Responding to the call: Population affinity estimation in a South African forensic context of identification - are we helping or harming?

Abstract

There is growing momentum in forensic anthropology to move away from using socially ascribed racial groups in analysis and reporting. This study evaluated whether and how anthropological estimation of population-affinity influences identification outcomes in forensic cases processed at the Forensic Anthropology Cape Town (FACT) laboratory in South Africa. A total of 172 cases (174 individuals) were examined. Estimated South African population-affinity distributions were: 37 % Mixed, 22 % African, 4 % European, 12 % displaying traits from multiple affinities, and 25 % indeterminate. Identification outcomes were assessed for 168 cases (six were known prior to assessment). Overall, 37 % were positively identified, 49 % remained unidentified, and 14 % had unknown status. Identification rates varied: African (51 %) and European (83 %) population-affinity groups showed higher rates of positive identification compared to Mixed (26 %) and multiple affinities (20 %). Notably, individuals with indeterminate affinity achieved a 44 % identification rate. Statistical testing indicated no significant association between estimated population-affinity and identification outcome (p = 0.974). Where comparison was possible, estimated population-affinity aligned with socially ascribed racial categories in 84 % of cases, with six misclassifications. Challenges were noted in distinguishing between South African Mixed and African population-affinities. Although the sample size was limited, our findings support discontinuing the inclusion of population-affinity estimation in South African forensic anthropology reports. In practice, law enforcement often translates anthropological affinity estimates into social racial categories, which in the South African context risks misclassification and undermines identification efforts.

Highlights

• •South African sample of anthropologically analysed medico-legal cases.
• •Population-affinity data with identification outcomes are presented.
• •Higher rates of identification for South Africans of European and African population-affinities.
• •Population-affinity estimation in South Africa is not critical for positive identification.
• •The discontinued use of anthropological population-affinity estimation in this context is supported.

1. Introduction

There is a call across many disciplines including biology, genetics and anthropology to move away from antiquated socially ascribed racial groups in research, as they fail to account for the range and continuum of human variation, along with the contributing factors for the existing variation [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11]]. The risks and concerns of their continued use is documented to reinforce racial stereotypes and biases especially for marginalised peoples and vulnerable communities [5,6,[11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26]]. Structural vulnerability can be caused by racial social marginalisation due to discrimination and oppression and as a result people can carry increased vulnerability, having disproportionately poor health, increased probability of dying early and violently [[27], [28], [29], [30], [31]]. This means that forensic anthropologists must look beyond demographics and consider their practice within a structural vulnerability framework and towards humanitarian action, one way to do this would be to take an antiracist stance [32,33]. It has been argued that these paradigms in forensic anthropology can improve the quality of outcomes and reduce inaccuracies, as well as elevate the discipline's standard of operation more generally [32,33].

In a letter to the editor in the Journal of Forensic Sciences, Bethard and DiGangi [17] noted some critical concerns and harm of connecting someone's social race to skeletal traits. They referred to a failure “to educate medicolegal stakeholders that skeletal ancestry and skin colour are not the same thing,” (pp 1792) and raised awareness of racial bias embedded within the police institutions and investigative processes. They requested practitioners to critically reflect on their practice.

In this current paper, we respond to the call for data on identification rate outcomes. We assess if and how these may be impacted by the inclusion of an estimated population-affinity in the forensic anthropology report, and how these data may be (mis)interpreted as racial categories by service providers like the police. We chose to use the word population-affinity over ancestry due to the compelling arguments by Winburn and Algee-Hewitt [34] and others [[35], [36], [37], [38]], a best matching population method is utilised. This was accomplished through a review of anthropological data pertaining to forensic cases submitted to the Forensic Anthropology Cape Town (FACT) laboratory, from Forensic Pathology Service (FPS) facilities in the Western Cape (WC) province, South Africa [39,40]. First, we provide background to the medicolegal death investigation system in South Africa.

Forensic mortuaries and the police in South Africa are faced with an overwhelming number of unidentified persons. Up to 10 % of decedents who are examined at forensic mortuaries in South Africa remain unidentified annually and undergo pauper burial [[41], [42], [43], [44]]. Suggested reasons for this burden have included high mortality rates, migration (both internal and cross-border, especially when undocumented), lack of ante-mortem dental and medical records and post-mortem changes such as decomposition or burning [39,[41], [42], [43], [44], [45], [46], [47], [48]]. Cases involving highly decomposed or burnt human remains can be referred to forensic anthropologists for demographic profile estimations and/or trauma analyses.

Forensic Anthropology Cape Town (FACT) is a service provider that assists the state with the identification of human remains and is based at the University of Cape Town [39,[49], [50], [51], [52], [53]]. It is one of two forensic anthropology laboratories in the WC province, [39,49]. Medicolegal forensic anthropology cases are commonly referred to FACT by the FPS facilities and the South African Police Service (SAPS) [39,47,[50], [51], [52]]. The inclusion of a forensic anthropologist is at the discretion of the forensic pathologists or the police [47]. FACT does on average 15–20 cases per annum [39]. When they are included sometimes it is direct consultation of a service and other times, they are more involved in a transdisciplinary collaboration in medico-legal death investigations, which is linked to improving case outcomes [54].

Initially, our study was an assessment of identification and regional taphonomy research based on retrospective examinations of anthropological assessments in the Western Cape (WC) province, South Africa [39,41]. Whilst population-affinity was a variable originally included in the study; we decided to critically re-evaluate our results to address several questions: Is population-affinity estimation necessary for forensic anthropology in South Africa? Does the knowledge of estimated South African population-affinity contribute to or impede positive identification? Are there any trends associated with different South African population-affinities regarding positive identification or case outcomes?

2. Materials and methods

This was a retrospective, descriptive, cross-sectional study, including all cases that were referred to FACT by FPS and SAPS facilities in the WC from January 1, 2006 to December 31, 2018. This study used data from the FACT repository (HREC REF: R012/2019), which contained FACT case reports and affidavits, and was carried out with institutional ethical approval (HREC REF: 263/2019).

The demographic estimations regarding age-at-death were juvenile (0–17 years), young adult (18–35 years), middle-old aged adult (36+), adult (18+). In South Africa there are three primary racial groups whereby people self and state-identify they are bureaucratically defined as black (African), coloured (Mixed) and white (European). Thus, the population-affinity was recorded as African, European, Mixed and ‘multiple’: Individuals of African population-affinity in South Africa descend from Bantu-speaking sub-Saharan agropastoralists who migrated into the country around 2000 years before present [55,56]. South Africans of European population-affinity descend from Europe predominantly Britain, Netherlands, France, Portugal, Italy, Greece, Germany [[55], [56], [57], [58]]. South Africans of Mixed population-affinity are a biologically heterogeneous group of people with variable admixture and descendants of indigenous African San, Khoekhoe, Bantu-speaking peoples, migrants from parts of Europe, Asian and Madagascan Cape enslaved people or migrants, they in fact may not be mixed [56,59]. The term ‘multiple’ was a classification used for decedents who were estimated to belong to either one population-affinity or another, for example African or Mixed. These individuals displayed traits from multiple population-affinities or had phenotypes not typical of one population-affinity over another.

The SAPS Investigating Officers were contacted in 2021 to ascertain if any of the decedents were positively identified. Where individuals had been positively identified, legal documentation was perused to ascertain the known population-affinity of the decedent, such that the accuracy of the anthropological estimations could be assessed. However, we were limited in this regard as (i) no ante-mortem population-affinity information was available - only a social race category, and (ii) while the forensic anthropologist would report an estimated population-affinity, the state/government/police would essentially ‘equate’ or transform this estimation into a social racial category in their investigation. South Africans of European population-affinity corresponded to the ‘white’ racial category; Mixed population-affinity to ‘coloured’; African population-affinity to ‘black’. In instances where individuals were estimated to belong to multiple population-affinities (e.g., African or Mixed population-affinity), the racial category was defined as either ‘black’ or ‘coloured’. Given these limitations, we could not determine the accuracy of the population-affinity estimations, but rather the concordance of the social race category as it was derived from the forensic anthropology report and the ante-mortem racial category reported by the next-of-kin. The limitation in this approach is acknowledged, which will be discussed later. It was the only means to assess if population-affinity estimations were interpreted in a ‘helpful’ way for the identification of unknown decedents.

All data were recorded and compiled into Microsoft® Excel (Version 16.30 for Mac 2019). IBM SPSS® Statistics software (Version 25.0.0.0) was used for statistical analyses and to generate tables and graphs. Descriptive statistics were computed for all the data. Additionally, to determine the distribution of variables, comparisons between categories within each demographic variable were investigated using Chi-squared tests and Fisher–Freeman–Halton exact tests. The alpha value for the statistical tests conducted was 0.05.

3. Results

A total of 172 forensic cases were examined, involving 174 individuals. Across population-affinities, most decedent individuals were male (Table 1) and most were adults (Table 2). Juveniles comprised 12 % (20/174) of the cases with most individuals in this category having indeterminate population-affinity (Table 2).

Table 1.

Composition of individuals examined by the Forensic Anthropology Cape Town laboratory by population-affinity and sex.

Population-affinity	Sex n (%)			Total n (%)	Number of individuals who were ultimately Identified n (%)
	Female	Male	Unknown
African	14 (36 %)	25 (64 %)	0	39 (22 %)	20 (33 %)
European	0	6 (100 %)	0	6 (3 %)	5 (8 %)
Mixed	19 (29 %)	45 (69 %)	1 (2 %)	65 (37 %)	17 (28 %)
Multiple	6 (30 %)	14 (70 %)	0	20 (12 %)	4 (7 %)
Indeterminate	10 (23 %)	26 (59 %)	8 (18 %)	44 (25 %)	15 (24 %)
Total	49 (28 %)	116 (67 %)	9 (5 %)	174	61

Table 2.

Composition of individuals examined by the Forensic Anthropology Cape Town laboratory by population-affinity and age-at-death.

Population-affinity	Age-at-death
	Juvenile n (%)	Young-Adult n (%)	Middle-Old Adult n (%)	Adult n (%)	Indeterminate n (%)	Total n (%)
African	5 (13 %)	7 (18 %)	17 (44 %)	10 (25 %)	0 (0 %)	39 (22 %)
European	0	0	5 (83 %)	1 (17 %)	0 (0 %)	6 (4 %)
Mixed	3 (5 %)	9 (14 %)	44 (67 %)	9 (14 %)	0 (0 %)	65 (37 %)
Multiple	3 (15 %)	6 (30 %)	8 (40 %)	3 (15 %)	0 (0 %)	20 (12 %)
Indeterminate	9 (20 %)	6 (14 %)	20 (45 %)	7 (16 %)	2 (5 %)	44 (25 %)
Total	20 (12 %)	28 (16 %)	94 (54 %)	30 (17 %)	2 (1 %)	174

A total of 61 cases had information on the case status and were positively identified. Of these 22 had information regarding place-of-birth and/or intended burial. This was important to examine local (national) vs non-local (foreigner) as a potential reason for lack of identification. Of these cases 86 % (19/22) were local: 45 % (10/22) were born in the were born in the WC; 36 % (8/22) in the Eastern Cape; 5 % (1/22) in the Northern Cape provinces. The remaining three individuals were foreign internationals 14 % (3/22) from Ghana, the Netherlands and Zimbabwe.

The population-affinity profile of the cases was 37 % (65/174) of decedents being Mixed population-affinity, 22 % (39/174) African population-affinity and 4 % (6/174) European population-affinity. Individuals who displayed traits from multiple population-affinities backgrounds accounted for 12 % (20/174) of the sample. Population-affinity was indeterminate for 25 % (44/174) of decedents, often due to lack of a cranium, fragmentation, or immaturity.

Since six decedents were already identified at time of examination and were referred for trauma analysis only, information regarding positive identification was reviewed for 168 individuals. Positive identification was reached for 37 % (61/168) and 49 % (84/168) were unidentified at the time of the study. For the remaining 14 % (23/168), identification status could not be ascertained from the Investigating Officers. There was a significant association between estimated population-affinity and rates of identification (N = 168, p = 0.005, Cramér's V = 0.233) with higher percentages of individuals estimated to be of African population-affinity (51 %, 20/39) and European population-affinity (83 %, 5/6) identified compared to those of Mixed population-affinity (26 %, 17/65) and multiple population-affinities (20 %, 4/20) (Table 1). Importantly, individuals with indeterminate population-affinity estimations were positively identified in 44 % (15/44) of cases, indicating that a population-affinity estimation was not a critical indicator for identification outcomes. Further, a Chi-Square test showed that estimating population-affinity was not significantly associated with identification outcome (χ² (2, N = 168) = 0.05, p = 0.974, Cramér's V = 0.018) (Table 3).

Table 3.

Breakdown of population-affinity estimation for identified cases examined by the Forensic Anthropology Cape Town laboratory.

Case Identification	Population-affinity Estimated
	No n (%)	Yes n (%)	Total n (%)
No	20 (24 %)	64 (76 %)	84 (50 %)
Yes	15 (25 %)	46 (75 %)	61 (36 %)
Unknown	6 (26 %)	17	23 (14 %)
Total	41 (24 %)	127 (76 %)	168

For those identified, only 38 had both a population-affinity estimation and a known social racial category. The social racial categories and population-affinity agreed for 84 % (32/38) of these cases (Fig. 1). Of these, 16 % (5/32) corresponded to European, 38 % (12/32) to African, 34 % (11/32) to Mixed and 12 % (4/32) to multiple population-affinities. When an individual was estimated to be of European population-affinity, it corresponded racially with ‘white’ in 5/5 cases (100 %), African with ‘black’ 12/14 (86 %), Mixed with ‘coloured’ 11/15 (73 %) and multiple with ‘black’ or ‘coloured’ 4/4 (100 %). In four of the misclassifications, decedents were estimated as Mixed population-affinity, which was interpreted as ‘coloured’ by the SAPS - but ante-mortem information indicated that three ascribed racially as black and one was white. The remaining two misclassifications were estimated to be of African population-affinity (and interpreted as ‘black’ by the SAPS), however, ante-mortem information indicated they racially ascribed as ‘coloured’.

Fig. 1.

Individuals examined by the Forensic Anthropology Cape Town laboratory from (a) cases positively identified with documented social race compared with estimated population affinity and (b) a breakdown of those where the results were discordant.

4. Discussion

These results represent a small sub-sample of all unidentified decedents from medicolegal death investigations in the region, as they are only those where forensic anthropology services were requested, and only from a single service provider over a limited time frame. Additionally, the sample of those with known identification was even smaller, and we acknowledge these limitations. Despite this, the results are a starting point of data analyses and production responding to the call by Bethard and DiGanagi [17].

Several studies have examined the contributions that forensic anthropologists have made in forensic casework globally [43,50,[60], [61], [62], [63], [64], [65], [66], [67], [68], [69]]. Regardless of population-affinity, males nearly doubled females in the cases (Table 1). According to Statistics South Africa [70] the sex ratio in the WC is nearly equal 51 % female and 49 % male, illustrating a societal sex balance. While not all unidentified cases are homicide, research into male violence and homicide is disproportionately high compared to women in South Africa [39,[71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86]]. These results are consistent with a systematic review of published articles on unidentified deceased individuals around the world where it was reported that 77 % of unidentified decedents were biologically male [46]. There were also disparities, which were suggested to be due to differences in regional demographics, murder rates, environmental conditions and/or socioeconomics [65], which highlight the importance and value of regionally specific research.

When the cases were examined by age-at-death, across population-affinities, most were adults. These results are consistent with those from the recent systematic review of the literature that found most unidentified decedents were between 30 and 55 years old [46,69]. Indeterminate population-affinity was high among juveniles at 20 %, an unsurprising result as estimating population-affinity anthropologically in South Africa is generally based on the assessment of fully developed cranial characteristics [87]. Thus, with juveniles estimating population-affinity would rarely be possible. However, the link between age and estimating population-affinity may not have been correlated with identification, like Hughes [69] juveniles and adolescents had higher rates of identification.

South African studies have either hypothesised or shown that a proportion of unidentified decedents are non-local foreign immigrants, or undocumented migrants [39,[41], [42], [43], [44],46,48,50,60,88]. Therefore, information pertaining to the decedents place-of-birth and intended place-of-burial were analysed in this study. From the 22 decedents that had information regarding place-of-birth and/or intended burial only three were foreign internationals. The other 19 were local nationals with most having the same place of birth and intended burial. Less than half of the decedents (10/22; 45 %) were residing in the WC province permanently, supporting the view that migration contributes to the number of unidentified remains. However, migration status does not appear to impact the odds of identification. These data suggest that this was due to internal migration within South Africa as opposed to foreign immigrants.

In the FACT reports/affidavits, the methodologies used for estimating population-affinity were frequently unreported, but FORDISC - a personal computer forensic discriminant functions computer programme - was not used [89]. FORDISC has been shown to be challenging and provides low accuracy for some populations including South Africans due to significant overlap between groups within the population [4,44,[90], [91], [92]]. In conversation with FACT practitioners, non-metric morphological traits were used e.g. those outlined by İşcan and Steyn [87]. Non-metric methods have been noted to exhibit poor scoring consistency and limited in their effectiveness, as they primarily evaluate variation in two dimensions [93,95]. Consequently, the application of these methods in forensic casework have been suggested to be problematic [94], potentially leading to inaccurate estimations and hindering the identification process. However, the genetic diversity in Africa is well established and in particular for the South African population [59,[95], [96], [97], [98], [99], [100], [101], [102], [103], [104]]. And most methods, including metric ones, do not account well for this level and range of variation [4,44,60]. Studies assessing population-affinity variation utilising geometric morphometric techniques have demonstrated a correlation between legally/peer-reported social race and cranial morphology, and distinct morphological variations between population-affinities, albeit with significant areas of overlap [93,105,106]. Geometric morphometric methods are highly repeatable and may be more reliable for use in forensic cases [93]. Genetic phenotyping and estimation of biogeographical population-affinity using single nucleotide variants in the DNA may also offer solutions [[107], [108], [109], [110]]. However, it will not address the disconnect between biological population-affinity and socially ascribed race. If population-affinity estimations are to be continued, then regional improvements and more research are required [111].

Population-affinity was estimated for 74 % (130/174) individuals’, which was lower than the 99 % and 91 % found in other studies [62,107]. Cranial fragmentation and immaturity contributed to indeterminate population-affinity estimations for 25 % (44/174) of decedents. The results of this study suggest that estimating population-affinity may not be necessary, as alleged (suspected) identity for cases was the most significant factor (81 %) contributing to positive identification [39,112].

Furthermore, individuals with indeterminate population-affinity were positively identified in 44 % (15/34) of cases. Where identification occurred social race and population-affinity estimation group agreed in 84 % of the cases. Albeit a bit lower than other studies in the 90 percentiles, it is still high [34,67,68,113]. Whether population-affinity was estimated or not did not significantly impact identification outcomes (χ² (2, N = 168) = 0.05, p = 0.974, Cramér's V = 0.018). These findings challenge the traditional reliance on population-affinity estimation as a key component in forensic identification processes. When compared to other elements estimated of the demographic profile for FACT cases with positive identification sex was accurately estimated in 98 % with only one misclassification [39]. Sex was indeterminate for only 5 % of decedents this was due to fragmentation, missing critical elements and immaturity [39]. These high rates of accuracy for sex add credence to the idea that population affinity estimates are not critical for identification. Despite the small sample sizes these results raise the question of whether population-affinity should be estimated for forensic anthropology, as is being debated and discussed elsewhere [6,8,17,110].

In the South African context, the term ‘coloured’ is a bureaucratically acceptable racial category that many South Africans claim as their socio-cultural identity. It is a biologically defined heterogenous group of people called ‘Mixed’ population-affinity in this study as defined in the materials and methods section [56,59]. In the South African context these terms are commonly used, bureaucratically sanctioned and its meaning is understood by South Africans. Admixed persons may fall under this category, but not all persons who self-classify as ‘coloured’ are admixed and may even be from homogenous population groups. For example, all San and Khoekhoe people were in the past and are currently categorised under this racial group and display African population-affinity [56,59]. In South Africa these racial terms while still utilised by government and are sanctioned it is notable they were created and applied under the apartheid regime. In a decolonial framework their continuance in South Africa is becoming increasingly contested [33]. The racial categorisation of ‘Mixed’ has been questioned in the forensic anthropological literature [34,68]. The equivalent term in our paper is multiple - for individuals who show population-affinity to multiple-membership signatures.

Individuals of European and African population-affinity had higher rates of identification compared with those of Mixed population-affinity. This is notable as in South Africa people of European population-affinity only make up 7.3 % of the national and 16.4 % of the Western Cape Provincial population [71]. Appel et al. [114] in the United States also found similar results of higher identification rates for people of European population-affinity. Due to the small sample size in our study the association between the estimated population-affinities and identification rates should be interpreted with caution. The reasons for this are not entirely clear. We hypothesise that it may relate to individuals of European and African population-affinity being reported as missing less frequently in the region compared to those of Mixed population-affinity, who make up most of the population in the WC province. There may be fewer missing persons records for police to match. However, it may be that racial disparities/biases exist in policing, and more resources are put into cases involving South Africans of European and African population-affinity. While this has not been clearly demonstrated in a South African context, similar trends have been observed in other parts of the world [115]. South Africa's historical racial legacies suggest it may very well be a factor.

To examine informativeness of the population-affinity estimations, social race categories derived from population-affinity estimates and ante-mortem records were compared for the positively identified individuals (n = 38). There were six misclassifications, which is problematic and emphasises the limitations with ‘deriving’ race from population-affinity; particularly for individuals in the region of Mixed and African population-affinity (Fig. 1). Our results highlight the challenge detecting population-level differences between Mixed and African population-affinities as both are primarily descendants from Africa with past and present admixture, which is supported by other research showing morphological and biological similarity [4,20,56,57,59,60,93,94,[96], [97], [98], [99], [100], [101], [102], [103], [104],116,117]. On the other hand, the comparison of social race categories is also flawed and does not indicate the accuracy of anthropological population-affinity estimations, as an individual could socially/legally identify as a particular group, however, their skeletal morphology may reflect other aspects of their population-affinity.

Racism and racial divides are historically embedded in South Africa linked to the colonial history and apartheid regime of racial segregation that ended in 1994. The bureaucratically and currently legislated racial terminology in South Africa were created and formally established under the apartheid government. Many concerns over the continued use of these terms in democratic South Africa have been raised [4,14,20,117], especially for their lack of inclusivity and historical legacies of prejudice and discrimination, segregation and othering [4,14,20,[117], [118], [119], [120], [121], [122], [123]]. Their continued risk and use in forensic casework, biological research and sciences for purporting the existence of race as a socially ascribed and constructed variable as having biological substance is dangerous [20,21,117]. Local forensic practitioners in South Africa have been calling for decolonisation of the discipline to improve scientific integrity in its service to society [123,124].

The presented results illustrate a disconnect between social racial categories in South Africa that do not necessarily conform to the distinct broad biological, genetic, and geographical population-affinity origins, meaning a disconnect between skeletal traits and population affinity that are translated into social racial categories. Regarding racial versus population-affinity categories in forensic anthropology in South Africa. Morris [4] stated the only way there is value added is if the categories used are understood and interpreted correctly by the police and legal officials. Population-affinity data will be interpreted, organised and ratified into racial socially ascribed categories, thus, leaving the interpretation to the police and legal advisors may result in disadvantage to the decedent, even cause prejudice and discrimination to be introduced into a medicolegal death investigation [7]. More awareness is required within the medicolegal death investigation service that skeletal population-affinity may not correlate to one's race or skin pigment. Alternatively, estimating population-affinity is perhaps no longer necessary [17,18,99,104] and as found in this study, identification results may be unaffected or improve without its inclusion.

Some forensic anthropologists are calling for the discontinuation of population-affinity estimations [1,6,7,17,18,32,124,125]. Smay and Armelegos [1] discuss transforming the discipline and recount the history of race in biological anthropology as well as debunk the arguments for its continued use. Human ethical mandates regarding research integrity would be that the researcher adheres to balancing principles of beneficence and nonmaleficence, meaning the outcomes benefit equitably and fairly, while not introducing unnecessary harm or risk [20]. There were instances in this study where population-affinity was estimated, but the person was not identified. It is unclear if this was due to population-affinity estimations being inaccurate, or if this was due to population-affinity estimations being incorrectly equated with race, or other reasons entirely. Either way, the question remains as to whether harm was done by estimating population-affinity in the first instance, and whether it may have impeded identification, rather than facilitated it. Therefore, for these individuals, population-affinity estimations may be hindering their identification and thus harming more than benefitting. If population-affinity was not estimated, different avenues for investigative leads may be used and followed thus, increasing chances of identification. Others argue the importance of statistical tracking these data to ensure that racial disparities are not invisible or hidden statistically [111]. Our priority is to assist with identification and the data regarding identified and unidentified cases will not be lost of statistics if the population-affinity is not estimated in the forensic case reports. Rather investigators may rely on alternative and more informative data and information facilitating identification.

M'charek et al. [6] describe the concerns of race in forensic identification as othering and equates race as biology, reduced to bodily characteristics. An antiracist framework is a move towards racial equity and fairness in forensic anthropology [32,33], and we argue for this action to improve and create more equitable case outcomes for decedents. Despite antiracist movements, some have called for value in estimating population-affinity in biological anthropology [109,126,127]. As the ultimate goal is identification, many more are calling for increased inclusivity, equitable, holistic and trans-disciplinary approaches to medicolegal death investigations, which provides closure and justice equitably for all decedents [[23], [24], [25], [26], [27], [28], [29], [30], [31],54].

5. Conclusions

Identification of the deceased is important for criminal, social, humanitarian, ethical and civil reasons. We respond to the call for data on identification rate outcomes, if and how these may be impacted by the inclusion of population-affinity in the South African forensic anthropology affidavit or report. This was accomplished through a review of anthropological data pertaining to forensic cases submitted to FACT, from all FPS facilities in the WC. The population-affinity profile represents a subset of decedents in the WC province, providing a local sample perspective of anthropologically analysed unidentified bodies. In this study, individuals were mostly of Mixed or African population-affinities. Individuals of European and African population-affinities had significantly higher rates of identification compared with those of Mixed population-affinity. More research is needed to examine whether these findings hold true in larger samples and different contexts (e.g. other forensic laboratories is the country) and to understand the reasons behind the population-affinity disparities in identification rates. Do population-affinity estimations obstruct how the information provided to law enforcement is used, therefore, hindering or harming the chances of certain groups of people being identified? When population-affinity was not estimated more cases were positively identified for people of black and white social racial groups. Population-affinity estimations were not a significant factor contributing to positive identification, and there were challenges detecting local population-level differences between Mixed and African population-affinities. Upon critical reflection, we support the advisement of others in the discipline for population-affinity estimation to be discontinued, especially as the police tend to ‘equate’ or transform population-affinity into social racial categories and in 6/38 the social race category derived from population-affinity was not the individual's actual reported race. While correlation between legally/peer-reported social race and cranial morphology and distinct morphological variations between population-affinities have been demonstrated, there are significant areas of overlap particularly between people of Mixed and African population-affinity in South Africa. This fact needs to be built into forensic anthropology reports to ensure accurate interpretation of the results.

CRediT authorship contribution statement

Victoria E. Gibbon: Writing – review & editing, Writing – original draft, Visualization, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Athi Baliso: Writing – review & editing, Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Laura J. Heathfield: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Methodology, Formal analysis.

Ethical approval

Ethical approval for this study was granted by the Human Research Ethics Committees of the University of Cape Town (HREC REF: 263/2019), it also used data from the ethically approved FACT repository (HREC REF: R012/2019), which contains FACT case reports and affidavits.

Data availability statement

Data analysed during this study are included in this published article.

Funding sources

Baliso was supported by the University of Cape Town Vice Chancellor Master's Research Scholarship, Ada and Bertie Levenstein Bursary, University of Cape Town's Financial Aid Scheme and the South African National Research Foundation Master's Innovation Scholarship; Gibbon acknowledges support from the South African National Research Foundation. Their contribution towards this research is hereby acknowledged. Opinions expressed and conclusions arrived at, are those of the authors and not necessarily to be attributed to the NRF.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.