Doctoral Dissertation Research: The conditional independence assumption in juvenile dental age estimation

根据骨骼元素估算个体的年龄可以让我们了解人类历史上的生物文化经历、身份和种群动态。使用牙齿发育信息的方法对青少年年龄估计特别有用。本博士论文研究项目探讨牙齿之间的发育关系以及这些关系如何影响青少年牙齿年龄估计的错误率。更好地了解人类牙齿相关性的变化可以改善生物考古和法医研究的年龄估计方法，以及对人类生长和发育的临床理解。该项目支持STEM领域的研究生培训、开源牙齿年龄估算软件的开发，以及为研究人员和从业人员举办的在线培训研讨会。为了简化数学模型，通常假设在考虑了实际年龄的影响后，牙齿发育特征是相互独立的。这种条件独立假设的问题在于，如果条件独立不是生物学上有效的，那么得出的年龄估计就会有偏差，错误率也会高于预期。这在生物考古学和法医人类学中都是一个问题，因为有偏见的年龄估计会损害我们对过去人口的理解，而未知的年龄区间错误率不符合专家证词的道伯特标准。这些问题通过三个具体目标来解决：1)描述发育中牙齿之间的相关性程度，2)描述发育中牙齿之间相关性的可变性程度，以及3)开发，验证和测试一个包含牙齿之间相关性的开源发育年龄估计程序。这些目标采用贝叶斯统计模型和r中的聚类分析来实现。模型拟合在世儿童的牙齿发育分数，并使用现有的死者数据库中的牙齿发育分数进行验证，以便开发年龄估计软件。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Estimating the age of individuals based on skeletal elements can inform our understanding of human biocultural experiences, identities, and population dynamics across the history of our species. Methods that use information about dental development are particularly useful for juvenile age estimation. This doctoral dissertation research project examines the developmental relationships between teeth and how these relationships affect error rates in juvenile dental age estimation. A better understanding of human variation in dental correlations can improve both methods of age estimation for bioarchaeological and forensic research as well as clinical understanding of human growth and development. The project supports graduate training in STEM, the development of open-source dental age estimation software, and online training workshops for researchers and practitioners. In order to simplify mathematical modeling, it is often assumed that dental developmental traits are independent from one another after accounting for the effect of chronological age. The problem with this conditional independence assumption is that if conditional independence is not biologically valid, resulting age estimates will be biased and have error rates that are larger than expected. This is a problem in both bioarchaeology and forensic anthropology because biased age estimates can compromise our understanding of past populations and unknown age interval error rates do not meet the Daubert standard for expert testimony. These issues are addressed through three specific aims: 1) characterize the extent of correlations between developing teeth, 2) characterize the degree of variability of correlations between developing teeth, and 3) develop, validate, and test an open-source developmental age estimation program that incorporates the correlations between teeth. These aims are approached using Bayesian statistical models and cluster analysis in R. Models are fit to dental development scores of living children and validated using dental development scores from existing decedent databases in order to develop the age estimation software.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.