Local Communities' Attitudes and Support Towards the Kgalagadi Transfrontier Park in Southwest Botswana - MDPI
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
sustainability Article Local Communities’ Attitudes and Support Towards the Kgalagadi Transfrontier Park in Southwest Botswana Naomi Moswete 1 , Brijesh Thapa 2, * and William K. Darley 3 1 Department of Environmental Science, University of Botswana, Gaborone 00704, Botswana; MOATSHEN@mopipi.ub.bw 2 Department of Tourism, Hospitality and Event Management, University of Florida, Gainesville, FL 32611, USA 3 College of Business and Innovation, University of Toledo, Toledo, OH 43606, USA; wd4361@gmail.com * Correspondence: bthapa@hhp.ufl.edu; Tel.: +1-352-294-1656 Received: 16 January 2020; Accepted: 12 February 2020; Published: 18 February 2020 Abstract: Protected areas are of national importance and have developed into sources of benefits while in other situations have sparked conflicts among stakeholders, including residents from adjacent local communities, and park authorities. In this study, we examined community residents’ attitudes towards the Kgalagadi Transfrontier Park (KTP) in the Kalahari region (SW Botswana). This study assessed factors that influence support for, or opposition to, the KTP. A questionnaire with semi-structured questions was used to gather information from head of households (N = 746) in nine villages in the Kalahari region. Overall, positive attitudes and support for the KTP as a transfrontier park were documented, though tangible benefits were limited. Further based on analyses, literacy, proximity, and employment status were key variables that influenced support. In addition, any increase in residents’ perceived benefits, land ownership, conservation awareness, and local benefits resulted in increased support for KTP. The implications indicated that communities near the KTP (Botswana side) need to be consulted, while further communications between the KTP management and authorities and adjacent villages are required to initiate effective community conservation programs. Additional programs and community outreach initiatives would also enable positive attitudes and support of KTP. Keywords: conservation; Transboundary Park; residents; attitudes; Kalahari; southern Africa 1. Introduction Parks and protected areas are generally associated with benefits and related values (monetary, pride) by local residents due to improved quality of the environment, and other social and economic benefits including employment [1–10]. However, local residents’ attitudes towards protected areas in the developing world have been mixed [1,3,11–17]. Research has identified various factors that influence negative attitudes such as, human-wildlife conflict [14,18–22], land claims [23–25], restrictive policies and access regulations to collect non-timber forest products (e.g., nuts, wild mushrooms, berries, seeds, medicinal plants and herbs, etc.), and livestock grazing [18,24,26–30]. Conversely, positive attitudes have been influenced by community and personal economic benefits largely derived from tourism [4,22,31–37]. As evident, there are challenges faced by stakeholders (i.e., local residents, resource managers, park authorities, tourism planners, developers, and conservation organizations) with respect to the balance of conservation priorities and livelihood needs [7,38–42]. In general, rural people in developing countries experience hardships and have had difficulties with minimal resources due to climate change, limited agricultural production as a result of Sustainability 2020, 12, 1524; doi:10.3390/su12041524 www.mdpi.com/journal/sustainability
Sustainability 2020, 12, 1524 2 of 17 unreliable rainfall and recurring droughts, and population growth in villages flanking protected areas [6,10,13,21,26,34,43–47]. In southern Africa, livelihood activities with sole dependence on forest and rangeland resources have caused, and in some instances, exacerbated soil and land degradation [48,49]. Subsequently, instances of conflicts over natural resource use between different stakeholders that include park authorities and adjacent local communities [15,24,50–52] have also fueled unsustainable livelihood activities such as illegal hunting, overharvesting of rare species of flora and fauna [32]. In response, rural communities have resorted to new livelihood ventures, such as park-based community ecotourism and wildlife safaris enterprises in the form of Community-based Organizations (CBOs) or Trusts near or in Protected Areas (PAs) [5,47,52–55]. In Botswana, the government has introduced the concept of community based natural resource management (CBNRM) and community based organization (CBOs) (e.g., wildlife based Trusts) as a strategy to diversify rural livelihoods and reduce competition for the same resources among stakeholders [24,49]. It is through such initiatives that rural communities are encouraged to establish CBOs (Trusts) to develop community-based tourism enterprises from which they collectively plan, make decisions, manage and operate tourism enterprises and share benefits [5,56,57]. Thus, many rural communities especially those found in or near resource rich (i.e., fauna, flora, and cultural-heritage) protected areas have formed CBOs/Trusts which are comprised of one or several villages with equal rights of ownership and management. For instance, local residents of Khwai, Sankuyo and Mababe villages in the Okavango Delta region were found to benefit from ecotourism ventures via their CBOs/Trusts [54]. In the southern Kalahari region, marginalized communities with CBOs/Trusts accrued benefits from ecotourism activities tied to wildlife in PAs [33,47,58]. Likewise, residents of Khawa village and Ngwatle, Ukhwi and Ncaang settlements, all are located in the Wildlife Management Areas (WMAs) derived benefits from CBNRM - safari hunting operations through their CBOs/Trusts [56]. WMAs are areas reserved by the government for wildlife uses and other conservation activities. Permitted land uses are for consumptive or non-consumptive wildlife utilization. These areas are situated in the buffer zones of PAs and mitigate land uses conflicts, and are used for migratory corridors for wildlife [59]. Overall, tourism activities that occur in all types of protected areas create opportunities (e.g., tour guides, entrepreneurial activities (e.g., beadwork for souvenirs)) as well as lead to increased competitiveness as destinations [3,5,35,36,60,61]. Hence, it is important that local people are actively involved in all spheres of decision-making with regards to community-based ventures [33,37,62], and conservation of resources (i.e., fauna, flora, and cultural heritage) [7,53,63]. 2. Site Context The southwest Kalahari region is popularly known for and is associated with the Kgalagadi Transfrontier Park (KTP) that is conterminous with Botswana and South Africa (see Figure 1). KTP is the first transboundary protected area to be created in southern Africa [48,64,65], and has become important for conservation as well as sources of livelihood for local people that reside within or adjacent to it [5,7,41,52,58]. In Botswana, national parks and game reserves were created to safeguard and maintain wildlife resources, preserve biodiversity, integrate conservation and development activities, foster ecological education and promote park-based tourism to benefit environmental resources and people [41,66,67]. The government’s commitment to conservation and preservation of the natural and cultural resource base, and the promotion of sustainable utilization of such assets are evident [29,68]. The protection and preservation of wildlife resources inside and/or in the Wildlife Management Areas (WMAs) or buffer zones of PAs have created increased numbers of wild animals in some parts of Botswana [69]. There has also been benefits in terms of improved grass cover and biomass for wild animals and for grazing domesticated animals—goats and sheep. It is also in these WMAs that local people have established CBOs through which they venture in tourism—e.g., photographic tourism and community camping sites for wilderness or nature-based tourists.
Agriculture in the Kgalagadi region only benefits a number of small to medium scale commercial farmers [48,64,71], and there is over exploitation of tree resources, especially adjacent to villages and settlements [39,47,72]. Thus, recent recommendations include the necessity for alternative livelihoods in which rural Sustainability people 2020, 12, 1524 could use rangeland in a sustainable manner to benefit themselves and 3 ofthe 17 environment [29,32,41,50]. Figure Figure 1. 1.Map MapofofBotswana Botswana depicting depicting Wildlife Wildlife Management ManagementAreas Areas(WMAs) (WMAs)andandthethe location of of location Kgalagadi Transfrontier Kgalagadi Transfrontier Park Park (KTP) in (KTP) inBotswana southwest southwest Botswana (P.G. (P.G. Koorutwe). Koorutwe). Additionally, In the Kgalagadi, the case of the Kgalagadi region studieshas high have unemployment identified along with that rangelands havelimited opportunities supported to a diversity initiate of income wildlife andgenerating livelihoodactivities activities[47]. suchThe as KTP game offers opportunities ranches (Tsabong,toMaubelo further capitalize on and and Maralaleng developTrust/CBO villages community-based located inecotourism initiativessouth) Tsabong, Kgalagadi [32,33,47,61]. The recent [32], subsistence emphasis hunting, andon sustainable gathering [70]. ecotourism initiatives Agriculture demands in the Kgalagadi that local region onlypeople’s benefitsinvolvement a number ofin tourism small ventures, to medium especially scale those commercial that reside farmers in or near [48,64,71], and PAs therebe increased, is over particularly exploitation of tree among theespecially resources, most affected [16,29]. adjacent Given and to villages the importance [39,47,72]. settlements in the establishment Thus, recent of the KTP for transboundary recommendations include the conservation, necessity forlocal residents alternative have also livelihoods had in expectations which rural peoplefor could development use rangelandopportunities for their in a sustainable mannerrespective to benefit communities themselves and [48,65]. the Additionally,[29,32,41,50]. environment there is limited research that pertains to conservation, tourism, and local communities adjacent to PAs inthe Additionally, southwestern Kgalagadi regionBotswana [19,32,33,47]. has high unemployment Hence,along the purpose of this with limited study was to opportunities examineincome initiate local communities’ attitudes[47]. generating activities andThe support KTP towards the Kgalagadi offers opportunities Transfrontier to further Park.onMore capitalize and specifically, develop to identify factors community-based which influence ecotourism the[32,33,47,61]. initiatives level of support Theorrecent opposition emphasis of KTP. This study on sustainable only focusedinitiatives ecotourism on communities demands onthat the local Botswana sideinvolvement people’s of KTP due to inthe paucity tourism of research. ventures, Additional especially those information that reside in about the South or near PAs beAfrican perspective increased, has been particularly amongdetailed elsewhere the most [7,25,30,58,70]. affected [16,29]. Given the importance in the establishment of the KTP for transboundary conservation, local residents have also 3. Methods had expectations for development opportunities for their respective communities [48,65]. Additionally, there is limited research that pertains to conservation, tourism, and local communities adjacent to 3.1. Study PAs Site in southwestern Botswana [19,32,33,47]. Hence, the purpose of this study was to examine local communities’ attitudes The Kgalagadi region andissupport known for towards the Kgalagadi its unique, Transfrontier large and relatively Park. pristine More specifically, ecosystem, with large-to identify factors which scale migratory routesinfluence for wild the level ofand ungulates support or opposition predatory mammalian of KTP. This study carnivores. Theonly focused region has on communities on the Botswana side of KTP due to the paucity of research. Additional desert features such as salt pans, calcrete rimmed fossil valleys, and undulating and crisscrossing information about the South sand dunes Africanthroughout scattered perspective its haslandscape been detailed elsewhere [7,25,30,58,70]. [41,48,59,65,72]. The attractiveness of southern Kalahari and the greater KTP includes unique natural attractions—birdlife and social weaver nests. 3. Methods Other desert tourism attractions include cultural heritage with ethnic songs, music, dances, traditions, local food, poetry, folklore, handicrafts, religion, language, and traditional costumes [47]. 3.1. Study Site The study site is distinctive as it boasts of unspoiled wilderness, desert-adapted wildlife (e.g., The Kgalagadi elephants), region isofknown and handicrafts for its Kalahari the diverse unique, large peopleand relatively that include pristine ecosystem, San/Basarwa, with BaHerero, large-scale migratory routes for wild ungulates and predatory mammalian carnivores. The region BaKgalagadi, and many others [47,73]. The architecture of dwellings is unique, and is an attraction has desert features such as salt pans, calcrete rimmed fossil valleys, and undulating and crisscrossing sand dunes scattered throughout its landscape [41,48,59,65,72]. The attractiveness of southern Kalahari and the greater KTP includes unique natural attractions—birdlife and social weaver nests. Other desert tourism attractions include cultural heritage with ethnic songs, music, dances, traditions,
Sustainability 2020, 12, 1524 4 of 17 local food, poetry, folklore, handicrafts, religion, language, and traditional costumes [47]. The study site is distinctive as it boasts of unspoiled wilderness, desert-adapted wildlife (e.g., elephants), and handicrafts of the diverse Kalahari people that include San/Basarwa, BaHerero, BaKgalagadi, and many others [47,73]. The architecture of dwellings is unique, and is an attraction in its own way as appreciated at the Trailblaizers cultural village for tourism, situated about 10 miles from the village of Ghanzi in northern Kalahari. The Kgalagadi district is sparsely populated (49,049) with a density of 0.38 people per square kilometer [74]. The population is comprised of six ethnic groups, namely Bangologa, Basarwa, Baherero, Batlharo, Coloureds, and Nama. Residents overwhelming live in the communal areas mostly in and around the villages of Matsheng, Kang and Tsabong [48]. On average, the village/settlement size consists of 198 inhabitants. Within the district, there are more people and settlements in the southern Kgalagadi region (59%) than the north (41%) [48]. The economy is principally based on raising small scale livestock and nominal crop farming, while traditional livelihood activities inclusive of subsistence hunting and gathering are also evident [39,50]. For this study, nine village/settlements were selected from the districts: Kang, Ncaang, Ukhwi, Zutshwa, Tshane (Kgalagadi North) and Khawa, Struizendam, Bokspits, Tsabong (Kgalagadi south). 3.2. Data Collection The targeted respondent was the head of the household. In the event this person (father or mother) was not available, then any member of the family (18 years or older) who had lived in the village or settlement for at least 12 months was requested. The survey questions were translated into the national language—Setswana. The translation was checked and verified for consistency by an expert in English and African languages. The questionnaire was translated back to English [75], as it is the official language and used as the medium of instruction at schools and government institutions. Responses to the survey questions (45–60 minutes approximately) were conducted verbally by the lead author who is a native of Botswana. Collectively 746 responses were completed for a response rate of 75% (see Table 1). Table 1. Selected villages, population, distance from KTP and sample households. Approx. Total Total Village 30% of Household Distance from Village/Settlements Households Population (N) Households Sampled the Park Fence (n) * (km) North Kgalagadi Ncaang 175 43 13 37 250 Ukhwi 453 114 34 59 90 Zutshwa 469 118 35 55 75 Tshane 858 209 63 89 160 Kang 3744 913 274 122 (82) ** 280 South Kgalagadi Khawa 517 128 39 75 21 Struizendam 313 76 23 44 23 Bokspits 499 122 37 53 53 Tsabong 6591 1608 482 212 (145) ** 300 Total (9) 13,619 3331 1000 746 * Household estimate = total population/4.1; Note: ** Initially sampled 30% of the total households and subsequently extracted 30% of this sample (number in parentheses) for data collection. (Source: GoB, 2001).
Sustainability 2020, 12, 1524 5 of 17 3.3. Operationalization of Variables The questionnaire had items that measured various constructs and issues. First, attitudes towards KTP (independent variable) were operationalized with 13 items adapted from the literature [22,26,34,76]. The items focused on conservation priorities, land ownership, perceived benefits, resource use and management. Each was operationalized using a five-point Likert-type scale anchored by 1 (strongly disagree) to 5 (strongly agree). In addition, knowledge about KTP was operationalized with three items (Yes/No/Don’t Know) such as: KTP provides opportunities for community development programs/projects; community campsites outside KTP accrue more money from visitors; and many visitors who visit KTP stay in my district. Support for KTP (dependent variable) was measured via 5 items in a Likert-type scale anchored by 1 (strongly oppose); 2 (Oppose); 3 (Neutral); 4 (support) to 5 (strongly support). The items focused on support or opposition towards KTP regulations and guidelines, management staff, transfrontier status, conservation area, buffer zones and wildlife management areas. The items were adapted from the literature [43,77,78]. Finally, ancillary items that relate to age, gender, education, household income, employment, residency, ethnicity, household size, sources of income, and occupation were also measured. Among these, four items such as literacy (educated/uneducated), proximity to KTP (distance in kilometers), employment status (formal, part-time, self-employed, unemployed, retired), and length of stay (residence in the area) were included as covariates, as past research has demonstrated its applicability at this site [33]. 4. Results 4.1. Profile of Respondents With respect to the sample, 45% of the respondents were from Tsabong and Kang, 20% from Tshane and Khawa, and 35% from Ncaang, Ukhwi, Zutshwa Struizendam and Bokspits. Only 45% were males and 55% females, as reflective of the time of day for the household survey since most men were engaged in agricultural and livestock related work, and women largely focused on domestic chores. Forty one percent were in the 18-30 age category, 40% between 31 and 50 years old, and 19% above 51 years old. About 21% had primary education while 16% had no formal schooling. Only 18% reported a high school education. With respect to distance, 38% lived close (21–99 km) to the park whereas 62% lived further away (100-300 km). About two-thirds had lived at their current address since birth, while a third of the respondents had lived between 1 to 10 years. For income, 42% noted less than P1000 total household income per month, 31% between P1001 and P3500, and 26% reported over P3500 (US $1.00–BWP 10.00 based on May 2017 USA-BWP conversion). The primary source of income was through formal employment (31%) (e.g., security guard, family welfare nurse) and self-employment (24%) (e.g., souvenir production for commercial purposes). In addition, 25% reported to be unemployed and were dependent on government aided welfare support programme. The study area had mixed race/ethnic groupings with Bakgalagadi as the majority, followed by Batlharo, Bangologa, and San/Basarwa. 4.2. Conservation Attitudes and Support for KTP Some of the key findings relate to the fact that nearly all household heads (98%) agreed (strongly agreed and agreed responses combined) that KTP should be protected to benefit future generations, while 92% agreed to the importance in the protection for survival of plants. The majority (91%) agreed that it was essential for the government to devote more money toward a strong conservation program for KTP. However, 87% agreed that if hunting and cattle grazing were allowed in KTP, then wild animals would all disappear. Moreover, the majority (85%) agreed that unlimited access to natural resources (e.g., collecting fuel wood, medicinal plants, herbs, etc.) inside the KTP would lead to loss of and extinction of some rare species. About 70% disagreed (strongly disagreed and disagreed responses combined) that conservation and protection of KTP had taken land from the community. Similarly, 71%
Sustainability 2020, 12, 1524 6 of 17 disagreed that farmers did not have land to cultivate and graze their livestock due to KTP, while 20% agreed. Respondents were also asked to put forth their views about whether it would be better if some parts of the land in KTP were allocated to communities to utilize for agriculture. A sizeable percentage of respondents (69%) disagreed that parts of land from KTP should be allocated for agriculture, while 23% agreed (see Table 2). Table 2. Residents’ attitudes (percentages) towards the Kgalagadi Transfrontier Park (N = 746). Strongly Strongly Statements Disagree Neutral Agree Disagree Agree KTP should be protected for benefit of our future generations. 0.4 0.4 1.3 42.9 55 KTP protection has taken our land from us *. 17.8 51.7 9.5 17.8 3.1 It is important to protect KTP for survival of plants. 0.8 3.6 3.2 68.0 24.3 Farmers don’t have land to cultivate and graze livestock due 19.2 51.7 9.0 15.8 4.3 to KTP*. Staff from KTP has done nothing for villagers’ lives *. 6.4 36.2 20.0 29.8 7.6 It is better if some parts in KTP be allocated to the local people 19.7 49.7 7.4 19.2 3.6 to use for agriculture *. If hunting and grazing in KTP is allowed then wildlife will 2.4 6.3 3.6 57.4 30.0 disappear. If there is unlimited access to forest resources in KTP (Firewood, medicinal plants, forest foods) they will all 3.4 6.7 5.4 61.7 22.9 disappear. It is important for government to devote more money toward 1.2 3.9 4.4 62.9 27.5 a strong a conservation program for the KTP. KTP provides jobs for people from the village. 5.4 31.1 12.1 42.4 9.1 KTP is being managed for the local people. 4.2 35.3 16.4 35.3 8.7 I am happy to have my village next to KTP. 0.9 6.4 9.5 66.2 16.2 It is important to protect KTP for the survival of wildlife. 1.6 1.5 2.9 61.9 32.0 * Item reverse coded prior to analysis. Similarly, majority of respondents (95%) expressed support (strongly support and support responses combined) for the protection of KTP as a conservation area. A sizeable proportion (72%) were supportive of KTP as a transfrontier park. Also, the level of support for the current management staff was noted by 65% of the respondents. Although the majority were supportive of KTP as a transfrontier park, there were others (21%) who opposed it (strongly oppose and oppose responses combined). Additionally, a large number (78%) were supportive of the creation of KTP buffer zones and Wildlife Management Areas, while 73% supported regulations and guidelines that maintained KTP as a transfrontier park (see Table 3). Table 3. Residents’ level of support (percentages) for Kgalagadi Transfrontier Park (N = 746). Strongly Strongly Statements – I Support Oppose Neutral Support Oppose Support KTP as a Transfrontier Park. 9.4 11.9 6.7 51.7 20.0 Current management staff at KTP. 1.1 14.6 19.2 51.5 13.5 Creation of buffer zones and WMAs. 2.1 6.2 13.0 58.8 19.6 Regulation and guidelines for KTP. 4.3 7.6 15.3 54.2 18.5 Protection of KTP as a conservation area. 0.4 1.2 2.8 65.5 29.8
Sustainability 2020, 12, 1524 7 of 17 4.3. Data Analysis First, the 13-item independent attitudinal measures were assessed for appropriateness to conduct data analysis via Bartlett’s test of sphericity and Kaiser-Meyer-Oaklin (KMO) measure of sampling adequacy. Bartlett’s test of sphericity was highly significant (approximate chi-square = 1386.647, df = 78, p < 0.001), and KMO was 0.752, which exceeded [79] recommended cut-off of 0.60. Likewise, for the five-item dependent variable measure (i.e., support) Bartlett’s test of sphericity was also highly significant (approximate chi-square = 753.445, df = 10, p < 0.001), and KMO was 0.699. These two measures for the independent and dependent variable measures suggested that the data was suitable for factor analysis. Second, a principal component analysis (PCA) was performed using Varimax rotation with Kaiser normalization. Four factors (i.e., land ownership, conservation awareness, local benefits, and resource use) were generated and accounted for 53.14% of the total variance which exceeded the minimum cut-off of the 50% [79,80]. The four factors explained 20.09%, 12.27%, 10.03%, and 7.75% of the variance, and had eigenvalues of 3.00, 1.59, 1.30, and 1.01, respectively (see Table 4). The single factor for the criterion or dependent measure explained 46.71% of the total variance. Factor loadings 0.40 or greater were considered significant [81]. The reliability score ranged from 0.56 to 0.70, and were deemed acceptable [81]. Considering these scores, [82] recommend reporting the mean inter-item correlation for the items, and suggests an optimal range of 0.20 to 0.40. For this study, the ranges of inter-item correlation means were 0.241 to 0.440, and noted to be acceptable. Although reliabilities were lower than desired, they were not low enough to justify discontinuation given the range of mean inter-item correlations, number of items, and the unique nature of the data [81–84]. The items within each factor were computed as independent index, respectively. In addition, an index was created for the three items that measured perceived benefits about KTP which were: 1.) KTP provides opportunities for community development; community; 2) Community campsites accrue more money from visitors; and 3) Many visitors who visit KTP stay in my district. Table 4. Factor loadings and reliabilities. % of Mean No. of Scale Constructs and Items Loading Variance Inter-item Reliability Items Mean * Explained Correlation 10.92 Land Ownership 3 20.09 0.440 0.70 (2.58) Farmers don’t have land to cultivate and 0.806 graze livestock due to KTP. KTP protection has taken our land from us. 0.802 It is better if some parts of land in KTP is allocated to local people to 0.705 use for agriculture. 16.96 Conservation Awareness 4 12.27 0.242 0.56 (1.81) KTP should be protected for the future of 0.701 our new generation. It is important to protect KTP for the 0.601 survival of wildlife. It is important for the government to devote more money towards a strong conservation 0.570 program for KTP. It is important to protect KTP for 0.479 survival of plants.
Sustainability 2020, 12, 1524 8 of 17 Table 4. Cont. % of Mean No. of Scale Constructs and Items Loading Variance Inter-item Reliability Items Mean * Explained Correlation 10.76 Local Benefits 4 10.03 0.241 0.56 (2.74) KTP provides jobs to people from the village. 0.765 Staff from KTP have done nothing for 0.682 villagers’ lives. 8 of 16 KTP is being managed for the local people. 0.641 8 of 16 8 of 16 8 of 16 8 of 16 I am happy to have my village 8 of 16 I am happy to have my village next to KTP. 0.415 0.415 next to KTP. I am happy to I am havehappy my village Itoam have happymy village Itoam havehappy my village to have my village 8.01 8.01 I am happyResource to have my village 0.415 0.415 0.415 0.396 7.750.415 next to KTP.next Resource toUse KTP.next Use to KTP.next 2 to KTP. 0.415 2 7.75 0.57 0.396 (1.52) 0.57 (1.52) next to KTP. 8 of 16 8.01 8.01 8.01 8.01 IfIfhunting and huntingResource and grazing Resourcegrazing inResource Use KTP in KTP is Use is Resource allowed 2 UseResource 2 Use 7.752 7.75 0.396 2 7.75 0.396 0.57 7.758.01 0.3960.57 0.396 0.57 0.57 Use 2 7.75 0.789 0.396 0.57 ( 1.52) ( 1.52) ( 1.52) ( 1.52) allowed Ithen am then wild happy wild animals animals to will will disappear. 0.789 (1.52) If hunting disappear. andIfhave my hunting grazing village and inIf KTP hunting grazing is and in If KTP hunting grazing is 0.415 and in KTP grazing is in KTP is If hunting next to KTP. and grazing in KTP is 8 of 16 allowed then allowed wild animals then allowed wild will to animals then allowed wild willanimals then0.789 wild willanimals 0.789 will 0.789 0.789 IfIfthere thereisisunlimited allowed unlimited then wild animals access access to will forest 0.789 8.01 disappear. resources in disappear. KTP Resource (e.g., disappear. firewood, Use(e.g., disappear. medicinal 2 7.75 0.775 0.396 0.57 forest resources disappear. in KTP (1.52) Iplants) am happy If there then toIfhave ismedicinal unlimited there they myis village access will unlimited If there also to isaccess disappear. unlimited If there to isaccess0.775to access to unlimited firewood, If there hunting is unlimited and grazing plants) access in KTPtothenis 0.415 next to KTP. forest resources forestin resources KTP (e.g., forest in resources KTP (e.g., forest in resources KTP (e.g.,in KTP (e.g., they forest allowed will alsoDependent resources then disappear. wild in animals KTP (e.g., Measure will 0.775 0.789 0.775 0.775 0.775 8.01 firewood,Resource medicinal firewood, plants) Use medicinal firewood, then plants) medicinal firewood, 2 thenplants) medicinal 0.775 thenplants) 7.75 then 0.396 0.57 firewood, disappear. Dependent medicinal Measure plants) then (1.52) they will also they disappear. will also theydisappear. will also theydisappear. will also disappear. 19.08 19.08 they If therewillisalso hunting unlimited KTP KTP anddisappear. grazing Support Support access in KTP to is Dependent 5 Dependent 5 46.71 .327 46.71 0.70 .327 0.70 Dependent Measure Dependent Measure Measure Measure ( 3.16) (3.16) allowed Dependent forest resources then wild Measure in animals KTP (e.g., will 0.789 Regulations and guidelines 0.775 Support 19.08 19.08 19.08 19.08 Regulations firewood, disappear. KTPand medicinal Support KTPto guidelines plants) thento maintain Support 5 KTPKTP KTP Support 5 46.715 46.71.327 5 46.71 .327 0.70 46.71 19.08 .3270.70 .327 0.70 0.70 KTP Support 5 46.710.789 .327 0.70 (3.16) (3.16) (3.16) (3.16) maintain as they If a KTP Transfrontier willisalso there as a Transfrontier disappear. unlimited Park. accessRegulations to toguidelines 0.789 ( 3.16) Regulations Park. Regulations and guidelines and Regulations andtoguidelines andtoguidelines to Regulations Current forest Dependentandin management resources guidelines Measure staff to at KTP. 0.753 0.789 maintain KTP as aKTP maintain (e.g., Transfrontier KTP maintain as a Transfrontier KTP maintain as a Transfrontier KTP0.789 0.775as a Transfrontier 0.789 0.789 Current management maintain KTP as a staff at KTP. Transfrontier 0.753 0.789 19.08 firewood, Park.as a medicinalPark. plants) Park.then Park. KTP KTP SupportNational Transfrontier National Park. 5 46.71 0.701 .327 0.70 KTP Park. they as a Transfrontier will also disappear. 0.701 (3.16) Current management Current management staff Current at KTP. management staff Current at KTP. management staff 0.753 at KTP.staff 0.753 at KTP. 0.753 0.753 Park. Current Regulations Protection management and of KTP staff guidelines as attoKTP. aNational conservation area. 0.753 0.614 KTP as Dependent a Transfrontier KTP Measure as a Transfrontier KTP as a Transfrontier National KTP as a Transfrontier National National Protection of KTP KTP as a Transfrontier maintain KTP as a as a National Transfrontier 0.701 0.789 0.701 0.701 0.701 19.08 Park. Creation of KTP Park. KTP buffer Support Park. zones and Park. WMAs. 5 0.614 0.701 46.71 0.527 .327 0.70 conservation area. Park. (3.16) Protection ofProtection KTP as a ofProtection KTP as a ofProtectionKTP as a of KTP as a Creation Protection Current Regulations ofofKTPKTPbuffer management and zones as astaff guidelines and attoKTP. * Scale 0.614standard 0.753 deviation 0.614 in brackets. 0.614 0.614 conservation conservation area. conservation area. conservation area. 0.527 area. 0.614 WMAs. conservation KTP maintain KTParea. as a Transfrontier National as a Transfrontier 0.789 Creation of KTP Creation buffer ofzones KTP Creation buffer andof zonesKTP Creationbuffer andof0.701 zones KTP buffer and zones and Creation of KTP buffer zones and * Scale standard Park. 0.527 deviation in 0.527 brackets. 0.527 0.527 WMAs.Furthermore, WMAs. management Protection Current WMAs. several of KTP as astaff at KTP. WMAs. techniques WMAs. 0.527 were used to ensure against multi-collinearity. First, a correlation 0.753 analysis was Furthermore, conservation KTP area. as a Transfrontierconducted several National along techniques* Scale with were a0.614 * Scale standard standard * Scale used deviation standard collinearity deviation * Scale to ensure in deviation brackets. standard * Scale diagnosis. in brackets. indeviation brackets. standard indeviation The bivariate against multi-collinearity. brackets. in brackets. correlations First, a ranged between 0.701 correlation Creation Park. 0.01 and of analysis KTP 0.49, buffer and waszones were conducted and below along 0.70 with [80] a collinearity (see Table diagnosis. 5). The Subsequently, bivariate cutcorrelations off points were earmarked Furthermore, Furthermore, severalFurthermore, techniques severalFurthermore, techniques were several 0.527used techniques were several to ensure usedtechniques were to against ensure usedweremulti-collinearity. to against ensure used multi-collinearity. toagainst ensure First, multi-collinearity. against a First, multi-collinearity. a First, a First, a WMAs. Furthermore, Protection ranged of KTP0.01 between asseveral a and 0.49, techniques and werewere below used 0.70 to [80]ensure (see Tableagainst multi-collinearity. 5). Subsequently, cut offFirst, pointsa toconservation determine correlation area. multicollinearity—tolerance correlation analysis was correlation analysis conducted was correlation analysis * Scale conducted along standard 0.614 was with analysis value conducted along a collinearity was with of less conducted along a 0.10, collinearity diagnosis. with along a or a variance collinearity diagnosis. The withbivariate a inflation collinearity diagnosis. The correlations bivariate factor diagnosis. The correlations bivariate The(VIF) value correlations bivariate correlations correlation were earmarked analysis was conducted to determine along withdeviation multicollinearity—tolerance in brackets. a collinearity valuediagnosis. of less 0.10,The orbivariate a variance correlations inflation ranged ofCreation above between ranged 10 [84] 0.01 between (p. and ranged 0.49, 164). 0.01 between and ranged and The were 0.49, 0.01 between below and and tolerance were 0.49, 0.70 0.01 below and [80] and (see were 0.49, 0.70Table below and [80] were 5). (see 0.70 Subsequently, Table below [80] (see 5).0.70 Subsequently, Table [80] cut (see 5). off Subsequently, points Table 0.98 or less with 0.98off points cut 5). off Subsequently, points cut off points cut ranged factor (VIF)of KTP between value buffer 0.01 zones and of above and 0.49, 10 and [84] were (p. 164). below The 0.527 0.70value tolerance[80] (see for value each Table for 5). of each of the variables Subsequently, the variables cutwasoffwaspoints 0.98 or were WMAs. earmarked Furthermore, were earmarked to determine were several earmarked to multicollinearity—tolerance determine techniques were earmarked to determine multicollinearity—tolerance to determine multicollinearity—tolerance valuemulticollinearity—tolerance of lessvalue 0.10, of orless avalue variance 0.10, ofor lessinflation a value variance 0.10,oforless ainflation variance 0.10, or ainflation variance inflation as the were highest, earmarked to which determine greaterwere thanused is multicollinearity—tolerance 0.10.to The ensure less with .98 as the highest, which is greater than 0.10. The variance inflation factor (VIF) was 1.432 against variance value multi-collinearity. inflation of less 0.10, First, (VIF)a was 1.432 (highest), factorinflation or a variance factor (VIF)factor correlation value(VIF) of above analysis factor value10 (VIF) of[84] *above factor value (p. standard Scale 164). 10 (VIF) of[84] above The value (p. tolerance 164). 10of[84] above deviationThe in(p. value tolerance 164). 10 [84] for brackets.The (p. each value tolerance 164). of for the Theeach variables value tolerance offor thewas each variables value 0.98 offor the orwas each variables 0.98 of the orwas variables 0.98 orwas 0.98 or factor (VIF) (highest), which value which was was below of was above below theconducted 10 the[84] cut-offcut-offalong (p.of164). of with 10. 10. The a collinearity tolerance Collectively, Collectively, value thethediagnosis. for each ofThe multicollinearity the bivariate variables assumption correlations waswas0.98notor less with ranged .98 betweenless as with the highest, 0.01 .98 less and aswith the which 0.49, highest, .98 and less as is with weregreater the which highest, .98 below than as is0.70 the greater which 0.10. highest, [80] The than (see 0.10.multicollinearity is greater variance which Table The than isinflation 5). greater variance 0.10. The than factor Subsequently,inflationassumption variance 0.10. (VIF) cutThe factor was inflation off variance 1.432 was1.432 (VIF)factor was not inflation (VIF) violated. factor was 1.432 (VIF) was 1.432 less with In violated. .98addition, as the highest, which validity discriminant is greater than 0.10.was assessment Theperformed, variance inflation and thefactor bivariate(VIF) waspoints 1.432 correlations In addition, (highest), which Furthermore, were earmarked discriminant (highest), was which below (highest), several totodetermine was the validity which cut-off techniques(highest), below was the of wereassessment 10. which cut-off below Collectively, used was the ofto10. was cut-off below performed, Collectively, ensurethethe ofmulticollinearity 10. cut-off Collectively, against theof and the multicollinearity 10. Collectively, assumption the bivariate multicollinearity multi-collinearity. assumption the was First, correlations multicollinearity not a assumption was not ranged assumption was not was not (highest), ranged which from 0.01was below 0.49, themulticollinearity—tolerance while cut-off of 10.were reliabilities Collectively, from 0.56the value of Thus, less 0.10, to multicollinearity 0.70. orassumption a variance discriminant inflation waswas validity not violated. correlation from factor 0.01 (VIF) Inanalysis violated. toaddition, value0.49, of Indiscriminant wasviolated. addition, while above conducted 10 Inviolated. [84] discriminant addition, validity (p.along reliabilities 164). Inassessment discriminant with were Theaddition, validity a from discriminant assessment was validity collinearity tolerance 0.56 valueperformed, assessment wasvalidity diagnosis. to 0.70. for each performed, and of The Thus, the assessment thewas bivariate performed, and the bivariate discriminant variables was correlations bivariate performed, and correlations was 0.98 thecorrelations validity or bivariate and was thecorrelations bivariatefor verified correlations violated. In the verified for addition, discriminant validity assessment was performed, and the bivariate correlations constructs. ranged ranged less with from between .98 ranged as0.01 0.01 the to from 0.49, ranged and highest,0.01 while 0.49, to from whichand0.49, reliabilities ranged 0.01 were is while to greaterfrom 0.49, reliabilities below were 0.01 than while from 0.70 to[80] 0.10. 0.49, reliabilities were 0.56 The while (see to from 0.70. Table variance reliabilities were 0.56 Thus, 5). to from 0.70. discriminant were 0.56 Thus, Subsequently, inflation factor to from 0.70. discriminant cut (VIF)validity 0.56 Thus, off was to1.432 0.70. discriminant was points validity Thus, discriminant was validity wasvalidity was the constructs. ranged from 0.01 to 0.49, while reliabilities were from 0.56 to 0.70. Thus, discriminant validity was verified were for earmarked verified the was constructs. forverified to determinethe constructs. for verified the constructs. for Collectively, themulticollinearity—tolerance the constructs. value (highest), verified for which theTable 5.below constructs. Correlation cut-off matrix of 10. of study the variables, means andofstandard less 0.10, multicollinearity orassumption a variance inflation deviations. was not factor (VIF) violated. value of above In addition, 10 [84] validity discriminant (p. 164).assessment The tolerance was value for eachand performed, of the the variables correlations bivariate was 0.98 or Table Ƶ1 highest, 5. Correlation ƵTable matrix of study variables, means and standard deviations. Variables less rangedwith .98 as from Table the 0.01 Tableto 5. 5. Correlation 0.49, 2 which while Correlation isƵ3 greater 5. matrix Correlation Table reliabilities matrix 5.Ƶthan of study were of study X15. 4Correlation matrix Table variables, of 0.10. from study The 0.56 variables, X2ofand Correlation matrix means variables, study variance to 0.70. means X3 matrix standard means variables, of and study inflation Thus, and standard X4 deviations. standard means factor discriminant(VIF) deviations. X5standard variables, and was validity Y1 deviations. means and 1.432 was deviations. standard deviations. Covariates (highest), verified Variableswhich Variables was for theVariables below constructs. Ƶ11 the Ƶ2 Ƶ21 Variables cut-off of 10. Collectively, the multicollinearity Ƶ3 Ƶ2 Ƶ31 Ƶ4 Ƶ3 Ƶ2 Ƶ41 X1 Ƶ4 Ƶ3 Ƶ2X1 Variables X2X1Ƶ4 Ƶ3 X3 assumption X2X1Ƶ4 X4X3 X2 was X2X1 X5X4X3 X3 not X2 Y1 X5X4X3 X4 X5 Y1 X5X4Y1 Y1 X5 Y1 Literacy Variables( 1) 1.000 Ƶ1 discriminantƵ2 Ƶ3 Ƶ4 X1was performed, X2 X3 X4 X5 Y1 violated. In addition, validity assessment Covariates Covariates and the bivariate Covariates Covariatescorrelations Proximity (Ƶ2) 0.11 ** 1.000 Covariates Covariates ranged from Literacy ( 1)0.01 to 5. Literacy Table 0.49, 1.000 while ( Correlation 1)Literacy 1.000(reliabilities 1)Literacy matrix 1.000( were of study from 1.000 0.56 1) variables, to 0.70. means Thus, discriminant and standard deviations. validity was Employment Literacy ( 1) Literacy −0.20 1.000 verified for (Ƶ Proximity the 2) constructs. 1.000 Proximity 0.11 2−0.10 **(ƵProximity ) 1.000 ** **(Ƶ1.000 0.11 Proximity 2) 1.0000.11 **(Ƶ2) 1.000 0.11 ** 1.000 status (Ƶ Variables Proximity )) 2) 0.11 ( 31(Ƶ Ƶ**1 ** Ƶ2 1.000 Ƶ3 Ƶ4 X1 X2 X3 X4 X5 Y1 EmploymentEmployment −0.20 Employment −0.20 Employment −0.20 −0.20 Length Employmentof stay Proximity −0.50 −0.20 −0.10 ** 1.000 −0.10 ** Covariates 1.000 −0.10 ** 1.000 −0.10 ** 1.000 status (Ƶ3) Table status 0.11**(Ƶ 5. ) −0.09 status Correlation ** −0.10 1.000 3 * ****(Ƶ3 0.31 ) matrix 1.000** statusof** (Ƶ1.000 study 3 ) ** variables, means and standard deviations. (Ƶ4)(Ƶ status Literacy ( (3))1) ** ** 1.000 Length of2 stay Length−0.50 of stay Length−0.50 of stayLength−0.50of stay −0.50 Variables Length Proximity of stay (Ƶ 2) Ƶ −0.50 0.11 1 ** −0.09 Ƶ 1.000 2 * 0.31 Independent Ƶ3−0.09 ** * 1.000 Ƶ4 0.31 ** Measures −0.09 * 1.000 X1 0.31 −0.09 **X2* 1.000 0.31X3** 1.000X4 X5 Y1 Employment (Ƶ4) (Ƶ4**) −0.09(Ƶ*4**) 0.31 **(Ƶ4** ) 1.000 ** Perceived Employment) (Ƶ4status ** ** −0.20 −0.20 −0.10 ** 1.000 Covariates −0.07 −0.13 ** −0.10 ** −0.07 −0.04 1.000 Independent 1.000 Independent Measures Independent MeasuresIndependent Measures Measures benefits Literacy ((X1) status (Ƶ (33))1) ** 1.000 Independent Measures Perceived Perceived Perceived Perceived Land Perceived Length Proximity of stay (Ƶ2) 0.11 −0.07 −0.50 ** −0.13 1.000 −0.07 ** −0.07 −0.13 −0.07 ** −0.04 −0.07 −0.13 −0.07 ** 1.000 −0.04 −0.07 −0.13 **1.000 −0.04 −0.07 1.000−0.04 1.000 Length benefits (X1)benefits −0.07 (X1)−0.13 benefits −0.09** * (X1) ** benefits −0.07 0.31 ** −0.15 (X1) −0.04 1.000 1.000 Ownership benefits4) (X1) (Ƶof Employment stay 0.12** ** ** 0.20−0.09 −0.20 −0.50 *−0.07Land 0.31 **** 0.04 1.000 1.000 Land (X2)(Ƶ Land Land −0.10 ** 1.000 Land status ( 34)) ** −0.15 Independent −0.15 Measures −0.15 −0.15 Ownership Ownership 0.12 ** Ownership0.200.12** ** Ownership −0.07 0.20 0.12**−0.15 ** −0.07 0.20 0.12**0.04 ** −0.070.201.000 ** 0.04−0.07 1.000 0.04 1.000 0.04 1.000 Conservation Ownership Perceived Length of stay 0.12 −0.50 ** 0.20 ** −0.07 ** 0.04 ** 1.000 ** ** (X2) (X2) −0.07 −0.13 −0.09(X2) ** * −0.13* −0.07 0.31 ** (X2) ** −0.04 1.000 1.000 Awareness (X2) benefits (Ƶ4) (X1) −0.06 ** −0.06 0.01 0.23 ** 0.25 ** 1.000 Conservation Conservation Conservation * Conservation (X3) Conservation Land −0.13* Independent −0.13* Measures −0.13* −0.13* Awareness Awareness −0.06 Awareness −0.06 −0.06 −0.13*Awareness −0.06 −0.06 0.01 −0.06 −0.15 −0.06 0.230.01 ** −0.060.25 0.23 **0.01 ** 1.000 0.25 0.23 **0.01 ** 1.000 0.25 0.23 ** ** 1.000 0.25 ** 1.000 Awareness Ownership Perceived −0.06 0.12 ** −0.06 0.20 ** −0.07* 0.01 * 0.23 0.04** *0.25 1.000** * 1.000 (X3) (X3) −0.07 −0.13(X3)** * (X3) −0.04 −0.07 ** 1.000 (X3) (X1) (X2) benefits Conservation Land −0.13* −0.15 Awareness Ownership −0.06 0.12 ** −0.06 0.20 ** −0.07 0.01 0.23 0.04** 0.25 1.000** 1.000 * ** (X3) (X2) Conservation −0.13* Awareness −0.06 −0.06 0.01 0.23 ** 0.25 ** 1.000
ranged between ranged0.01 between and ranged 0.49, 0.01 between and ranged andwere 0.49, 0.01 between below and andwere 0.49, 0.70 0.01below and [80] and(see were 0.49, 0.70 Table below and [80] were 5). (see 0.70 Subsequently, Table below [80] (see 5).0.70 Subsequently, Table [80] cut (see 5). off Subsequently, points Table cut 5). offSubsequently, points cut off points cut off points were earmarked were earmarked to determine were earmarked to multicollinearity—tolerance determine were earmarked to determine multicollinearity—tolerance to determine multicollinearity—tolerance value multicollinearity—tolerance of less value 0.10,oforless avalue variance 0.10,oforless inflation avalue variance 0.10,oforless ainflation variance 0.10, or ainflation variance inflation factor (VIF)factor value(VIF) of above factor value10 (VIF) of[84] above factor value (p. 164). 10 (VIF) of[84] above The value (p.tolerance 164). 10of[84] above The(p. value tolerance 164). 10 [84] forThe (p. each value tolerance 164). of for the The each variables value tolerance offor thewas each variables value 0.98 offor the orwas each variables 0.98 of the orwas variables 0.98 orwas 0.98 or less with .98less as with the highest, .98 less aswith the which highest, .98 less as is with greater the whichhighest, .98 than as is the greater which 0.10. highest, The than is greater variance which 0.10. The than isinflation greater variance 0.10. The than factor inflation variance 0.10. (VIF) The factor was inflation variance 1.432 (VIF)factor was inflation 1.432 (VIF)factor was 1.432 (VIF) was 1.432 (highest), which (highest), was which below (highest), was the which cut-off (highest), belowwas the of 10.which cut-off below Collectively, was the of 10. cut-off below Collectively, thethe ofmulticollinearity 10. cut-off Collectively, theofmulticollinearity 10. Collectively, assumption the multicollinearity assumption the was multicollinearity not assumption was not assumption was not was not violated. Inviolated. addition,Indiscriminant violated. addition,Inviolated. discriminant addition, validityInassessment discriminant addition, validity discriminant assessment was validity performed, assessment was validity performed, and theassessment wasbivariate performed, and thewas correlations bivariate performed, and thecorrelations bivariate and thecorrelations bivariate correlations Sustainability 2020, 12, 1524 9 of 17 ranged from ranged 0.01 to from 0.49, ranged 0.01 while to from0.49, reliabilities ranged 0.01 while to from0.49, reliabilities were 0.01 while from to 0.49, reliabilities were 0.56while to from 0.70. reliabilities were 0.56 Thus, to from 0.70. discriminant were 0.56 Thus,to from 0.70. discriminant validity 0.56 Thus,to 0.70. discriminant wasvalidity Thus, discriminant wasvalidity was validity was verified forverified the constructs. forverified the constructs. forverified the constructs. for the constructs. Table 5. Correlation Table 5. matrix Correlation Tableof study 5. Correlation matrix Table Table variables, of 5. study 5.ofCont. Correlation matrix means variables, and study matrix standard means variables, of and study deviations. standard means variables, and deviations. standard means anddeviations. standard deviations. Variables Variables VariablesƵ11 Ƶ2 Ƶ21 Variables Ƶ3 Ƶ2 Ƶ31 Ƶ4 Ƶ3 Ƶ2 Ƶ41 X1 Ƶ4 Ƶ3 Ƶ2X1 Variables X2X1Ƶ4 Ƶ3 X3 X2X2X1Ƶ4 X4X3 X2X1 X5X4X3 X3 X2 Y1 X5X4X3 X4 X5 Y1 X5X4Y1 Y1 X5 Y1 Covariates Covariates Independent Measures Covariates Covariates Literacy ( 1)Literacy 1.000( 1)Literacy 1.000 ( 1)Literacy 1.000 ( 1) 1.000 Perceived Proximity (ƵProximity 2) 0.11 **(ƵProximity 2) 1.000 0.11 **(ƵProximity 2) 1.000 0.11 **(Ƶ2) 1.0000.11 ** 1.000 benefits −0.07 −0.13 ** −0.07 −0.04 1.000 Employment −0.20 Employment (X1) Employment −0.20 Employment−0.20 −0.20 −0.10 ** 1.000 −0.10 ** 1.000 −0.10 ** 1.000 −0.10 ** 1.000 status (Ƶ3) status**(Ƶ3) status**(Ƶ3) status**(Ƶ3) ** Land Length of stay Length−0.50 Length−0.50 of stay Length−0.50 of stay of stay −0.50 Ownership 0.12 ** −0.09 0.20* ** 0.31 −0.07 −0.09 ** * 1.000 0.31 −0.15 ** * ** −0.09 1.000 0.31 **0.04 −0.09 * 1.000 0.31 **1.000 1.000 4) (Ƶ(X2) (Ƶ4** ) (Ƶ4** ) (Ƶ4** ) ** IndependentIndependent Measures Independent MeasuresIndependent Measures Measures Conservation Perceived Perceived Perceived Perceived Awareness −0.07 −0.06 −0.13 −0.06 −0.07 ** −0.13** −0.07 −0.13−0.07 ** −0.04−0.07 −0.13 0.01 −0.07 ** 1.000 −0.04 −0.070.23 −0.13 **−0.04 **1.000 −0.070.251.000 **−0.04 1.000 1.000 benefits (X3)(X1)benefits (X1)benefits (X1)benefits (X1) Land Land Land Land Local −0.15 −0.15 −0.15 −0.15 Ownership Ownership 0.12 ** Ownership 0.20 0.12 ** ** Ownership −0.07 0.20 0.12 ** ** −0.070.20 0.12 **0.04 ** −0.07 0.201.000 ** 0.04−0.07 1.000 0.04 1.000 0.04 1.000 Benefits 0.13 ** 0.14 ** 0.05 ** 0.02 ** −0.40 ** ** −0.17 ** ** −0.32 ** 1.000 (X2) (X2) (X2) (X2) (X4) Conservation Conservation Conservation Conservation Resource −0.13* −0.13* −0.13* −0.13* Awareness Awareness−0.06 −0.05 Awareness −0.06 −0.06 Awareness −0.01 −0.06 −0.06 0.01 −0.06 −0.06 0.230.01 ** −0.06 0.25 0.23**0.01 ** 1.000 0.25 0.23 **0.01 ** 1.000 0.25 0.23 **** ** −0.176 1.000 0.25 ** 1.000 Use (X5) * −0.07 * −0.01 * 0.10 ** * 0.26** 0.35 ** 1.000 (X3) (X3) (X3) (X3) Dependent Measure KTP Support −0.16 ** −0.11 ** −0.07 0.04 0.30 ** 0.16 ** 0.25 ** −0.40 ** 0.16 ** 1.000 (Y1) Mean 0.60 0.62 2.46 28.51 0.37 3.64 4.24 2.69 4.00 3.82 Standard 0.49 0.49 1.20 19.84 0.32 0.86 0.45 0.68 0.76 0.63 Deviation * Correlation Significant at the 0.05 level (2 tailed); ** Correlation significant at the 0.01 level (2 tailed). 4.4. Regression Analysis Hierarchical regression was conducted with five predictor independent measures (perceived benefits, land ownership, conservation awareness, local benefits, and resource use) along with four covariates (literacy, proximity, employment status, and length of stay) on the outcome dependent measure (KTP support). Except for perceived benefits that employed an index, means of the predictor and outcome constructs were used in the analysis. The use of means has two advantages. First, “it provides a means of overcoming to some extent the measurement error inherent in all measured variables,” and second, the mean is able “to represent the multiple aspects of a concept in a single measure” [81] (p. 116–117). In the analysis, two models were produced and referred to as Model 1 and 2 (see Table 6). Model 1 presents the effects of the four covariates (literacy, proximity, employment status, and length of stay) on KTP support. The relationship was significant (F = 8.854, p < 0.001) as the variables explained 4.6% of the variance in KTP support. Literacy (beta = −0.183, p < 0.001), proximity (beta = −0.100, p < 0.01) and employment status (beta = −0.102, p < 0.01) were all significant and had negative relationships with KTP Support. However, length of stay failed to register a significant relationship. So, the next step was to control the four covariates, and identify if the five independent variables could predict a significant amount of variance in KTP support. Model 2 was highly significant (F= 23.281, p < 0.001) along with the change in the F value (F = 33.273, p < 0.001). Results demonstrated that with the control of the four covariates, the five independent variables predicted a significant amount of variance in KTP support. This new model explained 22.2% of the variance with the five independent variables that accounted for 17.7% of additional variance. Of the four covariates, only literacy had a significant effect on KTP support (beta = −0.116, p < 0.01). Essentially, any increase in the level of literacy resulted in support for KTP. Among the five independent variables, only resource use was not statistically significant. Perceived benefits (beta = 0.143, p < 0.001), land ownership (beta= 0.109, p < 0.01), conservation awareness (beta= 0.071, p < 0.05), and local benefits (beta= 0.272, p < 0.001) were all positively related to KTP support. In addition, local benefits had the strongest relationship followed by perceived benefits, land ownership, and then
WMAs. correlation analysis was Furthermore, conducted several techniquesalongwere with useda collinearity to ensure diagnosis. against The bivariate correlations multi-collinearity. First, a * Scale standard deviation in brackets. ranged between correlation 0.01 analysis and was 0.49, and conducted werealongbelow with 0.70 a [80] (see collinearity Table 5). diagnosis.Subsequently, Furthermore, several techniques were used to ensure against multi-collinearity. First, a The cut bivariate off points correlations were earmarked ranged between to0.01 correlation determine and analysis multicollinearity—tolerance 0.49, and was were below conducted 0.70 along [80]to with value a(see of less Table collinearity5). 0.10, or a variance Subsequently, diagnosis. Thecut inflation off points bivariate correlations Furthermore, several techniques were used ensure against multi-collinearity. First, a factor (VIF) werecorrelationvalue earmarked of to above determine10 [84] (p. 164). The tolerance multicollinearity—tolerance valuevaluefor each of lessof the 0.10, variables or a was variance 0.98cut or off points rangedanalysis betweenwas 0.01conducted and 0.49, and along werewith below 0.70 [80] (see a collinearity Table 5). diagnosis. The bivariate inflation Subsequently, correlations less with(VIF) factor .98 wereas the value highest, of above earmarked which to 10 [84]is(p. determine greater 164). than The 0.10. The value tolerance multicollinearity—tolerance variance inflation forTable each value ofof factor the (VIF) variables less 0.10, was orwas acut 1.432or inflation 0.98 variance ranged Sustainability between 2020, 12, 1524 0.01 and 0.49, and were below 0.70 [80] (see 5). Subsequently, off points 10 of 17 (highest), withwhich less were .98 factoras was the below highest, (VIF) value the cut-off which is of 10. greater Collectively, than of abovemulticollinearity—tolerance0.10. the The 10 [84] (p. 164). The tolerance multicollinearity variance inflation value for each assumption factor (VIF) was was notwas 1.432 earmarked to determine value of less 0.10,of orthe variables a variance 0.98 or inflation violated. (highest),Inless addition, which was discriminant below the validity cut-off of assessment 10. was Collectively, performed, the and the bivariate multicollinearity correlations assumption was not factor (VIF) value of above 10 [84] (p. 164). The tolerance value for each of the variables was 0.98 or1.432 with .98 as the highest, which is greater than 0.10. The variance inflation factor (VIF) was ranged from violated. In 0.01 to 0.49, addition, while discriminant reliabilities validity were from assessment 0.56 wasto 0.70. Thus, performed, the anddiscriminant the bivariate validity was was not conservation (highest), less with .98 aswhich awareness. the was below highest, Basically, which any theis cut-off greater increase of in 10. thaneachCollectively, 0.10. ofThe the variance independent inflation factor correlations multicollinearity variables assumption (VIF) (i.e.,was 1.432 perceived verified for ranged from the constructs. 0.01 violated. to addition, In 0.49, while reliabilities discriminant were validity from 0.56 to 0.70. assessment Thus, discriminant wasmulticollinearity performed, and the validity bivariatewascorrelations benefits,(highest), which land ownership, was below conservationthe cut-off of awareness, 10. Collectively, the and local benefits) resulted inassumption support forwas not KTP. verified for violated. the rangedIn constructs. from 0.01discriminant addition, to 0.49, while reliabilities validity were from assessment was 0.56 to 0.70.and performed, Thus,the discriminant bivariate validity was correlations Table 5. Correlation matrix of study variables, means and standard deviations. ranged verified from6.for0.01 theto constructs. 0.49, while reliabilities were from 0.56 to 0.70. Thus, discriminant Table Table Hierarchical 5. Correlation regression matrix of analysis study with variables, KTP means support as dependent variable. validity was Variables Ƶ verified for the constructs. 1 Ƶ2 Ƶ3 Ƶ4 X1 X2 and standard X3 deviations. X4 X5 Y1 Variables Ƶ1 TableƵ2 5. Correlation Ƶ3 matrix Ƶ4 of 1study Covariates Model X1 variables,X2 means X3and standard deviations. X4 Model 2X5 Y1 Literacy ( 1) 1.000Table 5. Correlation matrix ofCovariates study variables, means and standard deviations. Variables Ƶ1 B Ƶ2 Ƶ3 Beta Ƶ4 T-Value X1 X2 B X3 Beta X4 X5 T-Value Y1 Proximity Literacy(Ƶ( 2)1) 0.11 ** 1.000 1.000 Covariates Variables−0.20 Ƶ1 Ƶ2 Ƶ3 Ƶ4 Covariates X1 X2 X3 X4 X5 Y1 Employment Proximity (Ƶ 2) 0.11 Literacy ( 1** ) −0.10 1.000 ** 1.000 1.000 Covariates status Literacy (Ƶ ( Employment 31)) ** −0.20 −0.236 −0.183 −4.420 *** −0.150 −0.116 3.044 ** LiteracyProximity ( −0.50 1) (Ƶ21.000 ) −0.10 0.11 ** ** 1.000 1.000 Length status Proximity of stay (Ƶ 32)) ( Employment ** −0.130 −0.100 −2.762 ** −0.077 −0.059 1.725 Proximity (Ƶ ) 0.11 −0.09 **−0.20* 1.000 0.31 ** 1.000 (Ƶ4) of stay Length 2 ** −0.50 −0.10 ** 1.000 Employment Employment status Status (Ƶ ( 33)) ** −0.20−0.09 * −0.054 0.31 ** −0.102 1.000 Measures −2.685 ** −0.030 −0.057 1.643 (Ƶ4) Length of ** stay −0.50−0.10 ** Independent 1.000 Length ofstatus Perceived Stay (Ƶ ( 34)) ** −0.001 −0.09 * −0.028 0.31 ** 1.000 −0.660 0.001 0.020 0.522 Length (Ƶ−0.07 of stay 4) −0.13**** −0.50 −0.07 Independent −0.04 Measures 1.000 benefits Independent Perceived(X1) Measures −0.09 * 0.31 ** 1.000 Land benefits(Ƶ4) −0.07 ** −0.13 ** −0.07 −0.04 Independent 1.000 Measures benefits Perceived (X1) Perceived(X1) −0.15 0.278 0.143 3.964 *** Ownership 0.12 ** 0.20 ** −0.07 −0.07 −0.13 ** Independent −0.07 0.04 Measures −0.04 1.000 1.000 Land benefits Land(X2) Ownership (X2) (X1) ** 0.080 0.109 3.047 ** Perceived −0.15 Ownership Land 0.12 **−0.070.20−0.13 ** **−0.07−0.07 −0.04 0.041.000 1.000 benefitsAwareness Conservation Conservation (X1) (X3) ** −0.15 0.100 0.071 1.954 * (X2) Ownership 0.12 ** −0.13*0.20 ** −0.07 0.04 1.000 Awareness Local Land Benefits (X4) −0.06 −0.06 0.01 0.23 ** ** 0.25 ** 1.000 0.251 0.272 7.213 *** Conservation (X2) 0.12 ** * −0.07 −0.15 (X3) Ownership 0.20 **−0.13* 0.04 1.000 Awareness Resource Use (X5) Conservation −0.06 −0.06 0.01 ** 0.23 ** 0.25 ** 1.000 0.027 0.032 0.910 (X2) * −0.13* (X3) Awareness −0.06 −0.06 0.01 0.23 ** 0.25 ** 1.000 R2 Conservation 0.046 0.222 (X3) −0.13* * Adj. R2 Awareness −0.06 −0.06 * 0.0410.01 0.23 ** 0.25 ** 1.000 0.213 (X3) F Value 8.854 *** 23.281 *** df 4738 9733 ∆R2 0.177 ∆F 33.273 *** VIF
You can also read