Abstract
Predictive risk modelling is a computational method used to generate probabilities correlating events. The output of such systems is typically represented by a statistical score derived from various related and often arbitrary datasets. In many cases, the information generated by such systems is treated as a form of evidence to justify further action. This paper examines the nature of the information generated by such systems and compares it with more orthodox notions of evidence found in epistemology. The paper focuses on a specific example to illustrate the issues: The New Zealand Government has proposed implementing a predictive risk modelling system which purportedly identifies children at risk of a maltreatment event before the age of five. Timothy Williamson’s (2002) conception of epistemology places a requirement on knowledge that it be explanatory. Furthermore, Williamson argues that knowledge is equivalent to evidence. This approach is compared to the claim that the output of such computational systems constitutes evidence. While there may be some utility in using predictive risk modelling systems, I argue, since an explanatory account of the output of such algorithms that meets Williamson’s requirements cannot be given, doubt is cast upon the resulting statistical scores as constituting evidence on generally accepted epistemic grounds. The algorithms employed in such systems are geared towards identifying patterns which turn out to be good correlations. However, rather than providing information about specific individuals and their exposure to risk, a more valid explanation of a high probability score is that the particular variables related to incidents of maltreatment are just higher amongst certain subgroups in a population than they are amongst others. The paper concludes that any justification of the information generated by such systems is generalised and pragmatic at best and the application of this information to individual cases raises various ethical issues.
Notes
We could talk about assertions-of-beliefs which might be construed as knowledge-of-mental-states rather than knowledge-of per se. In this paper, however, I want to restrict our talk about knowledge to a more rigorous epistemic account that might satisfy modern epistemologists and that accounts for Gettier-like problems.
And even to such basic assumptions we must still be entitled to enquire ‘Why would you assert that the sun will rise tomorrow?’.
By nominal living conditions, I mean individuals or families that live under low socio-economic conditions. Such living conditions may exhibit characteristics such as overcrowded and or insufficient housing, poor education, low income, unemployment and so on. These characteristics are typical of the kind of risk variables modelled within the system in question. It should be noted that these factors are typically considered as being stereotypical in relation to child abuse.
I thank Dr. Brian Ballsun-Stanton for pointing this out to me.
Because this would be circular, traditional epistemologists, prior to the Gettier argument want to argue that justified true belief constitutes knowledge.
References
Ferraris, V., Bosco, F., Cafiero, G., D’Angelo, E., & Suloyeva, Y. (2014). Defining Profiling: Working Paper. From http://www.unicri.it/special_topics/citizen_profiling/WP1_final_version_9_gennaio.pdf
Goldman, A., & Beddor, B. (2015). Reliabilist epistemology, The Standford Encylopedia of Philosophy. From http://plato.stanford.edu/entries/reliabilism/
Hempel, K. (1962). Deductive-nomological vs. statistical explanation. In H. Feigl & G. Maxwell (Eds.), Scientific explanation, space and time, vol. 3, Minnesota studies in the philosophy of science (pp. 98–169). Minneapolis: University of Minnesota Press.
Hempel, K. (1965). Aspects of scientific explanation and other essays in the philosophy of science. New York: Free Press.
Kelly, T. (2014). Evidence. In E. N. Zalta (Ed.), The Stanford Encyclopedia of philosophy (Fall 2014 Edition). From http://plato.stanford.edu/archives/fall2014/entries/evidence.
Kim, J. (1988). What is naturalised epistemology? In J. Tomberlin (Ed.), Philosophical perspectives 2, epistemology (pp. 381–405). Atascadero: Ridgeview Publishing Co.
Laplace, P.-S., Marquis de. [1820] 1951. A philosophical essay on probabilities (trans: Truscott, F.W. and Emery, F. L.). New York: Dover Publications.
O’Connor, D. (2008). Attributions and cognitive closure; stereotypes of perpetrators and victims of child sexual abuse. All Volumes (2001–2008). Paper 13. From http://digitalcommons.unf.edu/ojii_volumes/13
Panattoni, L. E., Vaithianathan, R., Ashton, T., & Lewis, G. H. (2011). Predictive risk modelling in health: options for New Zealand and Australia. Australasian Health Review, 35(1), 45–51.
Salmon, W. (1989). Four decades of scientific explanation. Minneapolis: University of Minnesota Press.
Salmon, W. (1998). Causality and explanation. New York: Oxford University Press.
Striphas, T. (2015). Algorithmic culture. European Journal of Cultural Studies, 18(4–5), 395–412.
Talbott, W. (2008). Bayesian epistemology. In E. N. Zalta (Ed.), The Stanford Encyclopedia of philosophy. From http://plato.stanford.edu/archives/sum2015/entries/epistemology-bayesian
Vaithianathan, R., Malony, T., Jiang, N., De Haan, I., Dale, C., Putnam-Horstein E., & Dare, T. (2013). Vulnerable children: Can administrative data be used to identify children at risk of adverse outcomes? Centre for Applied Research in Economics, Auckland University. Retrieved from https://www.msd.govt.nz/documents/about-msd-and-our-work/publications-resources/research/vulnerable-children/auckland-university-can-administrative-data-be-used-to-identify-children-at-risk-of-adverse-outcome.pdf
Williamson, T. (2002). Knowledge and its limits. Oxford: Oxford University Press.
Woodward, J. (2014). Scientific explanation. The Stanford Encyclopedia of Philosophy. From http://plato.stanford.edu/entries/scientific-explanation/#WhaDoStaTheExp
Zdziarski, J. (2014). The importance of forensic tools validation. From http://www.zdziarski.com/blog/?p=3112
Acknowledgements
I wish to acknowledge and sincerely thank Dr Brian Ballsun-Stanton who read a draft of this paper and in true ANZAC and academic spirit made many inspiring and instructive comments. I also thank the two reviewers who challenged me to clarify and think deeper about many of the issues discussed in this article. Any remaining errors are mine alone.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Mckinlay, S.T. Evidence, Explanation and Predictive Data Modelling. Philos. Technol. 30, 461–473 (2017). https://doi.org/10.1007/s13347-016-0248-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13347-016-0248-9