I recently shared a post on social media about the findings of a study that I had co-authored about how horse owners determine if their horses have their social and behavioural needs met. Just over half of the participants had strongly agreed that the horses in their care had their social and behavioural needs met completely. The bulk of the paper was about how people justified their beliefs.

In one comment on my post, I was told that our time would have been better spent evaluating how healthy and happy the horses actually were.

The critique reflects a one-sided understanding of research but more importantly, it highlights a generalised cultural bias in favour of quantitative research that can also be identified in horse-related research. However, we should learn to appreciate and evaluate qualitative research on its own terms.

As I will explain in this article, this critic was comparing apples and oranges (quantitative and qualitative research) but when it comes to the relative value of different forms of research, it’s a matter of horses for courses.

Discussions about quantitative and qualitative research should not be about which one is more scientific, truthful or ‘better’. They should be about which approach is most appropriate.

Quantitative vs qualitative research

Qualitative and quantitative research are both based on data and can be conducted in a systematic and rigorous manner. But they differ according to their underlying philosophies, the forms of data that are valued and the criteria by which the findings are evaluated.

Equitation Science has gained rapid popularity in the past two decades. It is firmly grounded in what is known as a positivistic view of truth and knowledge, which is fairly synonymous with experimental science.

Experimental science is founded on several assumptions. The first is that there are real facts which exist independent from human experience and can be revealed through experimental techniques such as observation, definition and measurement.

These techniques further serve to remove any human influence and, therefore, come closer to an unbiased view of the world. The facts revealed by the experiment can be reduced to variables which can be controlled and manipulated for the purposes of gaining more knowledge.

This is why experimental science has developed many techniques for removing human bias (subjectivity), identifying individual units of behaviour and determining statistical significance (the measure of how reliable the finding is).

For example…

In 2014, Merkies and colleagues carried out an experimental study to determine if and how horses respond to stressed humans, and if horses differentiate between humans who are physically or emotionally stressed.

The experimental setting was a round pen where horses were exposed to humans from one of three groups: humans who were calm around horses, humans who were stressed by the presence of horses and humans who were stressed from doing a period of physical exercise. The humans were stationery and blindfolded.

Heart rate monitor data were used to measure fear in humans and stress in horses. Horse stress was also inferred from behavioural markers such as their speed, the height of their head relative to their withers, and where and how they stood in relation to the stationery humans. Some horses had their data collected in a round pen with no human.

These horses formed the ‘control group’ against which to benchmark the other three experimental groups. Numerical calculations were used to determine differences between the groups and identify which differences were statistically significant.

The researchers concluded – unexpectedly – that horses were less stressed in the presence of an emotionally or physically stressed human than a calm person.

Qualitative research

Qualitative methods have been designed to study all those things which do not lend themselves to observation, measurement or control – but which are equally important. As Einstein is reported to have said, ‘not everything that can be counted counts, and not everything that counts can be counted’.

The things that count for qualitative researchers are thoughts, feelings, attitudes, beliefs and behaviours that are relational and occur in complex environments (outside of laboratories), and which make most sense in non-numerical terms.

Qualitative research is grounded in a philosophy referred to commonly as social constructionism. In contrast to positivism, social constructionism asserts that knowledge about the world is inseparable from us as human knowers.

Therefore, specific qualitative research techniques have been developed to determine how humans think and talk about the world – and attribute meaning to their existence.

These techniques can be observational, but they frequently involve the researcher engaging with, talking to or ‘hanging out’ with people in interview, focus group or participant-observation studies, relatively.

Social constructionists might point out how the knowledge produced from the round penning experiment described above arose from human initiation, design and interpretation. However, they would not ask questions about the ways in which horses and humans influence one another in quite the same way.

Brandt, for example, interviewed women about how they related to horses. Women spoke to her about how they sense and make sense of horses. One participant noted how her knowledge of her horse’s emotional state could not be attributed to any observable signals, but was something she sensed through her own body. Another woman described how on the occasions where she overpowered her own reactions of fear, her horse’s nervousness would subside.

In this qualitative study, what women said was taken at face value. It didn’t matter that heart rate had not been measured, that data were women’s retellings of events that had already happened and which could not be verified or reduced to discrete variables to be manipulated in scenarios.

There was no need to ask if what was being discussed was real, because as a piece of qualitative research, the emphasis was on representations of truth and human-horse interrelations – all of them things which don’t make much sense in numerical terms or fit into statistical calculations.

The women spoke to Keri Brandt about things that they may not otherwise have given any conscious thought. They may have formed their opinions simply because they were responding to an interview question.

Qualitative researchers are aware of their impact on the data they produce and collect – and consequently, they take their subjectivity into consideration when interpreting their findings. They do not seek to remove themselves from data collection or control the topics of their research, as this would be fundamentally inconsistent with their social constructionist philosophy that humans cannot be separated from the knowledge we produce. In fact, Brandt’s own identification as a horse-rider (and a woman) may have made women more comfortable talking to her about their feelings and emotions.

Conversely, women may not have fully elaborated to Brandt because they may have assumed she already had shared knowledge. This is why qualitative research generally entails lengthy interactions with people to build the rapport that is necessary to identify research biases.

So, what does all this mean for research on (and with) humans and horses?

Different kinds of research questions suit different ways of understanding and asking about the world.

There are two main ways of conducting research – simplified into quantitative and qualitative research.
Some questions are best answered by quantitative research methods, such as the kinds of questions that are asked by equitation scientists. This includes determining the social and behavioural needs of horses.

However, there are other questions which would be entirely inappropriate to be addressed by quantitative methods, such as our consideration of how horse owners make decisions about whether their horses have their social and behavioural needs met.

The choice of which kind of science is ‘best’ is like comparing apples with oranges. What makes for a good apple does not necessarily make a good orange.

Many people are familiar with the criteria for judging the quality of experimental research. These are:

  • repeatability (would we get the same findings if we did the research again?),
  • generalisability (the extent to which findings from one study can be generalised to analogous behaviours and interactions across time, space and place),
  • reliability (the extent to which you trust your measures to capture the phenomena you are studying), and
  • validity (how much the data you measured can stand in for the topic of study, e.g. can heart rate stand in for the thing we call fear or nervousness).

Qualitative research, on the other hand, should be evaluated against criteria that is relevant to a social constructionist philosophy. These are usually credibility and trustworthiness (are the interpretations reasonable, plausible and supported by the data?)

Back to the future

The aim of the study mentioned at the start of this article about how horse owners determine if their horses have their social and behavioural needs met, was not to report objective measures of horse health. That would have involved a very different methodology with observation, measurement and testing.

Instead, we drew from an online survey which collected both qualitative and quantitative data through closed and open questions, respectively.

It is one thing to know ‘how many’ or ‘how much’ or ‘how often’, but those numerical answers are often made much more meaningful and useful when we also know ‘why’ and ‘why not’.

Our analysis of the common reasons why horse owners believed their horses were happy and healthy provided credibility and trustworthiness to our research findings and interpretations.

An equitation scientist can use quantitative research methods to determine the appropriate way for a horse to have their social and behavioural needs met, or to develop a scoring system to determine how much this knowledge is influencing horse management practices – but these facts and tools alone are not enough to improve horse welfare.

Horse owners make sense of and act upon seemingly straightforward scientific facts in highly complex psychological, social and cultural worlds. An understanding of these worlds is essential for changing human behaviour, especially where it has direct impact on animal welfare.

This is exactly the kind of insight that a social scientist can provide using qualitative research methods. It is a matter of knowing when to offer the apple, how to judge the orange and where you would be better off making a fruit salad.

References: