I research learning and the development of expertise in workplace and university contexts. At the core of seemingly diverse fields of study, such forensics, medicine, the natural sciences, is the process of learning to generalise from one instance to another; a new example of that identity, disease, biological species, geological structure or abstract concept. To become an expert with the categories at hand, novices must learn their basic relational structure—the commonalities or family resemblances that persist across many different instantiations. Understanding how people generalise, or transfer knowledge from their prior experiences to new instances, is fundamental to designing effective learning methods in university and work-based settings. The ultimate goal of my research is to develop a domain-general theoretical framework for efficiently creating expertise and facilitating transfer of learning.
Figure 1. Three family resemblance categories: birds, bees, and fingerprints.
Developing Expertise with Fingerprints
Much of my research to date has focussed on understanding the development of expertise in fingerprint identification: a domain where human examiners have the critical task of identifying people in connection with crime on the basis of their visual discriminations of fingerprint evidence. More specifically, fingerprint examiners compare pairs of fingerprints side-by-side and judge whether they belong to the same finger (e.g., Smith’s right thumb) or two different fingers (e.g., Smith and Jones’s right thumb). Prints from the same finger can look very different, due to variation in surface, positioning, pressure, movement, and moisture as a print is left behind. Conversely, prints from different people often look alike, due to the use of computer algorithms that help speed up the search for similar candidates. As a result, the task of distinguishing matching from non-matching fingerprints poses a real challenge for inexperienced novices, and experienced experts are not infallible.
We’ve compared the performance of fingerprint experts and novices on a series of cognitive tasks (Searston & Tangen, 2016, 2017), and tracked the performance of fingerprint trainees as they gained experience working in a fingerprint unit (Searston & Tangen, accepted). The results of our research suggest that fingerprint expertise is marked by an ability to retrieve and flexibly use information distributed across prior cases. That is, fingerprint experts appear to make use of their vast experience with how prints tend to look and vary to help resolve novel cases. We found, for example, that fingerprint experts are significantly more accurate than novices at detecting structural or stylistic information across a person’s fingers (Searston & Tangen, 2016; blog post). Fingerprint expertise also develops with, and is constrained by, experience in the domain, but it is surprisingly flexible to changing task demands (Searston & Tangen, 2017).
I am now pivoting from researching the development of expertise to researching how best to create expertise. As a basic first step in this direction, we’ve examined the benefits of feedback, contrastive, and elaborative learning methods for learning to discriminate fingerprints (Searston & Tangen, 2017). Further developing an empirical basis for how best to create expertise will help pave the way towards more reliable forensic evidence and help to guard against miscarriages of justice.
Developing Expertise with Natural Categories
A natural next step is to extend my research in the domain of fingerprints to other domains. For instance, I am currently working to better understand how people learn natural science categories in the context of higher education. How does testing, comparison, and elaboration focus learners on the most diagnostic and discriminating dimensions of the categories at hand? What are the benefits of providing learners with prototypical, definitional or rule-based information? What is the best way to present examples to learners for improved generalisation? Understanding how to most efficiently develop expertise across disciplines will provide a general empirical foundation from which we can draw on to enhance learning across university and work-based learning environments.