The Optimized Summed Scored Attributes Method for the Classification of U.S. Blacks and Whites: A Validation Study

Recently, Hefner and Ousley (2014) introduced the optimized summed scored attributes (OSSA) method that maximizes between‐group differences in U.S. black and white populations by dichotomizing six cranial morphoscopic trait scores. This study tests OSSA using an independent skeletal sample (Hamann‐Todd, n = 208) and positively identified forensic cases (Mercyhurst University, n = 28, and New York City Office of Chief Medical Examiner, n = 38). An evaluation of trait frequencies suggests shifting the heuristically selected sectioning point separating U.S. black and white populations from ≤ 3 to ≤ 4. We found a total correct classification of 73.0% (B = 50.9%, W = 89.2%) using the originally suggested sectioning point of ≤3, while the total correct classification increases to 79.2% (B = 80.2%, W = 78.5%) with a modified sectioning point of ≤4. With the increased total correct classification and reduced classification bias between ancestry groups, we suggest the modified sectioning point of ≤4 be used when assessing ancestry in forensic unknowns.     

A Validation Study of the Langley et al. (2017) Decision Tree Model for Sex Estimation

Langley et al. (2017) developed a sex estimation decision tree utilizing two traditional cranial traits (glabella and mastoid) and a new trait: zygomatic extension. This study aimed to test the reliability of their zygomatic extension scoring method and validate their sex estimation method. Ordinal score data were collected from 281 male and female U.S. White and Black individuals. The five traditional cranial traits were collected from physical specimens, while zygomatic extension was scored from 3D cranial models. Intra‐ and interobserver analyses carried out on a subsample of 30 individuals indicate good agreement between zygomatic scores. The decision tree correctly sexed 71.5% of the sample, but a strong sex bias (94.2% correct for females, 49.3% correct for males) severely limits the utility of this method. The Walker (2008) and Stevenson et al. (2009) methods produced higher accuracy rates (80.8% and 82.6%, respectively), although these methods also produced sex and ancestry biases.     

Assessing zygomatic shape and size for estimating sex and ancestry in a South African sample

Unidentified, decomposed and skeletonised human remains are frequently found in South Africa, therefore, standardised, reliable and relevant sex and ancestry estimation methods are required for forensic identification. This study assessed sex and ancestral variation in zygomatic size and shape in a South African population using geometric morphometric analyses. The zygoma of 158 South African individuals were sampled. Eight zygomatic landmarks were captured in 3-dimensions using a Microscribe G2 digitiser and assessed using procrustean geometric morphometrics. Shape and size differences were analysed using multivariate linear regression, discriminant function and canonical variate analyses. Males had significantly larger zygomas than females. Significant shape variation was found between ancestral groups. Bantu-speaking and Mixed ancestry individuals had narrower, shorter and more anteriorly projecting orbital margins, whilst Europeans had vertically elongated and receded orbital margins. European ancestral groups were most discernible from Bantu-speakers and Mixed ancestral groups. Ancestry estimation accuracies improved when ancestry was aggregated with sex. Pairwise ancestry-linked comparisons in females were as follows; Bantu-speakers (76%) from Europeans (72%), Bantu-speakers (71%) from Mixed ancestry (59%) and European (72%) from Mixed ancestry (63%). Similarly, ancestry-linked comparisons in males were as follows; Bantu-speakers (77%) from Europeans (81%), Bantu-speakers (53%) from Mixed ancestry (59%) and European (72%) from Mixed ancestry (82%). Size differences are putatively linked to variations in hormone-regulated growth and muscular robusticity between males and females. Shape variations between ancestral groups are likely attributable to the heterogenous genetic and ancestral origins of the South African population. It is challenging to distinguish between South Africa Bantu speakers and Mixed ancestry people due to Mixed ancestry individuals having variable genetic contributions from Khoesan, Bantu-speakers, Europeans and Asians. Bantu-speaking and Mixed ancestry people had zygomatic morphologies consistent with historical thermoregulatory adaptations to sub-Saharan climates, reported in African-descendants. Zygomatic morphology in European descendants suggests ancestral origins from colder climatic regions. This study demonstrated the utility of the zygoma in distinguishing between ancestral groups in South Africa, but further research is required to develop population-specific standards to distinguish between South African populations with shared African ancestry. The zygoma shows a promising ability to estimate sex and ancestry in South Africans, suggesting population specific standards for this bone may be of forensic interest.     

Evaluation of the (hu)MANid program for sex and ancestry estimation in a diverse,contemporary CT scan-based sample

Human remains from forensic and bioarcheological contexts are often fragmentary, requiring methods for estimating a forensic profile that are based upon limited skeletal features. In 2017, Berg and Keryhercz created an online application, (hu)MANid, that provides sex and ancestry estimation from mandibular morphoscopic traits and linear measurements. In this study, we examine the utility of the (hu)MANid application in a diverse, urban US adult sample (aged 20–45; n = 143) derived from computed tomography (CT) scans. We secondarily conduct a preliminary analysis of the program's utility in a sample of adolescents (aged 15–17; n = 40). Six morphoscopic, and eleven morphometric traits were recorded as directed by the literature associated with the (hu)MANid program. Percent correct classification and posterior predictive values were calculated for the sex and ancestry estimations output by the program; chi-squared tests were employed to compare self-reported and predicted ancestry. In the adult sample, sex was accurately predicted for 75.52% of the sample. Ancestry prediction, however, was less favorable ranging from 19.3% to 50% correct. For the adolescent sample, correct sex estimation (45%) did not surpass what could occur by chance alone, though ancestry prediction fared better than in the larger adult sample (percent correct prediction overall average: 47.5%, range 35.71%–71.43%). The (hu)MANid application shows utility for use with CT scan-derived adult samples for sex estimation, but caution is warranted for ancestry estimation and use with samples that may not have reached full adult maturity.     

Morphoscopic Trait Expression in “Hispanic” Populations

This study evaluates population variation of eight cranial morphoscopic traits using samples of known southwest Hispanics (n = 72), Guatemalans (n = 106), American Blacks (n = 146), and American Whites (n = 218). We applied the support vector machine (SVM) method to build a prediction model based on a subsample (20%) of the data; the remainder of the data was used as a test sample. The SVM approach effectively differentiated between the four groups with correct classification rates between 72% (Guatemalan group) and 94% (American Black group). However, when the Guatemalan and southwest Hispanic samples were pooled, the same model correctly classified all groups with a higher degree of accuracy (American Black = 96%; American White = 77%; and the pooled Hispanic sample = 91%). This study also identified significant differences between the two Hispanic groups in six of the eight traits using univariate statistical tests. These results speak to the unique population histories of these samples and the current use of the term “Hispanic” within forensic anthropology. Finally, we argue that the SVM can be used as a classification model for ancestry estimation in a forensic context and as a diagnostic tool may broaden the application of morphoscopic trait data for the assessment of ancestry.     

The rhomboid fossa of the clavicle as a sex and age estimator

The costoclavicular (rhomboid) ligament connects the first rib to the clavicle, stabilizing the pectoral girdle. It produces skeletal traits that may be tubercles, roughened impressions, shallow groove-like fossae, deep fossae, or leave no trace. A pit or depression at this site is often called a "rhomboid fossa." While these markings may appear pathological, they are normal variants of the clavicle. Using a large contemporary sample (N = 344:113 females, 231 males), we evaluated the presence of a rhomboid fossa as a sex and age indicator for unidentified skeletal remains. Logistic regression found significant relationships between the presence of a rhomboid fossa and sex and between presence of a rhomboid fossa and age. Fossae were more common in males (36% left, 31% right) than in females (3% left, 8% right). Posterior probabilities suggest that a fossa on the right clavicle is indicative of a male with 81.7% probability; a fossa on the left is indicative of a male with 92.2% probability. Younger individuals more commonly exhibited rhomboid fossae than older individuals, and the largest fossae were most common in males 20-30 years of age. However, the age effect was not conclusive and must be corroborated by other methods. A test of the sex estimation method on an independent sample (26 males, 23 females) found nine males and only one female with fossae present on the left clavicle. When the costoclavicular attachment exhibits an impression, a tubercle, or leaves no trace, this method cannot be used for sex estimation. When a clavicle exhibits a rhomboid fossa, it is likely from a male. The greater difference in fossa expression between the sexes on the left clavicle makes use of the left bone preferable. This technique can corroborate other sex estimates or provide an estimate for unknown individuals in the absence of other skeletal indicators.     

Secular Change in Morphological Pelvic Traits used for Sex Estimation

This research evaluates secular change in Phenice's (Am J Phys Anthropol, 30, 1969 and 297) three morphological traits of the pubis, as described by Klales et al. (Am J Phys Anthropol, 149, 2012 and 104): medial aspect of the ischio‐pubic ramus, subpubic contour, and ventral arc. Ordinal scores were collected for these traits and compared between a sample of innominates from the historical Hamann–Todd Collection (n = 170) and modern Bass Donated Collection (n = 129). Using the Freeman–Halton test, significant differences between temporal sample score frequencies were found for all traits in females and for the subpubic contour and ventral arc in males. Despite these findings, classification accuracy using logistic regression between the temporal periods remained low (68.7%). These results suggest that secular changes in trait expression are occurring; however, sex estimation methods using these traits and created with historical samples are still applicable to modern forensic cases. In fact, the secular changes occurring in these traits contribute to better classification accuracy between sexes in modern populations.     

Metric Assessment of the Pubic Bone Using Known and Novel Data Points for Sex Estimation,

Biological sex estimation of skeletal remains is essential in forensic and archaeological analyses. Anthropologists most often use the pelvis, which is the most sexually dimorphic element both morphologically and metrically. While nonmetric pubic bone features have been studied extensively, few metric studies have examined this individual bone for dimorphism. For this study, three observers examined three previously identified and ten novel measurements of the pubic body on a modern sample of isolated pubic bones from the Maricopa County Forensic Science Center (FSC), in Phoenix, Arizona (n = 400). A relationship between pubic body measurements and biological sex was demonstrated, with significant correlations. Discriminant function analyses found that five measurements, four of which were novel, discriminated between males (89%) and females (86%). Observer experience level did not significantly impact the results. These five measurements were reliable and show promise for inclusion in metric methods for assessment of sex.     

A Test of Age Estimation Methods on Impacted Third Molars in Males

This study tests the relationship between third molar impaction and its concomitant effect on age estimation methods. Data were collected on radiographs of males analyzed in the Defense POW/MIA Accounting Agency Laboratory (n = 114). Radiographs of skeletonized individuals were scored for dental development, and age was assigned based on the appropriate ancestry‐based method. Differences between identified age and estimated dental age were assessed to determine whether molar impaction affected root development and age estimations. Results indicate that impacted teeth tend to be underdeveloped and result in age estimates that are too low. While these results are of note to anthropologists and odontologists performing dental age estimates, more work is needed to explore the effect of impaction on development among a more diverse sample.     

Adult sex estimation from the minimum supero-inferior femoral neck diameter in a contemporary osteological sample from Buenos Aires,Argentina

Sex estimation of adult skeletons is an important step in forensic analysis. Although the femur has been metrically studied for sex assessment around the world, very limited information is actually available on modern populations of Argentina. In this paper, the estimation of sex based on the metric evaluation of the supero-inferior femoral neck diameter (SID) in a reference osteological collection from the contemporary Chacarita Cemetery of Buenos Aires City (Argentina), is evaluated. Protocols generated using SID in other three reference skeletal samples are also tested on this collection. One hundred and sixty-four individuals of both sexes and between 24 and 96 years old are analyzed. Inter and intra observer errors suggest that the replicability of the procedure is adequate. The sectioning point calculated from the direct measurements is 30.86 mm and results show a high degree of dimorphism. The proportions of correct sex discrimination and the likelihoods of correct allocation obtained with the direct measurements, along with the results of a discriminant function, a binary logistic regression and a Bayesian approach, are all higher than 0.85 (0.85–0.93 for females, 0.88–0.91 for males). When the formulae from other samples are used in the Chacarita Collection, the percentages of correct estimations range between 72.41% and 81.03% for females and between 80.46% and 88.50% for males, while the likelihoods are between 0.73 and 0.81 for females and between 0.79 and 0.82 for males. As the values for the estimations obtained using the statistical procedures generated in the present research are higher than those available for other collections, the method is more adequate to use in the analysis of contemporary skeletal remains from Buenos Aires and surrounding areas. The trends identified highlight the importance of population-specific metric methodologies in forensic contexts and deserve future testing in contemporary samples from neighbouring regions.     

Sexual dimorphism in America: geometric morphometric analysis of the craniofacial region

One of the four pillars of the anthropological protocol is the estimation of sex. The protocol generally consists of linear metric analysis or visually assessing individual skeletal traits on the skull and pelvis based on an ordinal scale of 1-5, ranging from very masculine to very feminine. The morphologic traits are then some how averaged by the investigator to estimate sex. Some skulls may be misclassified because of apparent morphologic features that appear more or less robust due to size differences among individuals. The question of misclassification may be further exemplified in light of comparisons across populations that may differ not only in cranial robusticity but also in stature and general physique. The purpose of this study is to further examine the effect of size and sex on craniofacial shape among American populations to better understand the allometric foundation of skeletal traits currently used for sex estimation. Three-dimensional coordinates of 16 standard craniofacial landmarks were collected using a Microscribe-3DX digitizer. Data were collected for 118 American White and Black males and females from the W.M. Bass Donated Collection and the Forensic Data Bank. The MANCOVA procedure tested shape differences as a function of sex and size. Sex had a significant influence on shape for both American Whites (F = 2.90; d.f. = 19, 39; p > F = 0.0024) and Blacks (F = 2.81; d.f. = 19, 37; p > F = 0.0035), whereas size did not have a significant influence on shape in either Whites (F = 1.69; d.f. = 19, 39; p > F = 0.08) or Blacks (F = 1.09; d.f. = 19, 37; p > F = 0.40). Therefore, for each sex, individuals of various sizes were statistically the same shape. In other words, while significant differences were present between the size of males and females (males on average were larger), there was no size effect beyond that accounted for by sex differences in size. Moreover, the consistency between American groups is interesting as it suggests that population differences in sexual dimorphism may result more from human variation in size than allometric variation in craniofacial morphology.     

Is Dwight Right? Can the Maximum Height of the Scapula Be Used for Accurate Sex Estimation?*

Abstract: This paper presents data from a sample of 803 individuals (308 females and 495 males) from the Hamann‐Todd collection testing Dwight’s century‐old assertion that maximum height of the human scapula can be used for sex estimation—males being larger than 170 mm, females falling below 140 mm. The results of this project show Dwight’s method has high accuracy when scapular height falls either above or below the sex specific demarcation points (96.81%), but a vast majority of both males and females fall in between. The overall accuracy of the method is just 29.27%. By empirically demonstrating the limited usefulness of Dwight’s technique, the author hopes the rote republication of this method in introductory texts on the subject will cease, and draw attention to the need for multiple methods of sex estimation as a response to the overlap in both size and shape between males and females.     

Ancestry Assessment Using Random Forest Modeling, ,

A skeletal assessment of ancestry relies on morphoscopic traits and skeletal measurements. Using a sample of American Black (n = 38), American White (n = 39), and Southwest Hispanics (n = 72), the present study investigates whether these data provide similar biological information and combines both data types into a single classification using a random forest model (RFM). Our results indicate that both data types provide similar information concerning the relationships among population groups. Also, by combining both in an RFM, the correct allocation of ancestry for an unknown cranium increases. The distribution of cross‐validated grouped cases correctly classified using discriminant analyses and RFMs ranges between 75.4% (discriminant function analysis, morphoscopic data only) and 89.6% (RFM). Unlike the traditional, experience‐based approach using morphoscopic traits, the inclusion of both data types in a single analysis is a quantifiable approach accounting for more variation within and between groups, reducing misclassification rates, and capturing aspects of cranial shape, size, and morphology.     

Development and Assessment of Postcranial Sex Estimation Methods for a Guatemalan Population

This study tests whether postcranial sex estimation methods generated from Hispanic, and mainly Mexican samples, can be successfully applied to other increasingly common migrant populations from Central America. We use a sample of postcranial data from a modern (1980s) Guatemalan Maya sample (n = 219). Results indicate a decrease in classification accuracies for previously established univariate methods when applied to the Guatemalan study sample, specifically for males whose accuracies ranged from 30 to 84%. This bias toward inaccuracies for Guatemalan males is associated with the smaller skeletal sizes for the Guatemalan sample as compared to the samples used in the tested sex estimation methods. In contrast, the tested multivariate discriminant function classification yielded less sex bias and improved classification accuracies ranging from 82 to 89%. Our results highlight which of the tested univariate and multivariate methods reach acceptable levels for accuracy for sex estimation of cases where the region of origin may include Guatemala.     

A reevaluation of the sex prediction accuracy of the minimum supero-inferior femoral neck diameter for modern individuals

The results of an independent test of the minimum supero-inferior femoral neck diameter as a sex predictor are presented. Seidemann et al. (1) generated discriminant functions for Caucasians, African-Americans, and a combined race sample from the Hamann-Todd skeletal collection. Jackknifed classification matrices and the use of independent, random validation samples indicated a sex prediction accuracy in the 90% range. This, combined with a high rate of preservation, makes the femoral neck a significant measure for forensic applications. However, the method has not been evaluated on a truly modern sample. Data were collected for 94 males and 49 females from the Documented Collection at the University of New Mexico. The sample consists of 94 Caucasians, 33 African Americans, three modern Native Americans, two Hispanics, and 11 individuals of unknown ancestry. All individuals were born after the turn of the century. We evaluate the accuracy of the discriminant functions generated from the Hamann-Todd control sample. For Caucasians, 83% were correctly classified, for African Americans 97% were correctly classified and for the combined race function 85% were correctly classified. This decrease in accuracy is the result of the increase in African American male and all female sample means. This effectively decreases the separation between males and females for the femoral neck diameter. We generate new discriminant functions from the modern data and jackknife the classification matrices. The Caucasian function was 84% accurate, the African-American function was 82% accurate and the combined sample function was 85% accurate. The femoral neck may provide a useful alternative to multivariate techniques for individuals who are poorly preserved.     

Sex Determination from the Second Cervical Vertebra: A Test of Wescott’s Method on a Modern American Sample*

Abstract: Numerous methods for establishing a biological profile exist; however, many of these methods rely on the recovery of several specific bones or on fragile skeletal elements that are sometimes irrecoverable. It is for this reason new methods utilizing other previously under‐documented bones should be established and tested by the forensic anthropological community. This study tests the accuracy of Wescott’s (J Forensic Sci 2000;45(2)) method for determining sex from the second cervical vertebra. Specimens were drawn from the donated skeletal collection curated at the Hamilton County Forensic Center (n = 57) and the William M. Bass Donated Skeletal Collection (n = 243). Both intra‐ and inter‐observer error rates were low and accurate classifications ranged from 78% (females‐Function 1) to 90.6% (males‐Function 5). Of the five functions, Function 4 achieved the highest overall accuracy, with 260 individuals (86.7%) falling into the correct category. Overall, this method is an effective classificatory tool for sex estimation.     

Estimating sex of the human skeleton based on metrics of the sternum

Estimating the sex of skeletal remains is critical in creating the biological profile of an adult individual. Previous research has shown that analyzing the sternum may lead to an accurate estimation of sex based on studies performed on various populations around the globe, such as Indian, European, African, Canadian and North American. The motivation of the current study is to develop classification functions and sectioning points for use in forensic investigations in the United States. The majority of previously published methods are population specific, meaning the data would not prove useful in the United States. For this study, sternal measurements were collected from the William M. Bass Donated Skeletal Collection located at The University of Tennessee, Knoxville following the metric definitions provided by Schwartz and Bass. This collection consists of recent skeletal remains with known age at death, ancestry and sex. Material for the present study consisted of 410 human sterna: 285 male (256 American White/29 American Black) and 125 female (120 American White/5 American Black). Previous studies analyze the applicability of Hyrtl's law in the estimation of sex. This law follows that the mesosternum is greater than twice the length of the manubrium in males and in females the length of the manubrium is greater than half the length of the mesosternum. In this study, comparisons of the proportion of the length of the manubrium to the length of the mesosternum were performed to determine if Hrytl's law is applicable in an American population. Comparisons of these measurements between individuals identified as American Black and American White were analyzed to determine whether this method could be used on both population groups. Further, discriminant function analysis was used to estimate sex and provide a population specific classification function for use in the United States. The discriminate function analysis produced an overall cross-validation classification rate of 84.12% for sex estimation. The cross-validation classification rate for males and females was 80.00% and 88.24%, respectively.     

The Use of Femoral Neck Axis Length to Estimate Sex and Ancestry

Having multiple reliable methods of estimating sex and ancestry from various skeletal features increases the likelihood of identifying skeletal remains. Femoral neck axis length (FNAL), as measured in living individuals, has been shown to vary by sex and ancestry. FNAL has not, however, been previously measured directly from skeletonized remains and investigated for its potential use in forensic anthropological applications. This research proposes a method for measuring FNAL from skeletal remains, determines the reliability and repeatability of the measurement, and assesses the validity of FNAL in sex and ancestry estimation. Results showed low interobserver error in the measurement of FNAL (TEM = 0.33 mm, R = 0.99). Significant differences in FNAL were found between sexes as well as between American Black, American White, and Native American groups. FNAL can correctly classify sex in ~86% of all cases and is considered valuable to sex estimation. The value of FNAL to ancestry estimation, however, is considered limited.     

Reliability of the Walker Cranial Nonmetric Method and Implications for Sex Estimation

The cranial trait scoring method presented in Buikstra and Ubelaker (Standards for data collection from human skeletal remains. Fayetteville, AR: Arkansas Archeological Survey Research Series No. 44, 1994) and Walker (Am J Phys Anthropol, 136, 2008 and 39) is the most common nonmetric cranial sex estimation method utilized by physical and forensic anthropologists. As such, the reliability and accuracy of the method is vital to ensure its validity in forensic applications. In this study, inter‐ and intra‐observer error rates for the Walker scoring method were calculated using a sample of U.S. White and Black individuals (n = 135). Cohen's weighted kappas, intraclass correlation coefficients, and percentage agreements indicate good agreement between trials and observers for all traits except the mental eminence. Slight disagreement in scoring, however, was found to impact sex classifications, leading to lower accuracy rates than those published by Walker. Furthermore, experience does appear to impact trait scoring and sex classification. The use of revised population‐specific equations that avoid the mental eminence is highly recommended to minimize the potential for misclassifications.