Forest Policy and Economics 11 (2009) 516–524

Contents lists available at ScienceDirect

Forest Policy and Economics

j ourna l homepage: www.e lsev ie r.com/ locate / fo rpo l

Determinants of access to forest products in southern Burkina Faso

PascalineCoulibaly-Lingani a,b,⁎,MulualemTigabu a, Patrice Savadogo a,b, Per-ChristerOden a, Jean-MarieOuadba b

a Swedish University of Agricultural Sciences, Faculty of Forest Sciences, Southern Swedish Forest Research Centre, PO Box 101, SE-230 53 Alnarp, Swedenb Centre National de la Recherche Scientifique et Technologique, Institut de l'Environnement et de Recherches Agricoles, Département Productions Forestières,03 BP 7047, Ouagadougou 03, Burkina Faso

⁎ Corresponding author. Swedish University of AgriculSciences, Southern Swedish Forest Research Centre, PSweden. Tel.: +46 40 41 5396; fax: +46 40 41 5398.

E-mail address: Pascaline.Lingani@ess.slu.se (P. Coul

1389-9341/$ – see front matter © 2009 Elsevier B.V. Aldoi:10.1016/j.forpol.2009.06.002

a b s t r a c t

a r t i c l e i n f o

Article history:Received 22 January 2009Received in revised form 18 May 2009Accepted 22 June 2009

Keywords:Forest dependencyGenderForest productsPoverty reductionProperty rights

There is an increasing understanding that forests and the forestry sector are key elements in povertyreduction strategies in Africa. However, issues of equity between various forest users are becoming a majorchallenge to environmental development, forest management and poverty reduction. This paper presents ananalysis of household representatives' socio-economic determinants and other constraints on accessingforest products, based on data collected through a questionnaire survey of 1865 respondents in sevendistricts of the Sissili province, southern Burkina Faso. Three logistic regression models were developed toexamine determinants of access to the forest for collecting fuelwood, grazing livestock and collecting non-timber forest products (NTFPs). The results showed that access to forest products is associated withindividual characteristics. Age, ethnicity, occupation and sources of income were significant determinants ofaccess to all types of forest products. Access to the forest for grazing livestock was further influenced bygender and household size, while access to NTFPs was influenced by gender, household size and educationlevel of the respondents. The formal forest law that precludes grazing in the forest, and customary rules andregulations pertaining to land tenure, were reported to be serious constraints to forest access for women andmigrant people. Understanding the factors influencing access to products from commonly-owned forestresources could form the basis for developing, modifying and targeting policy instruments that promoteequitable access. Policies should particularly encourage the direct involvement of vulnerable and margin-alized groups (women and migrants) in forest management activities.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

There is an increasing understanding of the importance of forestsand the forestry sector as a key element in poverty reductionstrategies in Africa (Oksanen et al., 2003). Over two-thirds of Africa's600 million people obtain a major proportion of their subsistence andsome cash income from a large and diverse set of forest products andforest-related activities (Arnold and Townson, 1998; Kaimowitz,2003; CIFOR, 2005). Plants and animals from natural forests, wood-lands and planted trees provide ecological services and are used byrural Africans for food, energy, medicine, animal feed, construction,furniture, agricultural implements and utensils. These uses of forestproducts allow livelihood diversification (Belem et al., 2007; Kristen-sen and Balslev, 2003; Kristensen and Lykke, 2003; Mamo et al., 2007;Shackleton et al., 2007; Taïta, 2003). Although the importance ofcommon property resources to rural West Africans is understood (Dei,1992; Williams, 1998; Becker, 2001; Beck and Nesmith, 2001), issuesof equity among various forest users is becoming a major challenge to

tural Sciences, Faculty of ForestO Box 101, SE-230 53 Alnarp,

ibaly-Lingani).

l rights reserved.

environmental development, forest management and poverty reduc-tion (Adhikari et al., 2004; Byron and Arnold, 1999; Rocheleau andEdmunds, 1997; Anderson et al., 2006). Thus, an understanding offactors influencing access to common property resources, such asforests, is paramount for identifying the most vulnerable groups andfor adapting natural resource management practices towards a moreholistic and equitable system.

Access is generally defined as “the ability to benefit from things —including material objects, persons, institutions, and symbols” (Ribotand Peluso, 2003). It encompasses both entering into a definedphysical property and obtaining products of a resource, in terms ofaccess and withdrawal sensu Schlager and Ostrom (1992). Severalfactors may influence an individual's ability to derive various productsfrom a given forest, such as gender, household size, education level,age of the household head, total household income and othercontextual factors. In addition, women are often constrained inaccessing and controlling land and forest resources, due to theconstruction of gender identities within households (Agarwal, 1997;Goebel, 1998). Furthermore, larger households clear more forest,because they have more workers and more mouths to feed (Godoyet al., 1997). A higher level of formal schooling is associated with lessforest cutting (Godoy and Contreras, 2001), due to higher opportunitycosts of time (Adhikari et al., 2004), and increased social status and

517P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

economic opportunities in this rural context (Adhikari et al., 2004;Agrawal and Gupta, 2005). The age of the head of household may bepositively related to forest resource utilization until a peak of physicalstrength is reached (Godoy et al., 1997). On the other hand, olderpeople may possess superior knowledge about various forestresources, and may utilize more medicinal plants and wild foods.Owing to improved off-farm employment opportunities (Angelsenand Kaimowitz, 1999) and access to credit (Godoy et al., 1997), totalhousehold income (wealth) may be associated with reduced forestclearance as a supplementary income-generating activity. On theother hand, the ownership of more assets may allow households toexploit more forest resources. Distances to forests and markets arecommon external forces that may often accelerate forest extraction(Angelsen and Kaimowitz,1999; Mamo et al., 2007). In contrast, forestpolicy and traditional rules governing forest usemay impose excessivecontrol over the use of forest resources in some developing countries.

In Burkina Faso, a national forest management program began inthe 1980s with the assistance of joint UNDP/FAO projects (Delnooz,1999; Ribot, 1999; Sawadogo, 2006). In the southern part of thecountry, two forest management domains were set up in the Sissiliand Ziro provinces, with the aim of supplying the main citiesOuagadougou and Koudougou, with fuelwood. Although there havebeen some studies on the socio-economic aspects of these forestmanagement activities (Delnooz, 1999; Hagberg, 2001; Yelkouni,2004), factors that shape access to forest products have not beenthoroughly investigated. Thus, the objectives of this research were toaddress the following questions. What are the key attributes ofrespondents that strongly influence access to the forests in Sissiliprovince for fuelwood collection, grazing livestock and extraction ofNTFPs? Are there other factors, related to formal and informal rulesand regulations, which constrain access to the forests?

2. Materials and methods

2.1. Site description

The studywas conducted in Sissili province (10° 57′ 31″N to 11° 55′38″N and 2° 48′ 47″W to 1° 24′ 13″W), which is located at an altitudeof ca. 300 m above sea level in southern Burkina Faso, West Africa(Fig. 1). The province is divided into seven administrative districts andcovers an area of 7111 km2, with a mean (±SE) annual rainfall of875.6±26.0 mm and a mean annual temperature of 28 °C, based ondata for the years 2000–2003. The total population of Sissili provinceis estimated to be 212628 inhabitants, 51.3% of whom are women(INSD, 2007) and there has been a high immigration rate since the1980s (Henry et al., 2003; Ouédraogo, 2006). The average populationdensity is 30 inhabitants per km2. The population is composed ofindigenous ethnic groups of Nuni, Sissala, Wala and migrant ethnicgroups of Mossi (farmers) and Fulani people (herders) from thecentral plateau and northern Sahelian parts of the country, respec-tively. The farming system is characterized by traditional subsistencecultivation of cereals (such as sorghum, millet and maize), tubers(yam and sweet potatoes), animal husbandry, and a lucrativeproduction system involving extraction of fuelwood and non-timberforest products, cultivation of cash crops (cotton and fruit-treeplantations) and ranching (Paré et al., 2008).

The natural vegetation in Sissili province includes Sissili Stateclassified forest, buffer zone forests bordering the Sissili forest, ForestManagement Units (FMU) and unprotected forests. The present studyfocused on the FMU and unprotected forests, which provide commoninterests for local people in terms of access to the forest for fuelwood,grazing and non-timber forest products. The FMU are blocks of foresthanded to co-operatives by the State for sustainable exploitation ofwood and non-wood products, based on management plans preparedby the State forestry authority and the co-operatives. About 112000 haof forest were intended to be used for forest management activities,

mainly fuelwood exploitation, honey harvesting and biological diversityconservation (Yaméogo Léon, personal communication, January 2008).Unprotected forests constitute bush lands in the vicinity of villages andtrees on fallow fields. These resources are available for use by localpeople, depending on traditional user rights and tenure.

2.2. Data collection

Detailed information on the determinants of access to the forest forcollection of fuelwood, extraction of non-timber forest products(NTFPs) and grazing livestock by local people, and other constraintsrelated to customary regulations and formal forest law were obtainedby conducting a household survey in 2008. The survey was conductedas follows. First, 30% of the villages in each of the seven districts of theprovince were randomly selected, resulting in 45 sample villages.Then a local census was conducted in each sample village to obtain anestimate of the number of households. Following the census, 30% ofthe households in each sample village were randomly selected for thesurvey, yielding a total of 1865 sample households. The respondentswere heads of the household, man or woman. Each respondent wasinterviewed separately and answered the questions on his own behalf.Care was taken to include 50% men and 50% women in the sample ofrespondents. A pre-tested semi-structured questionnaire was used forgathering information, and each interview lasted up to an hour. Theinterviews were conducted by trained field assistants. The checklist ofissues discussed during the interview contained the followingquestions. Do you collect fuelwood and NTFPs from the forestmanagement units and from the unprotected forest in the village?Do you graze your livestock in the forest management units and theunprotected forest in the village? Is there any constraint in accessingforest products? In addition, the following personal information wasrecorded for each respondent: household size, gender, age, residencestatus (indigenous versus migrant), education level, occupation andincome-generating activities.

2.3. Data analysis

In order to assess the significance of socio-economic determinantsof access to the forest for fuelwood collection, extraction of NTFPs andgrazing livestock, three stepwise binary logistic regression modelswere developed, using the forward elimination procedure. The logisticregression model is an appropriate statistical tool to determine theinfluence of explanatory variables on response variables, when thelatter have dichotomous outcomes and the former are continuous,categorical or dummy variables (Agresti, 1996; Peng et al., 2002). Inessence, the logistic model predicts the logit of the response variable(Y) from the explanatory variables (X). The logit is the naturallogarithm (ln) of odds of Y, and odds are ratios of probabilities (π) of Yhappening to probabilities (1−π) of Y not happening. The logisticmodel is specified as:

lnπ

1– π

� �= β0 + β1x1i + β2x2i + : : : + βkxki

where β0 is the intercept and β1, β2 … βk are the coefficients of theindependent variables x1, x2 … xk.

The response variables for the logistic regression applied herewereaccess to the forest for collection of fuelwood, extraction of NTFPs, andgrazing livestock, which were defined as binary variables with a value1 for respondents having access to the forest product or 0 otherwise.The categorization of variables was based on theoretical grounds, aswell as on the basis of the distribution of responses with respect to thevariables listed in Table 1. The model contained eight explanatoryvariables (gender, age, household size, ethnic group, residence status,level of education, source of income and occupation) that wereintroduced simultaneously, and the stepwise regression selected the

518 P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

best combination based on the most significant variables. Dichotomousexplanatory variables (gender and residence status) were coded byassigning 0 to one case and 1 to the other. For multinomial explanatoryvariables, coding was done by assigning 1.0 to the lowest number ofcases and sequentially higher values for others. Before performing the

Fig. 1. Location of t

logistic regression, multivariate correlation analysis was applied tocheck for co-linearity between the explanatory variables. The tolerancevalueswere all above the recommended threshold; therefore therewereno co-linearity problems. The significance of the logistic regressionparameters was assessed by the chi-square likelihood ratio and

he study area.

519P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

deviation test, and Hosmer–Lemeshow's and Wald's statistics. Thesignificance of differences in effects of other customary regulations andformal forest law-related constraints to access forest products related tothe investigated household characteristics were analyzed by χ2-tests(Tabachnick and Fidell, 1996). SPSS 16 software (Copyright SPSS forWindows, Release 2007 Chicago: SPSS Inc.) was used for all of thestatistical analyses.

3. Results

3.1. Profile of the respondents

The numbers of respondents with each set of household referenceperson's characteristics are summarized in Table 1. Respondents werecomposed of equal proportions of men (heads of their respectivehouseholds) and women (whowere either widows and heads of theirhouseholds at the time of the survey, or wives of the head of thehousehold). Most of the respondents were adult (25–45 years old),and the median household size was 6 to 10 family members. The

Table 1Description and summary statistics of the variables used in the binary logistic model.

Characteristics Number ofobservation

Fuelwood NWFP Grazing

Yes No Yes No Yes No

GenderFemale 904 461 469 723 207 512 418Male 908 453 482 725 210 514 421

Age classb25 172 91 84 141 34 105 7025–45 1048 563 499 845 217 627 43545–65 496 212 296 381 127 244 264N65 96 38 58 64 32 40 56

Household sizeV5 461 250 219 337 132 254 2156–10 880 431 472 715 188 510 39311–15 284 132 155 234 53 156 13116–20 96 40 57 78 19 42 55N20 91 54 39 71 22 57 36

Residence statusIndigenous 914 487 427 825 89 568 346Migrant 935 427 524 623 328 458 493

Ethnic groupsNuni 774 410 396 729 77 491 315Mossi 644 279 383 418 244 295 367Sissala 78 74 4 75 3 76 2Dagara 87 78 10 81 7 70 18Fulani 190 58 134 108 84 80 112Minority 39 15 24 37 2 14 25

Educational levelIlliteracy 1340 692 679 287 1084 607 764Primary school 198 102 99 24 177 89 112Secondary school 31 12 20 7 25 18 14Religious education 127 56 74 56 74 67 63Adult education 101 43 58 43 58 45 56Agricultural training 15 4 11 0 15 7 8

OccupationFarmer 598 412 186 543 55 454 144Herder 125 89 36 117 8 94 31Farmer+Herder 1112 409 703 772 340 469 643Farmer+Herder+Woodcutter

30 4 26 16 14 9 21

Source of incomeSNTFP+SCC+SL 842 365 512 684 193 379 498SNTFP 123 95 34 95 34 108 21SCC 201 170 32 173 29 170 32SL 93 50 44 67 27 54 40SNTFP+SCC 217 118 103 206 15 181 40SNTFP+SL 198 59 142 131 70 73 128SCC+SL 130 55 77 86 46 58 74Salary 9 2 7 6 3 3 6

SNTFP: selling of non-timber forest products; SCC: selling of cash crop; SL: selling oflivestock.

respondents were mostly from the following ethnic groups: Nuni(43%), Mossi (36%), Fulani (10%), Dagara (5%), Sissala (4%) and othersmall groups (2%) including Lele, Samo, Wala and Bwaba (hereaftercollectively referred to as the minority group). Among these ethnicgroups, migrant respondents were mainly Mossi, Fulani and Dagara,and they were mostly settled in Biéha (70%) and Niabouri (67%)districts. The indigenous respondents were Nuni and the Sissala. Themajorities of the migrants in the sampled villages live side by sidewith the indigenous people or in migrant quarters. Most of therespondents were illiterate; only 11% of respondents had a primarylevel education. The major source of income of the respondents wasselling of NTFPs, cash crops and livestock; very few respondentsgenerated their income through civil employment with a salary. Themajority of the respondents were farmers, but few of them wereinvolved in livestock husbandry and wood-cutting.

3.2. Determinants of access to forest products

As described above, three logistic regression models were devel-oped to explore the key determinants of the likelihood of access tofuelwood, NTFPs and grazing livestock in forests. Overall assessmentof the logistic regression model for the access to fuelwood, andHosmer–Lemeshow goodness-of-fit statistics, revealed that it pro-vided adequate fit to the data [χ2

(22, 1812)=247.568, pb0.001] with70% correct prediction (Table 2). The likelihood of having access tofuelwood was 8.3% and 3.1% higher for Sissala and Dagara, respec-tively, than for Nuni, Mossi, Fulani and minority ethnic groups(automatic reference group in the model). Respondents in the 25–45 age class had 1.3% more likelihood of access to fuelwood than otherage groups. Respondents engaged in farming or animal husbandry had1.9% and 3.5%more likelihood of access to fuelwood, respectively, thanthose engaged in both activities. In turn, those engaged in bothfarming and animal husbandry had better access than those involvedin combined activities of farming, animal husbandry and wood

Table 2Logistic regression predicting likelihood of reporting access to fuelwood. Note that thereference group in each explanatory variable is not included.

Variables βi s.e.βi

Waldχ2-test

d.f.

Sig. Oddsratio(eβ)

95.0% C.I. forodds ratio

Lower Upper

Age class 11.254 3 0.010b25 −0.021 0.151 0.020 1 0.887 0.979 0.728 1.31625–45 0.281 0.095 8.736 1 0.003 1.324 1.099 1.59445–65 −0.082 0.109 0.574 1 0.449 0.921 0.744 1.140

Ethnic group 55.933 5 0.000Nuni −0.390 0.147 7.040 1 0.008 0.677 0.508 0.903Mossi −0.541 0.151 12.780 1 0.000 0.582 0.433 0.783Sissala 2.112 0.454 21.660 1 0.000 8.261 3.395 20.101Dagara 1.133 0.334 11.506 1 0.001 3.106 1.614 5.979Fulani −1.196 0.220 29.516 1 0.000 0.302 0.196 0.466

Occupation 97.598 3 0.000Farmer 0.659 0.175 14.177 1 0.000 1.933 1.372 2.725Herder 1.244 0.218 32.632 1 0.000 3.468 2.264 5.314Farmer+Herder −0.305 0.162 3.528 1 0.060 0.737 0.536 1.013

Source of income 96.141 7 0.000SNTFP+SCC+SL −0.184 0.178 1.076 1 0.300 0.832 0.587 1.178SNTFP 1.238 0.254 23.701 1 0.000 3.448 2.095 5.674SCC 1.177 0.241 23.748 1 0.000 3.244 2.021 5.207SL 1.065 0.270 15.546 1 0.000 2.902 1.709 4.927SNTFP+SCC −0.283 0.209 1.830 1 0.176 0.754 0.500 1.135SNTFP+SL −0.266 0.220 1.469 1 0.226 0.766 0.498 1.178SCC+SL −0.370 0.237 2.431 1 0.119 0.691 0.434 1.100

Constant 0.129 0.246 0.276 1 0.600 1.138

SNTFP: selling of non-timber forest products; SCC: selling of cash crop; SL: selling oflivestock.Hosmer & Lemeshow Test: Chi-square=12.85, d. f.=8, p=0.117; −2 Loglikelihood=2075.91; Cox & Snell r2=0.21; Nagelkerke r2=0.29; overall percentageof right prediction=70.3%.

Table 4Logistic regression predicting likelihood of reporting access to non-wood forestproduct. Note that the reference group in each explanatory variable is not included.

Variables βi s.e. βi Waldχ2-test

d.f.

Sig. Oddsratio(eβ)

95.0% C.I. forodds ratio

Lower Upper

Gender(F=0, M=1)

0.252 0.072 12.300 1 0.000 1.286 1.117 1.480

Age class 9.238 3 0.026b25 0.140 0.194 0.523 1 0.470 1.151 0.786 1.68325–45 0.309 0.115 7.244 1 0.007 1.362 1.088 1.70545–65 0.012 0.128 0.009 1 0.924 1.012 0.788 1.301

Household size 18.495 4 0.001V5 −0.573 0.143 16.066 1 0.000 0.564 0.426 0.7466–10 −0.030 0.120 0.062 1 0.804 0.971 0.767 1.22911–15 0.247 0.164 2.261 1 0.133 1.280 0.928 1.76516–20 0.167 0.243 0.470 1 0.493 1.181 0.734 1.903

Ethnic group 139.003 5 0.000Nuni 0.326 0.217 21.143 1 0.133 1.385 0.906 2.118Mossi −1.423 0.207 0.068 1 0.00 0.241 0.161 0.362Sissala 1.568 0.532 0.535 1 0.003 4.798 1.692 13.601Dagara 0.303 0.416 16.407 1 0.466 1.354 0.599 3.060Fulani −1.453 0.260 4.291 1 0.000 0.234 0.140 0.390

Education 47.200 5 0.000Illiterate 0.397 0.267 2.213 1 0.137 1.487 0.882 2.508Primary −0.35 0.54 0.42 1 0.517 0.705 0.245 2.03Secondary −0.939 0.226 17.325 1 0.000 0.391 0.251 0.608Religious −1.409 0.263 28.592 1 0.000 0.244 0.146 0.41Adult −0.109 0.444 0.45 1 0.998 0.160 0 .

Occupation 56.535 3 0.000Farmer 0.562 0.196 8.205 1 0.004 1.754 1.194 2.576Herder 1.040 0.333 9.764 1 0.002 2.829 1.473 5.431Farmer+ −0.660 0.163 16.325 1 0.000 0.517 0.375 0.712

520 P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

cutting. Themore diverse the source of income of the respondents, theless was the likelihood of access to fuelwood.

With regard to access to forests for grazing livestock, the modelfitted the data well [χ2(22, 1812)=211.407, pb0.001] with 73% correctprediction (Table 3). Household representative's characteristics thatsignificantly determined access to the forest for grazing livestock,according to Wald statistics, were ethnic group, household size, age,gender, occupation and source of income (Table 3). The Sissala ethnicgroup was more likely (15.3%) to have access to the forest for livestockgrazing than other ethnic groups. However, the Mossi and Fulanigroups had a lower likelihood of access to the forest for grazing thanthe minority groups (the reference in the model). Respondents withsmall (V5 individuals) and very large (N20 individuals) householdswere less likely to have access to the forest for livestock grazing thanmedium-sized households (having 1.1% more chance). Adults (45–65 years old) heads of households also had more chance of forestaccess for grazing livestock. Men had 1.2% more chance to access theforest for grazing livestock than women. The more diversified theoccupation and source of income of the respondent, the lower was thelikelihood of access to the forest resource for grazing livestock.

The regression model for access to NTFPs was significant overall[(χ2

(22, 1812)=336.229, pb0.001)] and the Hosmer–Lemeshow good-ness-of-fit test showed an adequate fit of the model to the data, with82% correct prediction (Table 4). Key determinants of access to NTFPswere ethnic group, education level, household size, age, gender,occupation and source of income (Table 4). Among ethnic groups,Mossi and Fulani were less likely to have access, while Nuni, Sissalaand Dagaraweremore likely to have access to NTFPs than theminority

HerderSource ofincome

45.949 7 0.000

SNTFP+SCC+SL

0.477 0.162 8.677 1 0.003 1.611 1.173 2.212

SNTFP −0.621 0.243 6.537 1 0.011 0.537 0.334 0.865SCC −0.096 0.244 0.154 1 0.695 0.909 0.563 1.467SL 0.703 0.269 6.807 1 0.009 2.020 1.191 3.425SNTFP+SCC 1.106 0.283 15.296 1 0.000 3.022 1.736 5.260SNTFP+SL 0.284 0.205 1.926 1 0.165 1.328 0.890 1.984SCC+SL −0.526 0.224 5.505 1 0.019 0.591 0.381 0.917

Constant 4.744 1657.839 0.000 1 0.998 114.856

SNTFP: selling of non-timber forest products; SCC: selling of cash crop; SL: selling oflivestock.Hosmer & Lemeshow Test: Chi-square=21.96, d. f.=8, p=0.501; −2 Loglikelihood=1488.68; Cox & Snell r2=0.22; Nagelkerke r2=0.33; overall percentageof right prediction=81.8%.

Table 3Logistic regression predicting likelihood of reporting access to forest land for livestockgrazing. Note that the reference group in each explanatory variable is not included.

Variables βi s.e.βi

Waldχ2-test

d.f.

Sig. Oddsratio(eβ)

95.0% C.I. forodds ratio

Lower Upper

Gender(F=0, M=1)

0.180 0.060 8.934 1 0.003 1.197 1.064 1.347

Age class 13.637 3 0.003b25 0.152 0.161 0.887 1 0.346 1.164 0.849 1.59525–45 0.326 0.099 10.795 1 0.001 1.385 1.140 1.68245–65 −0.045 0.113 0.161 1 0.689 0.956 0.766 1.193

Household size 11.476 4 0.015V5 −0.255 0.122 4.376 1 0.036 0.775 0.611 0.9846–10 0.018 0.100 0.032 1 0.858 1.018 0.837 1.23811–15 0.073 0.131 0.307 1 0.580 1.075 0.831 1.39116–20 −0.375 0.195 3.683 1 0.055 0.687 0.469 1.008

Ethnic group 39.321 5 0.000Nuni −0.075 0.169 0.197 1 0.657 0.928 0.665 1.293Mossi −0.467 0.172 7.323 1 0.007 0.627 0.447 0.879Sissala 2.731 0.619 19.431 1 0.000 15.345 4.557 51.674Dagara −0.122 0.296 0.171 1 0.679 0.885 0.495 1.581Fulani −0.576 0.230 6.264 1 0.012 0.562 0.358 0.883

Occupation 90.357 3 0.000Farmer 0.572 0.157 13.194 1 0.000 1.771 1.301 2.411Herder 1.028 0.209 24.160 1 0.000 2.796 1.856 4.213Farmer+Herder

−0.501 0.140 12.786 1 0.000 0.606 0.460 0.797

Source of income 130.866 7 0.000SNTFP+SCC+SL

−0.610 0.149 16.670 1 0.000 0.544 0.406 0.728

SNTFP 1.378 0.272 25.601 1 0.000 3.967 2.326 6.765SCC 0.843 0.226 13.935 1 0.000 2.324 1.493 3.620SL 0.429 0.247 3.029 1 0.082 1.536 0.947 2.490SNTFP+SCC 0.892 0.208 18.401 1 0.000 2.441 1.624 3.669SNTFP+SL −0.604 0.192 9.875 1 0.002 0.547 0.375 0.797SCC+SL −0.747 0.213 12.343 1 0.000 0.474 0.312 0.719

Constant −0.598 0.230 6.727 1 0.009 1.818

SNTFP: selling of non-timber forest products; SCC: selling of cash crop; SL: selling oflivestock.Hosmer & Lemeshow Test: Chi-square=9.22, d. f.=8, p=0.324; −2 Loglikelihood=1990.64; Cox & Snell r2=0.24; Nagelkerke r2=0.32; overall percentageof right prediction=72.5%.

group. Compared with respondents who had received agriculturaltraining, the likelihood of access to NTFPs was lower for all otherrespondents of all education levels, except illiterate individuals whohad similar probability of access. The larger the household size of therespondent, the higher was the likelihood of access to NTFPs.Respondents within the most active age band (24–45 years) had abetter chance of access to NTFPs compared with older respondents. Inaddition, men heads of households were more likely to have access toNTFPs than women heads of households. The likelihood of access toNTFPs was lower for respondents engaged in more than one activity,and for respondents who generated their income through sale ofNTFPs, sale of cash crops or sale of cash crops and livestock.Interestingly, the likelihood of access to NTFPs was high whenrespondents sold NTFPs and other agricultural products, such ascash crops and livestock, as shown by the higher odds ratio (Table 4).

3.3. Constraints for accessing forest products

The respondents stated that access to forest products is essentialfor their livelihoods. However, 6% of the rural respondents reportedthat they had difficulty in obtaining NTFPs and fuelwood. The majorconstraints to accessing forest products in Sissili province were the

521P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

forest law and traditional rules of land and resource tenure.Respondents considered that there were ambiguities in the rights ofcontrol and ownership of the forested lands and the resources theyoffer. The forests are located on lands that are characterized by theoverlap of multiple traditional and modern rights. Prevailing land lawdoes not recognize customary land rights; all land is considered Stateproperty, and anyone seeking access to land must apply for use rights.On the other hand, local communities do not recognize this Stateownership and regard themselves as the true owners of their land, byvirtue of their ancestral rights. While the State's monopoly of landownership is theoretical, the traditional rules are still observed by thelocal communities under the leadership of indigenous ethnic groupelders. The forest law, ‘Code Forestier’ adopted in 1997, specifies theaccess rights of users and the guiding principles for usufructuaryrights and practices. This law precludes livestock grazing frommanaged forest, thereby limiting access to forest pastures by livestockherders. Respondentsmentioned that the lack of official recognition oftraditional forest access rights has contributed to tenure instability. Ithas also encouraged a short-term mentality in the use of forestresources, for example through illegal harvesting of wood, non-woodforest products and charcoal processing, which compromises thesustainability of forest resources.

Respondents' perceptions of traditional rules and forest law, asconstraints to accessing forest products, varied significantly withrespect to their gender, residence status, level of education andoccupation (Fig. 2). The forest law was considered more restrictive bymen than by women, while women felt that the traditional rules weremore restrictive. Both the forest law and the traditional rules wereperceived as more restrictive for access by migrants than indigenouspeople. The uneducated respondents also felt these constraints morethan educated groups. Farmer-herders felt the forest law wasrestrictive, while both farmers and farmer-herders perceived thetraditional rules as restricting their access to forest products.

4. Discussion

The study revealed that respondents had differing constraints onaccess to forest products. Access to fuelwood appeared to be lesssensitive to gender than access to livestock pastures, while access toNTFPs was less likely for women than men. In Sissili province,fuelwood extraction is one of the main activities of forest manage-ment, which is done by wood cutters, who are mainly men, organizedin cooperatives. Women are much less involved in these cooperatives,but they are mainly responsible for gathering fuelwood for householdenergy consumption rather than selling, as men do. It should be notedthat in this study, access to fuelwood concerned collecting fuelwoodfor household consumption. Women and men had equal chance toaccess fuelwood for household energy consumption.

In Sissili province, men and women were equally involved inharvesting NTFPs. However, the processing of some NTFPs, such asfruits of Vitellaria paradoxa C.F.Gaertn (shea nut) and Parkia biglobosaBenth. (locust bean), is mainly a female activity. V. paradoxa and P.biglobosa are the main indigenous tree species in the forest, bush andfallow lands that provide food directly, in the form of fruits, pulp,seeds and other edible parts (Lamien et al., 2007). Shea nuts aretransformed into butter, while the seeds of locust bean are fermentedto make a spice (soumbala), used for consumption or sale in local,national and international markets (Teklehaimanot, 2004). In addi-tion, fruits and flowers from Tamarindus indica L., Bombax costatumPellegr. & Vuillet, Detarium microcarpum Guill. & Perr. and honey aregathered by the respondents. Although women in rural householdsare major gatherers of NTFPs for consumption or sale elsewhere(Agarwal, 1997; Tiani et al., 2005), the likelihood of access to NTFPswas 1.2% lower for women than men. This is because men haveexclusive control over some NTFPs, notably the fruits of P. biglobosa inthe forest, which puts women at a disadvantage. The ability of men to

benefit from NTFPs extraction comes from customary rights. Withregard to livestock grazing, women had 1.3% less chance than men toaccess the forest for grazing. This could be because many women-headed Burkinabé households do not possess a large herd of cattle.Those with few animals use concentrate feeds to fatten them; thusthey are less likely to go to the forest for free grazing of their livestock.

Access to forest resources may be positively related to the age ofthe head of the household, as long as his or her physical strength issufficient (Godoy et al., 1997). However, Mamo et al. (2007) foundthat age of the respondent was negatively related to household percapita income and relative forest income. Our results show that theage-forest access relationship depends on the type of forest productsthe respondents are gathering. For example, access to NTFPs and ageclass showed a positive relationship, while access to fuelwood andlivestock grazing was significantly higher only for the active age group(25–45 years) compared with other age groups. Older individualshave difficulty performing laborious agricultural tasks andmay turn toresource collection activities that demand less physical labor, such asharvesting NTFPs (Cavendish, 2000). Respondents with larger house-holds (11–20 family members per household, in our case) had higherlikelihoods of access to NTFPs, whereas those with medium to largehouseholds had a better chance of access to the forest for livestockgrazing. These findings support the view that fairly large householdsize may be the motivation for dependency on forest products (Godoyet al., 1997; Jha, 2008). Individuals from larger families may find itdifficult to access alternative sources of subsistence, and thus becomedependent on forest resources. With very large family size (N20 in ourcase), the dependency on forest resources decreased, most likely dueto a low worker–consumer ratio of the household. Our results areconsistent with those of a study in Ethiopia, where dependency on theforest was found to be positively related with household size (Mamoet al., 2007).

Our study shows that the ethnicity, but not residence status, of therespondents significantly influenced access to forest products. How-ever, it could be that residence status is embedded in the ethniccomposition of the villages. Generally, migrants appeared to havelower access to forest products than indigenous people. Migrantscultivate land acquired by gift or by informal renting agreements, butthe owner of the land may reserve the right to access some treeproducts, such as fruits of P. biglobosa and V. paradoxa (Boffa, 1999;Paré, 2001). This is supported by the logistic regression analysis,which indicated that the indigenous ethnic groups (Nuni and Sissala)have a better chance of access to NTFPs than themigrant ethnic groups(Mossi and Fulani). These, in turn, had a significantly lower chance ofaccess than the minority group (Lele, Samo, Wala and Bwaba). TheFulani and Mossi groups also had less chance of access to fuelwood,while these and the Dagara migrant ethnic group had less chance ofaccess to grazing pastures than the indigenous Sissala. While theindigenous Nuni group had less chance of access to the forest forfuelwood and grazing livestock, themigrant Dagara group had a betterchance of access to fuelwood and NTFPs. Generally, access to land andtenure is governed by a complex system of delegation of rights,involving ethnicity, inheritance and various sorts of compensations forexchange of land (Paré, 2001). Some migrant ethnic groups who haveresided in the village for a long time may be accepted and integratedinto the traditional system. This finding is similar to findings by DeGroote and Coulibaly (1998), who concluded that ethnic group is amajor determinant of access to forest resources.

In previous studies education has been shown to reduce forestdependency (Gunatilake, 1998; Mamo et al., 2007). A higher level ofeducation provides awider range of employment opportunities, whileparticularly making fuelwood collection unprofitable due to thehigher opportunity costs of collection (Adhikari et al., 2004). Ourresults indicate that education level did not influence respondents'access to the forest for fuelwood extraction and grazing livestock, butit substantially influenced individuals' access to NTFPs. Compared

Fig. 2. Countof the respondentperconstraintmentioned (FL: forest law,TR: traditional rules) togetherwith the results of theChi-squareanalyses (F=Farmer,H=Herders,W=Woodcutter,SNTFP = selling of non-timber forest products; SCC = selling of cash crop; SL = selling of livestock).

522 P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

523P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

with respondents who had received agricultural training (thereference category in the model), respondents with formal schoolingand adult education had limited access to NTFPs. It could be that theagricultural extension service promotes collection and processing ofNTFPs. For those with formal education, lower access may also beindicative of higher opportunity costs of time for collecting andprocessing NTFPs. Religious influence may also be a contributoryfactor to reduced access to NTFPs. Roots, herbs, leaves and animalsused in traditional medicine are important in the healthcare deliverysystem in the area, and religious bonds have been found to impact thechoices made by individuals seeking healthcare services (Tabi et al.,2006). Traditional medicine consists of health practices, approaches,knowledge and beliefs incorporating animal- or plant-based medi-cines and spiritual therapies, used either singularly or in combination,to treat, diagnose and prevent illness and to maintain well-being(WHO, 2002).

Respondents engaged in either farming or livestock husbandry,thus generating their income from selling cash crops or cattle, aremore likely to access the forest for fuelwood collection and grazinglivestock. In the study area, farmers cultivating cash crops were betteroff than farmers engaged in subsistence farming activities, whodiversify their income through off-farm activities, such as sellingNTFPs or wood. Several studies have shown that higher incomes,within relevant range of income in developing countries, are likely toincrease the dependency on forests (Angelsen and Kaimowitz, 1999;Escobal and Aldana, 2003; Adhikari et al., 2004; Mamo et al., 2007).Thus, this finding is consistent with several previous studies showingthat increases in income, within the relevant range of incomes indeveloping countries, are associated with increases in dependency onforests (Angelsen and Kaimowitz, 1999; Escobal and Aldana, 2003;Adhikari et al., 2004; Mamo et al., 2007). It is generally believed thatownership of more assets allows households to exploit more forestresources. Another interesting result is that respondents rely onselling of NTFPs to diversify their income. This implies that forestproducts play an important role in equalizing incomes of ruralhouseholds, as well as in sustaining the livelihood of the rural poorand in managing risks associated with weather, crop losses and otherunpredictable events.

Apart from household representative's characteristics, forest lawand traditional rules and regulations were identified as majorconstraints to accessing forest products in the study area. For mostrespondents, gender and residence status were perceived as factorsconstraining access. Women found themselves to be doubly dis-advantaged with respect to access to forest resources, because landinheritance practices and procedures for formalizing land rights oftendiscriminate against them. Traditionally, land rights were held by menor kinship groups controlled by men, and women did not have directrights over land. The prohibition of planting trees on rented lands is atraditional rule in force in southern Burkina Faso. Also, local customdictates that only those who have family rights to land (indigenouspeople) can plant trees. This is because planting trees on rented landhas been perceived as powerful evidence for land claims. Conse-quently, women and migrants are forbidden to plant trees, to preventthem from subsequently claiming the land (Fortmann and Bruce,1988). A recent study conducted in south-western Burkina Faso hasshown differences in numbers of trees on farm lands betweenmigrantand indigenous groups, due to differences in cultivated area (Ræbildet al., 2007), indicating a limitation on migrants. Furthermore,according to the customary rules, people have access only towoodlands that belong to their own lineages or families. Therefore,migrants are not always allowed to harvest valuable products, such asseeds of P. biglobosa, because of their great socio-economic impor-tance. This finding is consistent with a report by Campbell et al.(2001), in which it is argued that traditional customs and norms takeprecedence over the formal rule-based systems in determining theuse of common property resources, such as forests. The forest law

prohibits grazing in the forest, forbids animals being left overnight inthe forest, and only allows regulated cutting of wood. Herders shouldpay a fine of 0.15 Euro per animal to the management fund, if theircattle are found grazing in the forest (Hagberg, 1998). Since men aremainly responsible for cattle herding and fuelwood extraction andselling, they regard this as restrictive policy. The designation of sacredforests in some of the villages also prevents access to forest productsin some places. Sacred forests are sanctuaries where the communityelders are buried, places where sacrifices to ancestors are made, andother traditional rituals performed, and only indigenous married menare allowed to enter. The cultural values of these sacred forests mayrestrain livestock herders from deliberately setting fire to the forest toobtain new grass shoots for their cattle.

5. Conclusion

The results suggest that there are various types of forest users, withdifferent socio-economic status, perspectives, understanding andobjectives. Rural villages in Burkina Faso are not homogeneousentities that can be isolated and identified by a single objective orcommon interest. Thus, policies based upon assumptions of villagehomogeneity cannot deliver equitable access to common propertyresources. Since patterns of access to forest differed among respon-dents, local management institutions need to take these variationsinto account to ensure that the livelihood needs of the poor are met. Acoherent balance of formal and traditional tenure rights governingforest use appears to be lacking. More attention should be paid tosecuring user rights and to empowering forest users, by allowingdirect involvement of vulnerable and marginalized groups (womenand migrants) in forest management activities. The current forest lawfor managing forest resources is very restrictive, so we suggest thatcontrol over access to the forests should bemodified. In this context, itshould be noted that several recent studies in Burkina Faso havedemonstrated that a moderate level of grazing is not detrimental tothe establishment and growth of herbaceous (Savadogo et al., 2008)and woody species (Zida et al., 2009).

Acknowledgements

Funding for this study was provided by the Swedish InternationalDevelopment Cooperation Agency (Sida). We are grateful to the fieldassistants for their invaluable help in carrying out the survey.

References

Adhikari, B., Di Falco, S., Lovett, J.C., 2004. Household characteristics and forestdependency: evidence from common property forest management in Nepal.Ecological Economics 48, 245–257.

Agarwal, B., 1997. Environmental action, gender equity and women's participation.Development and Change 28, 1–44.

Agrawal, A., Gupta, K., 2005. Decentralization and participation: the governance ofcommon pool resources in Nepal's Terai. World Development 33, 1101–1114.

Agresti, A., 1996. An introduction to categorical data analysis. John Wiley & Sons, Inc.,New York. 290 pp.

Anderson, J., Benjamin, C., Campbell, B., Tiveau, D., 2006. Forests, poverty and equity inAfrica: new perspectives on policy and practice. International Forestry Review, 8(1), 44–53.

Angelsen, A., Kaimowitz, D., 1999. Rethinking the causes of deforestation: lessons fromeconomic models. The World Bank Research Observer, 14, 73–98.

Arnold, M., Townson, I., 1998. Assessing the potential of forest product activities tocontribute to rural incomes in Africa. ODI Natural Resource Perspectives 37(November).

Beck, T., Nesmith, C., 2001. Building on poor people's capacities: the case of commonproperty resources in India and West Africa. World Development, 29 (1), 119–133.

Becker, L.C., 2001. Seeing green in Mali's woods: colonial legacy, forest use and localcontrol. Annals of the Association of American Geographers, 91 (3), 504–526.

Belem, B., Nacoulma, B.M.I., Gbangou, R., Kambou, S., Hansen, H.H., Gausset, Q., Lund, S.,Raebild, A., Lompo, D., Ouedraogo, M., et al., 2007. Use of non wood forest productsby local people bordering the “Parc National Kabore Tambi”, Burkina Faso. TheJournal of Transdisciplinary Environmental Studies 6, 21.

Boffa, J.-M., 1999. Agroforestry parklands in sub-Saharan Africa. FAO ConservationGuide, vol. 34, p. 230.

524 P. Coulibaly-Lingani et al. / Forest Policy and Economics 11 (2009) 516–524

Byron, N., Arnold, M., 1999. What futures for the people of the tropical forests? WorldDevelopment 27, 789–805.

Campbell, B., Mandondo, A., Nemarundwe, N., Sithole, B., De Jong, W., Luckert, M.,Matose, F., 2001. Challenges to proponents of common property resources systems:despairing voices from the social forests of Zimbabwe. World Development, 29,589–600.

Cavendish, W., 2000. Empirical regularities in the poverty–environment relationship ofrural households: evidence from Zimbabwe. World Development 28, 1979–2003.

CIFOR, 2005. Contributing to Africa's Development through Forests Strategy forEngagement in sub-Saharan Africa. Centre for International Forestry Research,Bogor, Indonesia. 35 pp.

Dei, G., 1992. A Ghanaian rural community: indigenous responses to seasonal foodsupply cycles and the socio-environmental stresses of the 1980s. In: Fraser Taylor,D.R., Mackenzie, F. (Eds.), Development from within: survival in rural Africa.Routledge, London, pp. 58–81.

De Groote, H., Coulibaly, N., 1998. Gender and generation: an intra-household analysison access to resources in southern Mali. African Crop Science Journal 6, 79–95.

Delnooz, P. (1999) Gestion des ressources forestières: la communauté, l'État et lemarché. Etude des projets d'aménagement au Burkina Faso. Thèse de Doctorat enSciences de l'Environnement, Fondation Universitaire Luxembourgeoise. Tome 1,211 p., Tome 2, 545 pp.

Escobal, J., Aldana, U., 2003. Are non-timber forest products the antidote to rainforestdegradation? Brazil nut extraction in Madre De Dios, Peru. World Development, 31,1873–1887.

Fortmann, L., Bruce, J.W., 1988. Whose Trees?: Proprietary Dimensions of ForestryWestview, Boulder (USA).

Godoy, R., Contreras, M., 2001. A comparative study of education and tropicaldeforestation among lowland Bolivian Amerindians: forest values, environmentalexternality, and school subsidies. Economic Development and Cultural Change, 49,555–574.

Godoy, R., O'Neill, K., Groff, S., Kostishack, P., Cubas, A., Demmer, J., McSweeney, K.,Overman, J., Wilkie, D., Brokaw, N., Martinez, M., 1997. Household determinants ofdeforestation by Amerindians in Honduras. World Development, 25, 977–987.

Goebel, A., 1998. Process, perception and power: notes from “participatory” research inZimbabwean in a resettlement area. Development and Change 29, 277–305.

Gunatilake, H.M., 1998. The role of rural development in protecting tropicalrainforests: evidence from Sri Lanka. Journal of Environmental Management 53,273–292.

Hagberg, S., 1998. Money and meaning in forest management: a case-study fromBurkina Faso. In: Berger, A. (Ed.), Twice Humanity — Implications for Local andGlobal Resource Use. The Nordic Africa Institute and Forum for DevelopmentStudies, Uppsala, pp. 149–161.

Hagberg, S., 2001. In search of “nyo”: Lyela farmers' perceptions of the forest in BurkinaFaso. Africa: Journal of the International African Institute 71, 481–501.

Henry, S., Boyle, P., Lambin, E.F., 2003. Modeling inter-provincial migration in BurkinaFaso, West Africa: the role of socio-demographic and environmental factors.Applied Geography 23, 115–136.

INSD. (2007). Résultats Préliminaires du Recensement Général de la Population et del'Habitation de 2006. 51 p.

Jha, S. (2008). Household-specific variables and forest dependency in an Indian hotspotof biodiversity. Challenges for sustainable livelihoods. Environment, Developmentand Sustainability. doi:10.1007/s10668-008-9175-y.

Kaimowitz, D., 2003. Not by bread alone...forests and rural livelihoods in Sub-SaharanAfrica. In: Oksanen, T., Pajari, B., Toumasjukka, T. (Eds.), Forests in PovertyReduction Strategies: Capturing the Potential. European Forest Institute, Joensuu,Finland, pp. 45–63.

Kristensen, M., Balslev, H., 2003. Perceptions, use and availability of woodyplants among the Gourounsi in Burkina Faso. Biodiversity and Conservation 12,1715–1739.

Kristensen, M., Lykke, A.M., 2003. Informant-based valuation of use and conservationpreferences of savanna trees in Burkina Faso. Economic Botany 52, 203–217.

Lamien, N., Tigabu, M, Sita, G., Odén, P.C., 2007. Variations in dendrometric and fruitingcharacters of Vitellaria paradoxa populations and multivariate models for estima-tion of fruit yield. Agroforestry Systems 69, 1–11.

Mamo, G., Sjaastad, E., Vedeld, P., 2007. Economic dependence on forest resources: acase from Dendi District, Ethiopia. Forest Policy and Economics 9, 916–927.

Oksanen, T., Pajari, B., Toumasjukka, T., 2003. Forests in poverty reduction strategies:capturing the potential, executive summary. EFI proceedings No. 47. EuropeanForest Institute, Joensuu, Finland.

Ouédraogo, I., 2006. Land use dynamics in Bieha, Sissili Province, Southern Burkina Faso,West Africa. Umoja: Bulletin of the African and African American Studies 1, 18–34.

Paré, L., 2001. Negotiating rights: access to land in the cotton zone, Burkina Faso. LandTenure and Resource Access in West Africa. IIED, London. 28 pp.

Paré, S., Söderberg, U., Sandewall, M., Ouadba, J.M., 2008. Land use analysis from spatialand field data capture in southern Burkina Faso, West Africa. Agriculture,Ecosystems & Environment 127, 277–285.

Peng, C.Y.J., Lee, K.L., Ingersoll, G.M., 2002. An introduction to logistic regression analysisand reporting. The Journal of Education Research, 96 (1), 3–14.

Ræbild, a., Hansen, H.H., Dartell, J., Ky, J.-M.K., Sanou, L., 2007. Ethnicity, land use andwoody vegetation: a case study from south-western Burkina Faso. AgroforestrySystems 70, 157–167.

Ribot, J.C., 1999. Decentralization, participation and accountability in Sahelian forestry:Legal instruments of political-administrative control. Africa 69, 23–65.

Ribot, J.C., Peluso, N.L., 2003. A theory of access. Rural Sociology 68 (2), 1–40.Rocheleau, D., Edmunds, D., 1997. Women, men and trees: gender, power and property

in forest and agrarian landscapes. World Development 25, 1351–1371.Sawadogo, L., 2006. Adapter les approches de l'aménagement durable des forêts sèches

aux aptitudes sociales, économiques et technologiques en Afrique. Le cas duBurkina Faso. Center for International Forestry Research, Jakarta, Indonesia. p. 59.

Savadogo, P., Tiveau, D., Sawadogo, L., Tigabu, M., 2008. Herbaceous species responses tolong-term effects of prescribed fire, grazing and selective tree cutting in thesavanna-woodlands of West Africa. Perspectives in Plant Ecology, Evolution andSystematics 10, 179–195.

Schlager, E., Ostrom, E., 1992. Property-rights regimes and natural resources: aconceptual analysis. Land Economics 68 (3), 249–262.

Shackleton, C.M., Shackleton, S.E., Buiten, E., Bird, N., 2007. The importance of drywoodlands and forests in rural livelihoods and poverty alleviation in South Africa.Forest Policy and Economics 9, 558–577.

Tabachnick, B.G., Fidell, L.S., 1996. Using multivariate statistics. Harper Collins CollegePublishers, New York. 980 p.

Tabi, M.M., Powell, M., Hodnicki, D., 2006. Use of traditional healers and modernmedicine in Ghana. International Nursing Review 53, 52–58.

Taïta, P., 2003. Use of woody plants by locals in Mare aux Hippopotamus BiosphereReserve in western Burkina Faso. Biodiversity and Conservation 12, 1205–1217.

Teklehaimanot, Z., 2004. Exploiting the potential of indigenous agroforestry trees:Parkia biglobosa and Vitellaria paradoxa in sub-Saharan Africa. AgroforestrySystems, 61, 207–220.

Tiani, A.M., Akwah, G., Nguiebouri, J., 2005. Women in Campo-Ma' an National Park.uncertainties and adaptations in Cameroon. In: Carol, J., Pierce, C. (Eds.), TheEquitable Forest: Diversity, Community & Resource Management, pp. 131–149.

Williams, T., 1998. Multiple uses of common pool resources in semi-arid West Africa: asurvey of existing practices and options for sustainable resource management.Natural Resources perspectives, vol. 39. Overseas Development Institute, London.

World Health Organization (WHO), 2002. Traditional Medicine strategy 2002–2005.WHO, Geneva. 61 pp.

Yelkouni, M. (2004). Gestion d'une ressource naturelle et action collective: le cas de laforêt de Tiogo au Burkina Faso. Thèse de Doctorat en science économique.Université d'Aubergne - Clerrmont 1. Ecole Doctorale de Science Economique et deGestion. 330 p.

Zida, D., Tigabu,M., Sawadogo, L., Tiveau, D., Odén, P.C., 2006. Long-term effects of prescribedearly fire, grazing and selective tree cutting on seedling populations in the Sudaniansavanna of Burkina Faso. African Journal of Ecology 47, 97–108.