Data mining provides insights into views on horse and donkey welfare


A small international survey used data mining to evaluate people’s perceptions around the welfare of equines, small ruminants and turkeys.

Although the survey was carried out on a modest scale – too small to be representative – the Italy-based research team said text mining analysis seemed to be a promising way to investigate perceptions of animal welfare.

They said it emphasized the real perception of respondents, without the constraints deriving by close-ended questions.

Emanuela Dalla Costa and colleagues focused on horses, donkeys, sheep, goats and turkeys for their research.

The study team, writing in the open-access journal Animals, noted that people are increasingly sensitive regarding the way animals are kept and handled.

For each species, a survey comprising 14 open-ended questions was carried out online, with the target population being farmers, veterinarians, and animal owners.

A total of 270 surveys were collected from respondents from 32 different countries, using five different languages. Roughly a third of the survey results related to horses. They were mostly private owners and trainers.

When asked what horses required to be fit, healthy and productive, the data mining revealed the following words were used most frequently, in descending order: Feeding, water, shelter, exercise, freedom, care, proper training, health, pasture, space, company and forage.

For donkeys, the words were shelter, feeding, water, company, management, space, care and deworming.

The words “feeding” and “water” were the most frequently used in relation to all species.

It seemed that respondents generally considered the welfare principle of “good feeding” as the most relevant to guarantee fitness, health, and production.

Horse stakeholders used the word “forage” to highlight the need of fiber in the horse diet.

“Interestingly, for donkeys, the word ‘deworming’ was used,” the study team noted.

“The welfare principle ‘appropriate behavior’ was addressed by equine and turkey stakeholders using the words ‘company/social’; for horses, ‘proper training’ was also mentioned.

“Stakeholders of different species, except horses, frequently used the word ‘management’, which is a general term, not clearly linked with a specific welfare principle.”

In another question, respondents were asked: “Looking at your neighbor’s horse, which signs would you observe to assess accommodation, feeding, health, manifestation of normal and abnormal behavior?”

The most frequent word used by horse stakeholders to evaluate the conditions of accommodation “clean”, followed by “bedding”, “box”, “shelter” and “water”.

The text mining association showed that the term “clean” was associated with words referring to resource-based measures such as floors, mangers, drinkers, wells and water.

Discussing their findings, the study noted that, together with “feeding” and “water”, stakeholders recognized that providing “forage” is also necessary for horses. “The recognition of the importance of foraging is shown also by the use of other words such as ‘pasture, ‘grazing’, ‘roughage’, and ‘grass’.

“Horses naturally spend the majority of their time grazing, and it is reported that foraging food with high fiber content is not only a physical need (80% of the time, the gastric tract should be filled), but also a psychological need for them.

“In terms of signs to evaluate feeding conditions, horse stakeholders referred mostly to resource-based indicators, emphasizing how the quality of feeding (‘quality’, ‘clean’, ‘fresh’, and ‘appropriate’) is important.

“Nutrition is a critical component of horse health, and it is reported that good quality of feeding could be a reasonable target for decreasing the prevalence of pulmonary diseases in horses.

“Words referring to the principle ‘good housing’, such as ‘shelter’, ‘space’, ‘clean’, ‘air’, and ‘bedding’, were frequently used by stakeholders of different species.”

The authors said that, with a larger sample, it would be possible to undertake a more detailed analysis of welfare perception variations among species and to explore underlying reasons for these differences.

The full study team comprised Dalla Costa, Vito Tranquillo, Francesca Dai, Michela Minero, Monica Battini, Silvana Mattiello, Sara Barbieri, Valentina Ferrante, Lorenzo Ferrari, Adroaldo Zanella and Elisabetta Canali.

Text Mining Analysis to Evaluate Stakeholders’ Perception Regarding Welfare of Equines, Small Ruminants, and Turkeys
Emanuela Dalla Costa, Vito Tranquillo, Francesca Dai, Michela Minero, Monica Battini, Silvana Mattiello, Sara Barbieri, Valentina Ferrante, Lorenzo Ferrari, Adroaldo Zanella and Elisabetta Canali.
Animals 2019, 9(5), 225;

The study, published under a Creative Commons License,  can be read here.

Leave a Reply

Your email address will not be published. Required fields are marked *