The study was carried out purposively in the Jatholida village of Bogra district and Bhobokhali village of Mymensingh district. The study areas belong to Karatoya-Bangali Floodplain and Old Brahmaputra Floodplain, the two important Agro-ecological Zones (AEZs) of Bangladesh. Among the 30 AEZs there are some piedmont plain, tract (strip of high land), basin, haor, beel and some river floodplains, most of which are average productive due to soil reaction, hydrology and natural disaster. Haor is a bowl shaped vast area of water that becomes flooded in every monsoon. During rainy season it looks like a sea and the villages inside look like islands. It remains submerged for seven months. According to IUCN Bangladesh has about 400 haors. Beel is a large shallow depth water body that contains the additional water of rain and becomes dry during summer. During dry season it goes under cultivation and used as pasture to some extent. The selected river floodplains are highly productive and magnitudes of extension service providers are working in this area. The major crops of the study areas are different green vegetables followed by rice, jute, chilli, potato, eggplant, banana, mango, jackfruits, jujube, guava and litchi. As a result, the extension service is very important for the farmers of these areas. Another reason of the purposive selection of the study areas was close proximity and familiarity of the study site to the researchers. A list of 250 crop farmers from Jatholida village and 200 from Bhobokhali village was prepared with the help of the village mosque committee as they know the best about the occupation of the residents due to frequent interaction with them. These villages comprised crop farmers, fish farmers, livestock farmers, mixed farmers, day labourers and non-farm residents where crop farming is a dominant enterprise. However, 20% of them were selected randomly from the population of 450, as the sample of the study considering fund and time limitations of the researchers. Thus, the sample size was 90. In addition, six focus group discussions (FGDs) with crop farmers and 20 key informant interviews with academics, public and private extension staff and input dealers were executed to have deeper understanding about the research problem. Data were collected from the respondents by structured interview, in a face-to-face setting, in the year 2013, during winter and summer, the major crop seasons of Bangladesh. Summer is dominated with rice, summer vegetables and fruits. On the other hand winter is dominated with varieties of vegetables, some pulses, oil seeds and flowers. So, an equal proportion of the respondents were interviewed in both seasons to cover all types of crop farmers.
The independent variables of the study were ten socio-demographic characteristics of the respondent those are supposed to influence WTP for extension services. The variables were selected consulting relevant previous studies. The variables were age, gender, education, agricultural income, total income, proportion of the crop sold, farm size, farming experience, media contact, and innovativeness. Innovativeness is the degree to which an individual or other unit of adoption is relatively early in adopting new ideas than other members of a social system. On the other hand, adopter categories are the classification of the individual based on innovativeness (Rogers 1995). Appropriate scales were developed to assess age, gender, education, agricultural income, total income, proportion of the crop sold, farm size and farming experience. Media contact behaviour of the farmer was measured through a four point scale: frequently (3), now and then (2), seldom (1), not at all (0), against the use of seven important extension sources like block agriculture office at village, sub-district agriculture office at Upazila, NGOs, ideal farmers, neighbours and peers, input dealers and others. Innovativeness of the farmers was measured based on five adopter categories of E. M. Rogers with their corresponding scores. Farmers were asked to indicate their adoption behaviour as follows: innovators (5): invent ideas and search for new agricultural technology to the centre of diffusion; early adopter (4): adopt new technology at first when it comes to community; early majority (3): adopt the innovation before majority of the community people; late majority (2): adopt innovation after majority of the community people have adopted it; laggard (1): always confuse new technology and rely on traditional technology.
The dependent variables of this study were farmers’ WTP for the agricultural extension services and the amount farmers are willing to pay. WTP for advisory services can be determined by two ways: (a) direct /CVM and (b) indirect/demand and supply estimation (Ulimwengu and Sanyal 2011). CVM can be of different types. For example, open-ended, referendum, payment card and bidding (Anonymous 2012). However, in this study open-ended CVM was used in a face-to-face situation. First of all, the study sought willingness of the farmers through asking a binary choice ‘yes’ for willing and ‘no’ for not willing. However, the maximum WTP was assessed following open-ended question shortly thereafter in order to get a hypothetical quantitative response. CVM is usually used to estimate economic values of all kinds of environmental services. However, CVM has also been used in measuring farmers’ WTP for the extension services by many authors (Gautam 2000; Sulaiman and Sadamate 2000; Ajayi 2006 and Budak, Budak, and Kacira 2010).
In CVM, people are directly asked how much they would be willing to pay for specific extension service. People’s response would be hypothetical. Therefore, the method naming entails contingent [www.ecosystem valuation.org/contingent_valuation. htm]. WTP assumed the value 1 for willingness of the farmers to pay and 0 for otherwise. Descriptive statistics were employed to analyse the demographic characteristics of the farmers while binary logistic regression model was used following Foti et al. (2007) and Anim (2008), to assess the factors that determine the willingness of the farmers to pay for the extension services. The logistic model was based on the supposition that the likelihood of willingness to pay, Pi depends on a vector of known variables (Xi) and a vector (β, coefficient) of unknown variable.