Promoters and detractors: Exploring what horse owners appreciate in equine veterinary practice

Abstract

In the field of veterinary medicine, it remains unclear what clients find essential in order to remain satisfied with their veterinarian’s services. Moreover, it is argued that clients’ expectations may differ depending on the kind of animal they own, and that aspects like communication are vital for the client-veterinarian relationship. However, there is a lack of knowledge on what clients expect from their veterinarian, which can result in dissatisfaction with the provided service. In other words, there is no theoretical framework that describes what skills or knowledge a veterinarian should rely on in particular scenarios with a client. To fill this research gap, I performed several data analyses through natural language processing on the data from a survey taken by horse owners from the Netherlands and the United States. In this study, I conduct topic detection analysis on several open ended questions from a survey with the goal to identify what clients appreciate in their veterinarian, including the potential reasons why a client may leave a veterinarian’s service. First, the data is pre-processed to eliminate dispensable words to optimise the analysis. Second, an exploratory data analysis is conducted in order to get an overview of the participants that were recruited for the study, such as their demographics, the purposes for which they have horses in their care, and the number of times they require a veterinarian’s services. Next, I use term frequency-inverse document frequency (TF-IDF) to create word clouds in order to get a general overview of trends in the data. Furthermore, k-means clustering and the topic modeling method named Latent Dirichlet Allocation (LDA) are used to identify the most essential topics and themes from the data. Lastly, these clusters were manually evaluated in order to detect the themes present in the data. Through affinity diagramming and colour coding all the words in the clusters, I constructed frameworks according to Elte et al.’s [31] categories, which describe the varying kinds of client requirements and to what extent they overlap. I highlight several limitations of my chosen analysis methods. Additionally, I pose several future research directions based on the study’s findings.