Background of the study site To date, 37 PAs have been notified in Bangladesh,3 under the International Union for Conservation of Nature’s (IUCN) protected area management categories II and IV, covering nearly 11% forest area of the country (Dudley (2008). For the present study, we purposively selected Satchari National Park considering its particular richness in biodiversity, convenient access from the capital city Dhaka and nearby Sylhet , and its broad representativeness of the mainstream PAs of the country (Uddin et al. 2013). The park is also one of the PAs in which a co-management approach was pioneered in the country (Mukul et al. 2014). The word ‘Satchari’ comes from “seven streams” (locally called ‘chara’) and refers to the streams that flow through the forest. The area of the park is 243 ha which comprises of the forests of Raghunandan Hills Reserve within the Satchari Range (Uddin and Mukul 2007).. Administratively the park is located in Chunarughat Upazilla (administrative entity; subdistrict) of Habiganj district and is situated nearly 130 km north-east of capital city Dhaka. The vegetation of the park is ‘evergreen’, however large-scale conversion of the indigenous forest cover to plantations has changed its classic forest type entity (Uddin et al. 2013). Now only 200 ha of land contain natural forests; the rest is secondary raised forest. It is also one of the last habitats of critically endangered Hollock gibbons (Hoolock hoolock) in the country as well as in the sub-continent. The park has an undulating topography with slopes and hillocks ( locally called tilla) ranging from 10 to 50m in elevation. A number of small, sandy-bedded streams drain the forest, all of which dry out in the winter season after November. The annual average rainfall is 4162 mm. July is the wettest month having an average of about 1250 mm of rain, while December is the driest with no rainfall. May and October, the hottest months, have an average maximum temperature around 320C, while January is the coldest month when the minimum temperature drops to about 120C. The relative humidity of the study area is about 74% during December while it is over 90% during July-August. Sampling protocol A total of 19 villages with varied degrees of interests in the national park as well as in the reserved forest (RF) has been identified (Mollah et al. 2004). Of these villages, one village is located within the national park area and is inhabited by an ethnic community, the Tripura tribe. The other settlements that have stakes in the national park are located about 3-8 kilometers away from the park. Table 1 lists the degree of dependency that the various villages have on the RF as well as on the national park. Local people have traditionally collected various resources from the national park and adjacent RF. Many households, particularly poor households from the 19 villages studied, rely entirely or partially on the RF as well as on the park for collecting firewood, timber, bamboo, herbal remedies, and other NTFPs. Data collection We collected both quantitative and qualitative data. We randomly selected one village from each of the first four ‘forest dependence categories’ as identified by Mollah et al. (2004), including the only village inside the park, Tiprapara. We did not consider any village with only a minor degree of dependence on the PA and adjacent RF. After preliminary field visits and a reconnaissance survey, however, we performed a necessary ranking change between Deorgach and Ratanpur, given the fact that the forest is more critical to the people of Ratanpur than to the people of Deorgach. We conducted household survey in four sample villages—Tiprapara, Ratanpur, Deorgach, and Goachnagar—within a one-year time span (January to December, 2006). Before household surveys, we arranged some focus group discussions to construct a community map and a community profile of the study villages (Table 2). We also conducted a number of field visits to observe and verify the information recorded during our community mapping exercises. We interviewed 101 households in our four sample villages from a total households of 818 households within the survey villages.