Remodeling Forensic Skeletal Age: Modern Applications and New Research Directions

By Jieun Kim

Remodeling Forensic Skeletal Age: Modern Applications and New Research Directions presents a comprehensive understanding of the analytical frameworks and conceptual approaches surrounding forensic age estimation and the current state of the field. The book also includes a series of recommendations of best practice through chapter-examples that offer theory and guidance for data acquisition, technique and/or model development, and the assessment of impact of the adopted approaches. Written by leading, international experts, the book's contributors provide an introduction, conceptual understanding and taxonomy of statistical frameworks and computational approaches, including the Bayesian paradigm and machine learning techniques for age estimation.

• Discusses core concepts in age estimation, along with key terminologies
• Presents tactics on how readers can generate sound models that can be translated into forensic reports and expert testimony
• Provides a step-wise approach and best practice recommendations for data acquisition, considerations in sampling, exploratory data analysis, visualization, and sources of error for appropriate and reproducible research design
• Includes examples, theory and guidance on how to develop models for age estimation and reviews the impact of population-specific and universal approaches

• Language: English
• Publisher: Academic Press
• Release date: Apr 22, 2021
• ISBN: 9780128243916

Book preview

9780128243916_FC

Remodeling Forensic Skeletal Age

Modern Applications and New Research Directions

First Edition

Bridget F.B. Algee-Hewitt

Senior Research Scientist, Center for Comparative Studies in Race and Ethnicity, Stanford University, Stanford, CA, United States

Jieun Kim

Assistant Professor of Anatomy, Lincoln Memorial University, DeBusk College of Osteopathic Medicine, Knoxville, TN, United States

Table of Contents

Cover image

Title page

Copyright

Contributors

Looking back and forward: An introduction—Defining, refining, and [re]modeling age estimation: A trajectory for forensic anthropology

The challenge of advanced aging: Synthesizing theory, method, and forensic practice

New contributions to aging problems: Chapter summaries

Final thoughts: Aging beyond this volume

Section A: Longstanding problems of the population

Chapter 1: Using data from the US Korean War Dead and the Terry Collection to demonstrate problems of the common overlap methods

Abstract

Acknowledgments

Material and methods

Results

Discussion

Chapter 2: Testing for differences in senescence using score data to understand the effects of reference sample choices

Abstract

Samples and data

Methodological approach and analytical framework

Testing for sample effects using a single age estimator

Testing for sex and race effects using multiple age estimators

Discussion

Section B: Aging across the ages

Chapter 3: Subadult age estimation variables: Exploring their varying roles across ontogeny

Abstract

Acknowledgments

Background

Statistical explorations of correlation and conditional dependence

Conclusions

Chapter 4: Aging the elderly: Does the skull tell us something about age at death?

Abstract

Introduction

Material and methods

Statistical analysis

Results

Results per variable

Discussion

Chapter 5: Population variation in diaphyseal growth and age estimation of juvenile skeletal remains

Abstract

Introduction

Samples

Research hypotheses and predictions

Analysis

Results

Discussion

Conclusion

Chapter 6: Great expectations: The rise, fall, and resurrection of adult skeletal age estimation

Abstract

Acknowledgments

Good age estimates

Conventional procedure results

Resulting misconceptions

Is there hope?

A path forward

Recommendations

Section C: Computational methods come of age

Chapter 7: A volumetric approach to age estimation informed by voxel selection: Application to the spheno-occipital synchondrosis

Abstract

Acknowledgments

Declaration of interest

Methods

Results

Discussion

Conclusions

Chapter 8: The consecutive inference of ancestry and age from shape measures of the pubic symphysis

Abstract

Introduction

Data acquisition

Implementation

Statistical analysis

Discussion

Section D: Classic indicators rejuvenated

Chapter 9: The fallacy of forensic age estimation from morphometric quantifications of the pubic symphysis

Abstract

Acknowledgments

Introduction

A theorem about inverse regression

The fallacy here perfectly describes pubic symphysis data

The only solution is to change the problem

Concluding remark

Chapter 10: An application of the Bayesian SanMillán-Rissech acetabular aging method to an African American sample: Preliminary results

Abstract

Acknowledgments

Introduction

Samples and methods

Results

Discussion

Index

Copyright

Academic Press is an imprint of Elsevier

125 London Wall, London EC2Y 5AS, United Kingdom

525 B Street, Suite 1650, San Diego, CA 92101, United States

50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom

Copyright © 2021 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN 978-0-12-824370-1

For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Image 1

Publisher: Masucci, Stacy

Acquisitions Editor: Brown, Elizabeth

Editorial Project Manager: Kuhl, Mariana L.

Production Project Manager: Raviraj, Selvaraj

Cover Designer: Rogers, Mark

Typeset by SPi Global, India

Contributors

John Albanese     Department of Sociology, Anthropology and Criminology, University of Windsor, Windsor, ON, Canada

Bridget F.B. Algee-Hewitt     Center for Comparative Studies in Race and Ethnicity, Stanford University, Stanford, CA, United States

Clair L. Alston-Knox     Predictive Analytics Group, Melbourne, VIC, Australia

Mark D. Barry     High Performance Computing and Research Services, Queensland University of Technology, Brisbane, QLD, Australia

Jesper L. Boldsen     ADBOU, University of Southern Denmark, Odense, Denmark

Fred L. Bookstein

University of Washington, Seattle, WA, United States

University of Vienna, Vienna, Austria

Guillermo Bravo Morante

University of Granada, Granada, Spain

Konrad Lorenz Institute for Evolution and Cognition Research, Klosterneuburg, Austria

Hugo F.V. Cardoso     Department of Archaeology, Simon Fraser University, Burnaby, BC, Canada

Louise K. Corron     University of Nevada, Reno, NV, United States

Eugénia Cunha

University of Coimbra, Centre for Functional Ecology, Laboratory of Forensic Anthropology, Department of Life Sciences, Coimbra

National Institute of Legal Medicine and Forensic Sciences, Lisbon, Portugal

Susan R. Frankenberg     Department of Anthropology, University of Illinois at Urbana-Champaign, Urbana, IL, United States

Sara M. Getz     Department of Criminal Justice, University of Wisconsin, Platteville, WI, United States

Laura S. Gregory     Faculty of Health, Skeletal Biology and Forensic Anthropology Research Laboratory, School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia

Jieun Kim     Anatomy, Lincoln Memorial University, DeBusk College of Osteopathic Medicine, Knoxville, TN, United States

Lyle W. Konigsberg     Department of Anthropology, University of Illinois at Urbana-Champaign, Urbana, IL, United States

Nicolene Lottering

Faculty of Health, Skeletal Biology and Forensic Anthropology Research Laboratory, School of Biomedical Sciences, Queensland University of Technology, Brisbane

Faculty of Health Sciences and Medicine, Bond University, Robina, QLD, Australia

Donna M. MacGregor     Faculty of Health, Skeletal Biology and Forensic Anthropology Research Laboratory, School of Biomedical Sciences, Queensland University of Technology, Brisbane, QLD, Australia

George R. Milner     Department of Anthropology, The Pennsylvania State University, University Park, PA, United States

Stephen D. Ousley     Department of Anthropology, University of Tennessee, Knoxville, TN, United States

Michael H. Price     Santa Fe Institute, Santa Fe, NM, United States

Luis Ríos     Departamento de Biodiversidad, Ecología y Evolución, Universidad Complutense de Madrid, Madrid, Spain

Marta San-Millán

EUSES University School of Health and Sports, University of Girona, Girona, Spain

Group of Research on Clinical Anatomy, Embryology and Neuroscience (NEOMA), Department of Medical Science, University of Girona, Girona, Spain

Laure Spake     Centre for Research on Evolution, Belief and Behaviour, University of Otago, Dunedin, New Zealand

Kyra E. Stull

University of Nevada, Reno, NV, United States

University of Pretoria, Pretoria, South Africa

Peter Tarp     ADBOU, University of Southern Denmark, Odense, Denmark

Flávia Teixeira     University of Coimbra, Centre for Functional Ecology, Laboratory of Forensic Anthropology, Department of Life Sciences, Coimbra, Portugal

Svenja Weise     ADBOU, University of Southern Denmark, Odense, Denmark

Looking back and forward: An introduction—Defining, refining, and [re]modeling age estimation: A trajectory for forensic anthropology

Jieun Kim; Bridget F.B. Algee-Hewitt

Aging is a topic of considerable importance in biological anthropology. The steady number of publications and their emphasis on method development, testing, and validation provides clear evidence of sustained interest, while the history of debate over the validity of this theoretical work and their practical outcomes is testament to the depth of our collective investment. There is no denying how the reliable estimation of age-at-death from the human skeleton is of fundamental importance for the value that it brings to the field, and subfields, of biological anthropology, and to the areas of specialty and subject matter with which it interacts.

The study of age contributes to our understanding of morphological variation and the senescent process in modern population biology. As such, skeletal development and degeneration has implications that reach into the anatomical and medical fields as we, as skeletal biologists, can grapple with the effects of genetics, environment, and individual lifestyle factors on bone density, rates of fracture healing, and atypical skeletal expression. Estimating age is important among the personal identity parameters that are used in medico-legal case identification in forensic anthropology, as we seek to provide law enforcement, medical examiner, and nongovernmental agencies with the information that can help to link the unknown individual with the named person. In these circumstances, the anthropologist, working in the service of humanitarian aid and social justice, estimates skeletal age to assist in missing person, asylum seeker, and undocumented death cases. Finally, age-at-death is fundamental to the paleodemographic reconstruction of mortality profiles for skeletal assemblages in bioarcheology, allowing us to bring a better understanding of the life and death of past peoples.

Despite the importance of knowing about skeletal age and aging, achieving accurate, precise, and repeatable estimation continues to be a challenge. Several acknowledged but unresolved theoretical and methodological problems continue to actively constrain the optimal inference of skeletal age for both juveniles and adults. This volume was motivated by these challenges and the significant place that age estimation has held in the traditional field of physical anthropology and for the potential position that it holds in "the new biological anthropology (Fuentes, 2010)" as we seek out innovative research that supports collaboration and crosses disciplinary, theoretical, and methodological barriers.

It is in the spirit of this new vision which we hope the new biological anthropology will indeed embrace, we take up a discussion of this foundational research area. Our volume brings together experienced computational and/or forensic anthropologists as invited authors, whose contributed papers cross-cut old and new areas of research, with promise of impact in the field of biological anthropology and specifically on the practice of forensic anthropology. These works address long held, but poorly justified assumptions; fill gaps in knowledge and aging criteria, and provide new opportunities for research directions, whether in the form of data collection techniques, methods of analysis, or populations under study.

This volume is designed to motivate its readers, be they young thinkers, leaders in the field, or pioneers in their areas of research specialty, to continue to think critically and to more readily embrace thinking creatively about skeletal age estimation. We believe strongly that only in both looking back and forward, can we more fully define, more carefully refine, and maybe even completely remodel the trajectory for the theory, method, and practice of age-at-death estimation in forensic anthropology.

We are encouraged by how this collection of works demonstrates that, together, as a diversely trained and experienced community, it is possible to bridge the long acknowledged disconnect between computational and applied work in forensic anthropology. Through shared—same but different—foundations of knowledge, we gain fresh perspective into how best to improve estimation and meet the real-world forensic needs of aging in response to such challenges as changing demographics, disasters, and humanitarian crises. Now is the time to fully capitalize upon the insights that computational approaches to age-at-death estimation can provide in forensic casework contexts.

The challenge of advanced aging: Synthesizing theory, method, and forensic practice

For the past four decades, biological—including forensic—anthropologists have made a series of significant reformations to old ideas and innovations that motivate new research on skeletal growth and development and degeneration, as the field works to improve the accuracy, reliability, and precision of age estimation methods across the entire lifespan. This work, realized often most profitably through interdisciplinary efforts, has explored various anatomical sites of the skeleton as potential age markers that are informative of the ontogenetic or senescent processes and which offer reasonable predictions of chronological age at the time of death (Cunningham et al., 2016; DiGangi et al., 2009; Louise et al., 2017; Prince and Ubelaker, 2002; Ríos et al., 2008; Rissech et al., 2007; San-Millán et al., 2017; Watanabe and Terazawa, 2006). However, as the age-old problems of aging continue to wear on forensic practice, increasingly researchers are recognizing the need to turn to more innovative methodological solutions to resolve fundamental blockers in quantitatively rigorous ways that deliver both estimates and forensically useful evidence of age and explore new opportunities for broadening our thinking about how we choose and define a good age indicator to understanding the scope of the information that these different indicators, old and new, can convey (Chapters 6, 8, and 9). There is an exciting focus on novel quantification techniques that exploit imaging technologies to capture age data (Chapters 7 and 8; Navega et al., 2017; Stoyanova et al., 2017; Villa et al., 2015), and on the development of robust statistical approaches that more effectively and efficiently model skeletal age data (Anderson et al., 2010; Boldsen et al., 2002; Konigsberg, 2015; Stull et al., 2014). Continued efforts of forensic anthropologists to diversify the age groups, toward a uniform distribution, and populations by sex and ancestry affiliations, that make up the reference skeletal samples used to develop age estimation methods have motivated a number of population/ancestry-specific approaches (Chapter 2; Han et al., 2009; Katz and Suchey, 1989; Kim et al., 2019; Lottering et al., 2013; Nagaoka et al., 2012; Rissech et al., 2007; Sakaue, 2006) while questions of their utility have driven many subsequent validation studies (Baccino et al., 1999; Martrille et al., 2007; Murray and Murray, 1991; Rissech et al., 2012; Saunders et al., 1992).

These trends suggest that in recent years aging research has grown increasingly dynamic. Nonetheless, there remains common themes in age estimation that are widely well held in the field and impact perspectives and practice. Using multiple age indicators over a single age indicator yields more accurate age estimates. Age estimation is most accurate for sub- and young adults, while estimation error increases with age. Females tend to have more varying patterns of skeletal aging than males. The age-informative power of classic age indicators (e.g., the pelvic joint surfaces) decreases past the middle-aged. Most importantly, the highly variable nature of aging processes perpetuates higher degrees of inaccuracy and bias in age estimation, most severely for adults, even when sophisticated quantification or statistical modeling techniques are used to reduce technical or biological noise.

This latter issue of signal to noise in the presence of high variation is not exclusive to age-at-death estimation from the skeleton; indeed, it permeates all aspects of the biological study of aging in contemporary humans. It affects even the gold standard field of genomics in the context of especially DNA methylation to track individual epigenetic clocks and studies of clinical phenotypes, like serum contents, saliva, and blood pressure, of the living have chronically struggled to find a biological marker, or a combination of biomarkers, that can consistently accurately—if not perfectly—predict the normal aging process (Garinis et al., 2008; Hannum et al., 2013; Marioni et al., 2018; Sebastiani et al., 2017). The different genotypic and phenotypic age markers even have shown a decrease in predictive potential of such age indicators from the midlife and onward due to increased biological heterogeneity in the old-aged, which, thereby, creates an exacerbated underestimation of age for the older aged and elderly (Jylhävä et al., 2017). Forensic age estimation using molecular biomarkers generates similar patterns in magnitude and direction, where estimation error ranges from 4 to 10 years and old individuals are underestimated, to those values produced using skeletal remains in a more general bioanthropological context (Montesanto et al., 2020). While these findings impede our ability to achieve the predicted ages we may desire, such decreasing predictability of biomarkers is hypothesized as a favorable centenarian benefit of an abnormally benign aging process (Montesanto et al., 2020) that preserves the elderly by keeping biological age younger than calendar age.

The heterogeneity or variation in the aging or developing skeleton is known to be associated with not only chronological time but also a complex suite of life history variables, representing intrinsic and extrinsic factors, like energy metabolism, nutrition, pathophysiological stress, and activity levels, which are unique to each individual and either accelerate or decelerate normal aging processes (Mays, 2015). Recently, anthropologists have begun to explore the impacts of these factors, notably body size and activity, on skeletal aging and their potential for integration within forensic age-at-death estimation practice (Campanacho and Santos, 2013; Campanacho et al., 2012; Merritt, 2015; Wescott and Drew, 2015). While investigating the effects of these variables in shaping human variation and incorporating them into age estimation models are necessary, many practical stumbling blocks remain as reference collections rarely have such detailed metadata with which models can be developed. What is more, forensic anthropologists analyze unknown individuals without the antemortem data that might be indicative of any of the factors mentioned before. Nonetheless, the recent increased use of medical imaging data of living populations or medical examiner’s offices for the development of age-at-death estimation methods may resolve, in part, these difficulties (Boyd et al., 2015; Harth et al., 2010; Kacar et al., 2017; Wink, 2014).

Recognizing the limitations imposed by a canonical adherence to the traditional or classic indicators’ age to inferential power and the problems arising from improvement methods that combine indicators in unfounded ways (Chapters 1 and 2), novel skeletal elements or traits are increasingly sought (Chapters 4, 6, and 10) and alternative, exclusively statistical, approaches introduced, in many cases specifically to lessen the effects of, or ideally fully remedy, bias (Chapters 7–9). Recent work looking at the problem of underestimation for old individuals from a broader biological anthropology perspective has returned promising results that have forensic implications, as it suggests that the magnitude of error does not increase exponentially with age and error appears to even decrease in elderly individuals over 70 years (Chapter 6). Furthermore, the widely held belief that subadult age estimation is more accurate across the different stages of the ontogeny process is deemed to be invalid as subadults have different degrees of uncertainty and magnitude of error and variance depending on the stage of growth and development (Chapter 3), and it appears that prior reliance on population-specific models based on presumed ancestry was equally misguided (Chapter 5).

What we should take away from the persistence of these age problems is the immediate need for the forensic anthropological investigation in skeletal aging to be reconceived in a cross-disciplinary framework, performing under the paradigm of Biology of Aging, that is free to draw upon external ideas and strengths from innovations in driver fields like medicine, genetic, image processing, artificial intelligence in order to enrich our anthropological perspective, while free to continue disciplinary efforts in method refinement and development. Understanding skeletal aging as a complex multifactorial process, not solely as a function of time, and borrowing medical concepts and theories from genetics, orthopedics, geriatrics, and pediatrics will augment our understanding of the patterning of skeletal aging variation, and the factors that shape and reshape these patterns. In doing so, we should be more equipped to recognize and identify actionable solutions for the different magnitudes and directions of estimation error and bias as expressed between/within individuals and populations as well as among various age cohorts. Finally, adopting the tools of computational biology and operating within of machine learning in artificial intelligence contexts can only improve the effectiveness and efficiency of our data collection, morphological modeling, and age predictive algorithms.

While discussions of method innovation and theoretical advancement are [re]invigorating to the field of aging, which has suffered periods of stagnations and arguably even decline, what we cannot lose site of the purpose of age estimation in forensic anthropology is application, bracketed by the practical needs of law enforcements (or some comparable investigative agency or organization and the standards of admissibility upheld by the legal community). The intention then must be to translate these new advanced age-at-death methods into the real-world forensic case analysis through the mutual exchange of information between developers and end-users. These bilateral efforts can be realized by developers actively disseminating new knowledge and offering regular training via conference workshops, presentations, webinars, or short courses to diverse levels of audience, including graduate students, medico-legal professionals, academic colleagues, and independent forensic practitioners. Similarly, forensic practitioners must understand that the optimal process is interactive, where there is a willingness to continuously update their education in these novel methodological approaches and to keep their lab protocols up to date so they are compliant with the current best practice recommendations of the field. The accessibility of these new, often more complex, methods can be increased by active development of open-source software that complements the methods. There exists already several successful examples of this research application pipeline: Transition Analysis3 (Milner et al., 2020), forAge (Stoyanova et al., 2017), IDADE2 (Rissech et al., 2019), DXAge (Navega et al., 2017).

Lastly, is it absolutely essential that forensic anthropologists take an active role in shaping the trajectory of age estimation theory and methods, upon which the quality of any applied results is contingent. This is especially urgent, given low energy responses to calls for increased rigor in the forensic sciences and the slow movement to action with regards to instituting criteria for best practice. In 2013 the Scientific Working Group of Forensic Anthropology (SWGANTH) in the United States published a series of whitepapers outlining best practice guidelines for forensic casework involving human skeletal remains. Among them, the recommendations for age-at-death estimation emphasized developing and using age-at-death estimation methods specific to a target population, which have minimal subjective interpretation, quantifiable error/uncertainty of estimation, and well-documented shortcomings resulting, having been subjected to rigorous testing. Despite these (now outdated) recommendations, forensic scientists have favored a handful of traditional methods that are known to suffer from high observer and method error, rely upon problematic approaches to reporting (point estimates and too narrow/too wide intervals), integrate multiple age estimates in ad hoc ways, were mostly designed on skeletal collections representing 20th century Americans, and, because of the lack of consistent training opportunities are variable, often wrongly implemented and poorly taught (Falys and Lewis, 2011; Garvin and Passalacqua, 2012; Kimmerle et al., 2008). Different training backgrounds among forensic scientists and biogeographical diversity in skeletal aging processes challenge us to develop a standardized protocol of age-at-death estimation that meets the expectations of the forensic communities: such a model has yet to be established and remains as one of the most pressing needs in the field (Organization of Scientific Area Committees (OSAC) for Forensic Science, 2017).

As the nature of the genetic, biochemical, and skeletal biological mechanisms of aging is largely unknown (Goto, 2015), developing age-at-death estimation methods that offers increased accuracy, reliability, and precision demands ongoing investment with contributions from the research and applied communities. Perhaps for this reason, in the face of these challenges, it is an exciting time for research in age-at-death estimation. There is great potential to propel method and application forward by embracing the strengths of morphoscopic, computational, and algorithmic and 3D data capture techniques to deliver an integrative, thoughtful, and accessible quantitative solutions to age estimation for multiple populations. This edited volume provides readers with a window into the kind of research that will generate change in forensic age-at-death estimation and equip readers with the knowledge to promote new ideas and design their own path forward. The cumulative efforts of the forensic anthropologists, broadly defined, presented in this volume bring many perspectives to the discussion of skeletal development and aging that enrich our understanding of these complex processes, how they may be quantified, and what approaches permit their use in forensic aging contexts. Further to this point, this volume’s work also reflect the goals of justice system, identifying the immediate need for and great value in methods that emphasize the speed, accuracy, and scope of forensic analysis, which ultimately bolsters the administrating of justice (National Institute of Justice, 2020).

New contributions to aging problems: Chapter summaries

While osteological age markers are well characterized and aging methods well established, our ability to produce accurate and precise age-at-death estimates from skeleton remains a challenge for forensic case analyses. As a field, forensic anthropologists continue to ground best practice in historical precedent, relying on assumption rather than evidence for method efficacy and reliability, give preference to techniques that fail to adequately account for the complexity of the biology of aging and capture the resulting variation in the skeleton, and, ultimately, generate estimates that are unrealistic or inadequate for use in casework applications. In this volume, we present a book-length treatment of theoretical and statistical issues that have now shaped mainstream discussions (or debates) in age estimation literature, whether forensic or bioarcheological.

With chapters focusing on such fundamental areas for method implementation as specifying the population to combining indicators in overlap methods, contributors deliver authoritative research whose concluding statements can guide both future investigations in age methods and, perhaps most critically, direct forensic anthropological application toward best practice. Sampling across various areas of computational expertise, this volume also aggregates studies that present recent advances, in age estimation methodologies, which are both exciting for their innovation and important for the roadmap that they offer: demonstrating how using cutting-edge research can be implemented in forensic aging and giving insight into the future state of the art—from transition analysis to pipelines for ancestry and age inference. These and other works directly challenge and expand how we think about the senescent process, our methodological frameworks, and what information age indicators can and cannot convey.

What unifies this volume is a focus on long-standing aging problems—those that often lie latent within our methods and applications, are well known and have yet to be formally addressed or adequately resolved. Contributors are, therefore, motivated by the need to identify the most age-informative skeletal elements and traits, which better differentiate young, middle, and old individuals of various populations and provide greater granularity in estimates, address the prerequisites for method implementation, like prior knowledge of ancestry and the selection of population references or population-specific methods, that often result in ad hoc or best (but poor)-fit decisions, and, overall, thoughtfully attend to the developmental or degenerative processes in the skeleton, so that age is inferred with increased accuracy and estimates reflect biological, not just statistical or observational, reality.

This book is organized into four sections, (1) Long-standing problems of the population, (2) Aging across the ages, (3) Computational methods come of age, and (4) Classic indicators rejuvenated. In the section that follows we provide a summary of each chapter, underscoring how they discuss, each in very individualized ways, the major statistical and technical concerns in age estimation, from finding a scientifically sound way of combining multiple correlated age indicators or multiple age estimates from different age indicators, objective quantification of age-related skeletal changes with reduced observer error, and to the robust statistical analysis for traditional, score-based method revisions. The contributors of each chapter tackle, and propose possible solutions to, these issues by taking advantage of innovative data visualization and exploratory data analysis, robust statistical modeling techniques, state-of-the-art three-dimensional imaging technologies, and novel age indicators that were traditionally discounted for age estimation.

Section A. Long-standing problems of the population

In Chapter 1, Using data from the US Korean War Dead and the Terry Collection to demonstrate problems of the common Overlap Methods, Konigsberg suggests a scientifically sound way of combing age estimates derived from multiple age indicators specifically using a multivariate transition analysis approach while discouraging the use of overlap methods, a practice that have been favored by many forensic practitioners. By using datasets of the Terry collection and Korean War casualties and simulated target samples, the author compares age estimation trends and error structures of the overlap methods vs transition analysis using a Bayesian framework. Konigsberg shows that, whether one is taking a range of the minimum and maximum age or one standard deviation around the mean, the phase-based overlap approach perpetuates the problem of age mimicry. Furthermore, the author highlights the risk of a systematic bias where the young are overestimated while the old are underestimated when overlap methods based on an age-for-stage system are used. This chapter illustrates the importance of using an appropriate informed prior particularly when the likelihoods are weak. Konigsberg concludes by recommending the use of transition analysis instead of overlap methods as it combines age indicators based on statistically sound theory, provides formal quantification of error, and generates narrower age intervals compared to overlap methods.

In Chapter 2, Testing for differences in senescence using score data to understand the effects of reference sample choices, Frankenberg delivers both a critical statement on the population in age estimation as well as providing a practically useful step-by-step guide for exploratory data analysis before fitting a probit model for age-at-death estimation. The author uses a large-scale dataset (n > 1700) of the pubic symphysis, auricular surface, and ecto-cranial sutures to fit the univ-/multivariate cumulative probit regression model, and demonstrates the goodness-of-fit test for the probit regression and the optimization of ordinal stages. For model comparisons, Frankenberg suggests using [composite likelihood] Bayesian information criterion (BIC/CLBIC) to decide whether one should use a parsimonious or more complex model with additive or interactive terms (e.g., the impacts of race or sex) for age estimation. The author also highlights two important issues associated with a multivariate age estimation framework: the conditional independence and residual correlations among multiple age indicators. The analysis of eigenvalue structures of residual correlation matrices shows that other than age, sex, and population, additional unmeasured forces may affect the way the skeleton changes, further complicating the normal aging process and creating variation in aging skeletons.

Section B. Aging