Patrick Onghena is full professor of statistics and methodology at the Faculty of Psychology and Educational Sciences, KU Leuven (Belgium). After his doctoral and postdoctoral work on the statistical analysis of single-case data, he was appointed as a professor at KU Leuven, conducting research on single-case experimental designs and applied statistics, and teaching general methodology and statistics courses to students of educational sciences, psychology, and speech, language, and hearing sciences. He is the author of numerous articles about single-case experimental designs and applied statistics in international peer-reviewed journals and is a distinguished member of the Royal Flemish Academy of Belgium for Science and the Arts.
What attracted you to the field of single-case designs in the first place? Can you tell us about your first project using these methods?
I studied clinical psychology, was trained as a psychotherapist, and became very much interested in studying the processes and the effects of psychotherapy. What struck me in my psychotherapy practice was that meaningful relations between complaints, behaviours, experiences, memories, beliefs, and dreams could be discovered at the individual level without the need or desire to generalize these relations to other patients. However, many courses in my training took a group averaging approach to evaluate process and outcome of psychotherapy. That approach didn’t match my interest in individual human beings and my perspective on what helps in psychotherapy.
My first project was guided by my professor of statistics who knew about my interest in clinical case studies and single-case experimental designs. He challenged me to investigate the statistical power of randomization tests for single-case experimental designs. That challenge became the topic of my PhD dissertation (defended in 1994).
In your opinion, what are some of the exciting developments in the field of single-case designs that you have seen during your career? And that you see currently?
The most exciting developments in the field of single-case designs that I have witnessed during my career are the introduction of randomized single-case experimental designs and the possibility of causal inference at the individual level, the development of multilevel meta-analysis for replicated single-case experimental designs, the translation of behavioural methodology to medical settings and the reinvention of single-case designs as N-of-1 trials, the growing interest in mixed methods single-case research, and the research community’s endorsement of several authoritative guidelines for conducting and reporting single-case research.
Currently, I’m expecting a lot from developments in idiosyncratic assessment, how we can have repeated measurements that are demonstrably reliable and valid, sensitive to change, but robust to the confounding effect of the repetition itself, and merging Experience Sampling Methodology and Ecological Momentary Assessment with single-case designs. I’m also intrigued to see how the field is organizing itself, now it is becoming more mainstream, for example by the ICN or by the Small is Beautiful symposia.
What are some of the myths about single-case designs you have come across? Where do you think these myths come from and do you think they can be addressed?
There are many myths, but the most prominent myth is that single-case designs and N-of-1 trials are, by definition, less rigorous than group-comparison designs. A related myth is that you need group comparisons to make a causal claim, and hence that a group-comparison design is the design of first choice if you have a research question about an intervention’s effectiveness. In this myth, (a series of) single-case designs and N-of-1 trials are just a plan B because of lack of resources or for other pragmatic reasons. These myths miss the crucial point that single-case designs and N-of-1 trials have their own theoretical and methodological rationale. For example, N-of-1 RCTs have a different (complementary) conceptual definition of what constitutes “an effect” than group-comparison RCTs, and single-case designs focus on intraindividual variability while group-comparison designs have interindividual variability as their basis.
The origins of these myths are difficult to pinpoint and probably depend on the specific field of application. In psychology, these myths are partly tied to the successful scientific investigation of interindividual differences and to the eugenics movement in the beginning of the 20th century. Furthermore, the widespread adoption of between-groups analysis of variance designs borrowed from agricultural research set the stage for an experimental behavioral science in which human beings are the basic independent experimental units that can be used in the calculation of averages and standard deviations. However, although the average yield and its standard deviation in an agricultural field trial have an intrinsic meaning for food supply or food quality, the meaning of an average score or an average score difference calculated across participants is not always intrinsically meaningful for the psychology of individuals (or the effect at the individual level).
Whatever their origins, the myths are transmitted and consolidated by our teaching, training, and mentoring of young researchers. Consequently, in my opinion, they can also be debunked in our teaching, training, and mentoring. Single-case designs and N-of-1 trials shouldn’t be considered as the new holy grail but, alongside group-comparison designs, they deserve a place in our methods and statistics courses, and in our rigorous scientific practice.
What are some of your most recent applications of single-case designs, for example, which interventions have you been testing with single-case designs?
Recently, I have been involved in single-case research investigating the effect of parenting training on parenting behaviour and parental stress in mothers with borderline personality disorder, testing a digital behavioural health treatment for chronic pain, a study of the role of cognitive-behaviour therapy and gaming in decreasing anxiety in children with Autism Spectrum Disorder, and a study on the efficacy of online Memory Specificity Training in adults with a history of depression.
Looking to the future, what are your predictions about future trends/breakthroughs for the field of single-case designs?
In the near future, I expect major developments in the areas of idiosyncratic assessment and data analysis for single-case designs. As mentioned before, Experience Sampling Methodology and Ecological Momentary Assessment have a lot to offer to single-case methodology, and the wider availability of wearables and other digital technologies can foster the integration between single-case experimental designs, Experience Sampling Methodology and Ecological Momentary Assessment.
I also expect an expansion in the application of statistical techniques to data collected within single-case designs: machine learning, network analysis, time series analysis, multilevel models, Bayesian statistics, and distribution-free methods. Because data collected within single-case designs are just… data, we might expect that more and more statisticians and data scientists will get involved and offer promising new perspectives to analyze single-case data.
What do you think the field of single-case designs needs most?
The field of single-case designs needs interested and competent researchers who are leaders in their field and who can demonstrate the added value of single-case methodology to progress their field and substantive theory (e.g., as is happening in pain research at the moment). We need researchers who see the bigger picture and not blindly follow a prescribed routine for scientific discovery or justification. They can act as role models for the next generation of scholars.
How could the International Collaborative Network for N-of-1 trials and single-case designs make the most impact on the field?
I would encourage the ICN to continue its very worthwhile activities. Perhaps an effort could be made in the direction of obtaining more sustainable support for the ICN activities by applying for funding of this support. If there is more support, then the ICN could also expand its scope by establishing a scientific society or journal, or by associating with existing scientific societies or journals, or by organizing thematic international symposia that bring together researchers interested in N-of-1 trials and single-case designs from all over the world.