Questionnaire Application
The survey was administered using the Google Forms® platform, and farmer recruitment was conducted exclusively online. Survey links were distributed via the WhatsApp® messaging platform and through an email distribution list sourced from the database of the Study, Research and Extension Group in Animal Biometeorology at the Federal University of Uberlândia.
Before proceeding to answer the survey, each farmer was required to read and confirm their agreement with the provided consent form. The consent form explicitly outlined the study’s objectives and informed farmers that all collected data would be used exclusively for technical-scientific purposes, with strict adherence to principles of confidentiality and personal data protection. Additionally, only responses from farmers aged 18 and older who were actively engaged in dairy farming with lactating cows were considered for analysis.
Study Site Description
The study area is located in Southeast and Midwest of Brazil between latitudes 15°46’48” S and 23°32’56” S and longitude 41°3’59" W and 51°54’32” W (Figure 1). The farmers were invited to answer an online survey from February 2022 to September 2022.
Survey Instrument and Methodology
The questions were classified as structured/semi-open, that is, structured by presenting questions with pre-qualified answers (closed questions), and semi-open due to the closed answers that had the item “others”, which allowed explanatory observations when necessary. The questionnaire had 16 questions (see supplementary file) divided into three sections. Section 1 addressed issues related to farmers’ profile. In Section 2, information was requested regarding the characterization milk production, categorized by farm size, whether milk is the main source of income, daily volume of milk production, number of cows in lactation, dairy breeds, type of milking, milk destination and the availability of technical assistance. In Section 3, questions were raised concerning farmers’ knowledge of cattle behaviour and attitudes toward brushes for dairy cows.
In relation to section 3, knowledge was defined as the ability to comprehend and discern the topic, and attitude as the individual’s response to this topic (De Oliveira et al. 2020). Farmers’ knowledge was considered in terms of their understanding of animal behaviour related to herd fights, injuries caused by grooming habits, the need to reduce stress, and their familiarity with a brush for dairy cows. Their interest in installing the low-cost stationary brush was used as an indicator of attitude.
Eucalyptus-sisal Low-Cost Stationary Brush
The eucalyptus-sisal low-cost stationary brush for cows was made with an 8 to 10 mm sisal rope, which is wound around eucalyptus posts with a diameter ranging from 8 to 10 cm, and a height of 2.2 m, buried at a depth of 0.70 m. The sisal rope was fixed at the pole with staples, extending from the top to a height of 50 cm above the ground (Figure 2, A and B).
Statistical Data Analysis
Out of a total of 110 farmers, two participants did not agree with the consent terms, one did not answer all the questions, and one farmer resided outside the study area; therefore, they were excluded from the analysis. As a result, a sample of 106 completed questionnaires was considered for analysis, in accordance with the guidelines provided by Bartlett et al. (2001) for determining sample adequacy.
Sample size (n) was calculated using the equation:
With the responses obtained from the questionnaires, a descriptive analysis was conducted using absolute and relative frequencies.
To identify the questions associated with farmers' knowledge and interest in brushes, the multiple logistic regression model was applied. Logistic regression is a statistical model used when the dependent variable (or response) is binary qualitative, taking on values of 0 or 1, where 0 represents the absence of the event, and 1 represents the occurrence (Bonney 1987). Considered each individual variable in Section 3 as the dependent variable for model generation, while the variables from Section 1, 2 and 3 were treated as independent variables.
This model estimates the angular coefficient of an event happening based on one or more independent variables. It is represented by the following equation: