in activities such as agriculture, gathering of non-timber forest products and as storekeepers and 47% worked in governmental agencies as teachers, nurses or administrative personnel.

On average, the incentive associated with skills development (SKILL) was most preferred by the interviewees, followed by community forestry (COMFOR) while the incentive associated with the gathering of non-timber forest products under the supervision of national park officials NFOR was the least preferred (Table 1). Most of the youths preferred NFOR compared to adults (Table 2). Also, more female interviewees preferred COMFOR in comparison with their male counterparts, while more male interviewees preferred SKILL. Most of the interviewees in the Butatong and Wula villages preferred SKILL in comparison with those in the Bukalom village. On the other hand, most interviewees in the Bukalom village preferred COMFOR. NFOR was the least preferred among all of the villages.

Table 2. Choice of incentive strategies in relation to age, gender and community

	Age %		Gender %		Community %
Strategy	Youth	Adult	Male	Female	Butatong	Bukalom	Wula
NFOR	21.00	13.50	11.10	18.80	12.00	16.00	16.00
COMFOR	29.00	25.40	18.50	34.80	24.00	32.00	22.00
SKILL	50.00	61.10	70.00	46.40	64.00	52.00	62.00
Notes: NFOR = Gathering of non-timber forest products under the supervision of national park officials.
COMFOR = Community forestry; SKILL = Skill development.

Table 3. Multinomial logit model results for factors influencing choice of incentive strategies

	Community forestry (COMFOR)			Skill development (SKILL)
Variable	Coefficient	Marginal effect		Coefficient	Marginal effect
Constant	−3.54*			−0.01
	(2.15)			(1.16)
FARM SIZE	−0.13	−0.00		−0.11*	−0.01
	(0.09)	[0.99]		(0.06)	[0.89]
DISTANCE	0.01	0.00		0.05	0.01
	(0.14)	[1.01]		(0.12)	[1.05]
NONTIMB	−0.00	0.00		−0.00**	0.00
	(0.00)	[0.98]		(0.00)	[0.99]
HH_SIZE	0.14	0.01		−0.03	−0.01
	(0.14)	[1.15]		(0.09)	[0.97]
GENDER	−2.92**	−0.23		−0.03	−0.19
	(1.39)	[0.05]		(0.75)	[0.97]
EDUCAT	3.31**	0.12		2.07**	0.06
	(1.48)	[27.49]		(0.97)	[7.96]
MEMBER	3.49*	0.26		−0.32	−0.19
	(1.91)	[33.11]		(1.42)	[1.38]
B_MANGO	2.35**	0.05		2.02**	0.12
	(1.13)	[10.49]		(0.85)	[7.53]
COCOA	1.97	0.08		3.46***	0.34
	(1.73)	[7.14]		(1.41)	[31.63]
Log likelihood function			−61.31
Restricted log likelihood			−84.05
χ^2 statistic			45.48
Prob[χ^2 > value]			0.00****
McFadden Pseudo R ^2			0.27
% correctly predicted			79.83
Hausman–McFadden test: χ^2 statistic			6.27
P-value			0.39
Number of observations			137
Notes: Gathering of non-timber forest products under the supervision of national park officials (NFOR) is the omitted choice. *, **, ***, **** represent 0.10, 0.05, 0.01 and 0.001 levels of statistical significance, respectively. Standard error is in parenthesis and odds ratio is in square brackets. Marginal effects for discrete independent variables were calculated as the difference in the expected value of the dependent variable when the independent variable of interest takes the value zero and when it takes the value one while marginal effects for continuous independent variables were calculated at the mean of the independent variable of interest (Greene 2003).

To examine factors that might have influenced the interviewees' preferences for the incentive strategies, a multinomial logit model was estimated (Table 3). The result of the Hausman–McFadden test for IIA assumption was not statistically significant, which implies that the IIA assumption was not violated and the use of multinomial model for data analysis was justified. The result of the log likelihood ratio test was statistically significant, and approximately 80% of the interviewees were correctly predicted to be in the group to which they actually belonged according to the estimated model. This indicates that the model has a good fit. The variance inflation factor of all included explanatory variables did not exceed 1.4, which indicates that multi-colinearity (Chatterjee & Price 1991) was not a serious concern in the estimated model. The multinomial model requires that one of the dependent variables should serve as a base or constant (redundant variable), i.e. ‘no-change’ alternative. The NFOR incentive was used as the redundant variable, because it was the closest to the actual situation during the study period. According to the results of odd ratios and marginal effects (Table 3), the coefficients associated with education and extraction of bush mangoes are the most important variables for predicting the interviewees' choice of the COMFOR and SKILL incentives.

The results (Table 3) show that the interviewees who had at least primary school education, belonged to an environmental group and extracted bush mangoes for commercial purposes were more likely to choose the COMFOR incentive while males were less likely to choose this incentive. The odd ratios and marginal effects revealed that education, extraction of bush mangoes, membership of environmental group and gender are the most important variables for predicting choice of the COMFOR. The coefficients associated with farmland size, distance from the interviewee's home to the OD, income from non-timber forest products, household size and ownership of commercial cocoa farm were not statistically significant for choosing the COMFOR incentive.

Moreover, interviewees who had at least primary school education, extracted bush mangoes for commercial purposes and owned commercial cocoa farms were more likely to choose the SKILL incentive. The interviewees who had larger farmlands and derived more money from non-timber forest products were less likely to choose the SKILL incentive. The odd ratios and marginal effects show that extraction of bush mangoes, education and ownership of commercial cocoa farm are the most important variables for predicting the choice of the SKILL incentive. The coefficients associated with distance from the interviewee's home to the OD, household size, gender and membership of an environmental group were not statistically significant for choosing the SKILL incentive.

Discussion and conclusions

The findings from this study revealed that the current incentives to help promote local support for nature conservation in the Okwangwo Division are not preferred by most local people. For example, the majority of the interviewees preferred incentives that provide them access to skill development and community forestry. This could be a possible reason for the continued conflict of interests between local people and conservation efforts regarding use of land. The concerns raised by members of the focus group regarding the incentive strategies to be included in the study further revealed the importance of involving local communities in designing nature conservation projects that often have impacts on their livelihood activities. This can help in resolving the conflict of interests arising from different competing uses of land such as nature conservation and agriculture. It further highlights that, any incentive strategy that is incompatible with the tradition and values and which does not address the needs of the local people may not be very successful. The results of the study will provide nature conservation managers with a better understanding of how to design a more sustainable participatory approach to nature conservation.

The provision of skill development to local people might provide them with the highest motivation for supporting conservation efforts in the Okwangwo Division. A possible reason for this may be that skill development is new to the local people and they expect more benefits from it than the other incentive strategies. The already existing tension and mistrust between local people and conservation efforts (Ezebilo & Mattsson 2010b) may have influenced the preference for skill development. For example, the incentive that requires frequent contacts with the Okwangwo Division officials, i.e. gathering of non-timber forest products in the Okwangwo Division, was least preferred. Furthermore, local people often see conservation authorities as external forces who have come to dominate them and they do not expect gathering of non-timber forest products in the Okwangwo Division to be sustainable compared with other incentive strategies. This is in line with the social exchange theory (Ekeh 1974), i.e. people develop attitudes towards other people and things with regards to the anticipated benefits and costs that will accrue to them. Thus, local people tend to perceive the incentive that generates more sustainable net benefits positively (Napier & Napier 1991).

People who get benefits from non-timber forest products should be more interested in the incentive that provides increased access to the products. This might be the reason that the interviewees who get more money from non-timber forest products are less likely to choose skill development. The findings are similar to those of the study by Gadd (2005), who found that people have positive attitudes towards nature conservation when they obtain some direct economic benefits from it. People with formal education may receive more benefits from community forestry and skill development, because it may be easier for them to acquire new skills as well as to be involved in managing the community forest. On the other hand, the elite members of the local community are more likely to be educated and they may see community forestry as a means to control and benefit from more resources, as reported by Kellert et al. (2000). This could be the possible reason that the interviewees who had formal education were more likely to choose skill development and community forestry.

One of the objectives of the local environmental groups is to motivate people to plant more trees on their lands as well as to participate in decision-making for nature conservation. This may be a possible reason why the interviewees who belonged to an environmental group were more likely to choose community forestry. This collaborates with the findings by Pretty and Ward (2001). In their study of social capital and environment, they found that the membership of a social group had a positive influence on conservation. Bush mango is one of the major non-timber forest products collected and traded by local people in the study area (Ezebilo & Mattsson 2010a); therefore, interviewees who are involved in bush mango businesses believe that community forestry would provide them with the opportunity to grow more mango trees to boost their business. They may also believe that skill development would help them acquire more knowledge regarding ways to get more benefits from the bush mango business. Thus, they were more likely to choose skill development and community forestry.

In Nigeria as well as in many other African countries, men are often less involved in gathering forest products than women (Bayard et al. 2007). This is in line with the results of the present study. Thus, men are not well aware of the difficulties associated with gathering forest products to help them appreciate community forestry. This stresses the importance of gender issues when designing incentive strategies. Cocoa is one of the cash crops in the study area and is often processed to meet export standards. This implies that the people who own commercial cocoa farms believe that skill development will help them acquire more knowledge about better ways to process cocoa beans to meet the market demand. Surprisingly, the interviewees who had larger farmland were less likely to choose skill development, probably because they believed that the incentive would not provide them quick benefits in short term.

To improve local people's support for nature conservation, one can promote facilities that provide more people with access to formal education and bush mangoes. Education needs emphasis regarding ways of creating awareness and changing people's attitudes towards nature conservation as well as directing people to alternative income-generating activities that might reduce the pressure on forests. People who engage in bush mango business should be trained on sustainable use of forest resources, so that they can conduct their businesses with less negative impacts on forests. The present situation regarding the relationship of local people with conservation efforts can be improved by combining different incentive strategies, because the local communities in the buffer zone are heterogeneous. Thus, each individual, household and group has different values and interests, and therefore introducing a single incentive strategy would benefit only a fraction of the community. The local people who do not benefit from the incentive may frustrate conservation efforts, which may defeat the objectives of the incentive. On the other hand, an incentive strategy that encompasses different kinds of incentives has the potential to benefit more of the local people and motivate them to support nature conservation. For an incentive strategy to be sustainable, the benefits should be equitably distributed among the local people and future access to the incentive should be well guaranteed. Although offering local people direct extraction of non-timber forest products in a nature conservation area could lead to positive attitudes towards conservation, it must be accompanied with strict monitoring and control to prevent forest degradation. The findings should provide nature conservation managers with a better understanding of the general preferences of local communities regarding incentive strategy that might improve the relationship between local people and conservation efforts. This will help in planning and designing a nature conservation strategy to sustain local livelihood activities and nature conservation.

Acknowledgements

I thank all interviewees in Bukalom, Butatong and Wula villages that spent their valuable time for the interviews. I would like to thank Simon Apah, Peter Apah and Kingsly Akam who helped in administering the main and pre-test interviews, respectively. Thanks also go to the editors and anonymous persons who reviewed the manuscript for their useful comments.