We followed descriptive survey research design and used interview schedule as the instrument of the current research. The respondents of the study were the farmers who practiced floating agriculture. We purposively selected two Union each of the two Upazila (smaller administrative unit of Bangladesh), namely, Banaripara and Wazirpur, Upazila of Barisal District, as the study area. The selected Unions were Bisarkandi and Udaykati of Banaripara Upazila; and Satla and Otra of Wazirpur Upazila, respectively. All the farmers who practiced floating agriculture at the study areas (ie., four Union of two Upazila) were the population of the study. The total number of farmers practicing floating agriculture was 200 (Banaripara) and 185 (Wazirpur), that means a total of 385 which were the population of the study. Among the population, a total number of 140 farmers (75 from Banaripara and 65 from Wazirpur) were selected as the sample utilizing the Equation 1 developed by Kothari and followed by Hasan et al.. 2.1.1 Measurement of dependent variable The opinion of the farmers towards floating agriculture as a means of cleaner production is the dependent variable of this study. The interview schedule contains 20 statements which were administered for judging the farmers opinion. The opinion statements were aligned with the three areas of cleaner production including, environmental, economic and social and cultural areas. The farmers were asked to indicate the extent of their agreement on each of the 20 statement utilizing a Likert-type fivepoints scale like strongly agree, agree, undecided, disagree and strongly disagree with assigned scores of 5, 4, 3, 2 and 1, for positive statements, respectively and vice versa for negative statements. Different scales are used for measuring opinion of the respondents, although the Likert scale is the most widely utilized technique for opinion measurement. The Likert-type scales utilize fixed type and close form of responses to measure opinion or attitude. For conducting the present study, we employed five point Likert-scale and according the respondents were asked about their agreement or disagreement of each of the statements. 2.1.3 Measurement of independent variables There were eight independent variables of the study and those were farmers’ age, level of education, family size, farm size, family annual income, extension media contact, training participation on floating agriculture and knowledge on floating agriculture. Age of a respondent was measured by counting the years from the time of his/her birth to the time of interview. The level of education was measured by the number of years of schooling. Family size was measured by the total number of members including the respondent himself, spouse, children and other permanent dependents who lived together as family unit. The farm size possessed by the farmer under farm including share cropping and leased and homestead was the basis of measuring farm size and which was expressed in hectare for the current study. Family annual income of a respondent was determined on the basis of his total earnings from agriculture, service, business, and other sources. For measuring extension media contact of the respondent, a four-point scale i.e., not at all, rarely, occasionally and frequently was used and appropriate weights were assigned to quantify the variable as against five different types extension media and assigned scores were 1, 2, 3, and 4, respectively. Training participation on floating agriculture was measured by the total number of days that a respondent had encountered training experience in his entire life from different agricultural related organizations and from other organizations o floating agriculture. Meanwhile, the farmers’ knowledge on floating agriculture was calculated by answering 15 questions related to floating agriculture. The assigned score against each correct, partially correct and incorrect answer was 2, 1, and 0, respectively. 2.2 Statistical Analysis We utilized Statistical Package for Social Science (SPSS) version 16 for analyzing the data of this study. We calculated the mean and standard deviation to achieve the objectives of the study and used different categories for classifying the data. Different statistical tests like frequency count, percentage, mean, and standard deviation were applied to analyze and interpret the data based on the purpose of the study.