In their 2022 publication in this journal, Suber and colleagues attempt to apply crime mapping to the illicit trade of cultural objects from the Middle East to establish a causal relationship between conflict and heritage looting. The article calls for comments by readers on the methodological approaches and results (p. 559). This commentary addresses the article’s shortcomings, specifically highlighting its inadequate grounding in existing literature, methodological limitations, and problematic data approach.

Firstly, despite applying a spatial criminological approach, crime-mapping, to the illicit antiquities trade, the article does not cite much criminological research on this topic. Of the three sources cited specifically on crime mapping, two are general audience textbooks. A broad literature exists on crime mapping and crime science applied to conflict and analogous crime contextsFootnote ¹ and to cultural heritage specifically.Footnote ² The authors also ignore the more recent literature on satellite remote sensing and its application to looting since Brodie and Contreras’Footnote ³ piece on using Google Earth Pro. The field had already advanced significantly by 2022, with work using high-resolution satellite imagery to map looting,Footnote ⁴ tracking looting in conflict areas spatially and temporally,Footnote ⁵ and using open-source data and machine learning approaches.Footnote ⁶ This field has continued to advance at a rapid pace across these areas since 2022, making the missing discussion of the literature at the time even more problematic, as it misrepresents the field.Footnote ⁷ The article’s claim that “hardly any research” (p. 534) has been done to “map the most statistically significant locations of antiquities crime and correlate it with the analyses from satellite imagery” (p.534) neglects the substantial interdisciplinary research available.

Secondly, the study lacks crucial methodological and analytical details, including the synthesis of ASOR data, ethical considerations, and validation of incident locations. These omissions are particularly significant given the need for further rigorous empirical research in this highly politicized area.Footnote ⁸ Scholars who have worked extensively with the ASOR reports, both at the timeFootnote ⁹ and since,Footnote ¹⁰ understand that the methodology and coding practices used for heritage damage and destruction incidents carry substantive implications for the analyses of looting attempts or trafficking of antiquities. For example, the authors fail to explain their choice of grid cell size for the kernel density estimation (KDE) hot spot analysis, a decision that substantially impacts the results, as larger grid cells increase the likelihood of finding statistically significant hot spots but reduce analytical utility, while smaller grid cells offer more precision but may yield fewer significant results. The importance of these missing discussions is evident in their statistical analysis. It appears from Appendix 2 that, while they had the geolocation of incidents and the ability to produce grid cells (see Figure 11), they chose to do a more aggregate statistical analysis at the district level. While negative binomial regression is a valid approach, it ignores the spatial variation that they previously found in the hotspot analysis (pp. 547–48). It does not appear that their model can account for this underlying spatial variation (for example, through a spatial fixed effect). Additionally, negative binomial regression coefficients are more appropriately discussed as incident rate ratios (IRRs) rather than log odds.Footnote ¹¹

The authors’ apparent omission of a time component in their ASOR-CHI data analysis is problematic. This oversight contradicts the dynamic nature of the Syrian conflict (pp. 537–38) and ignores research published at the time,Footnote ¹² emphasizing the importance of a time component in mapping antiquities looting. This importance has been continually reiterated in recent scholarship.Footnote ¹³ The claim of missing time variables (p. 548) is questionable, as ASOR-CHI reports typically include both publication dates and estimated incident dates.Footnote ¹⁴ This raises further questions about their methodological approach: conducting a hotspot analysis without temporal data compromises the reliability and accuracy of the results as “hotspots” are statistically significant concentrations of crime events in space and time. These methodological flaws significantly undermine the study’s value and findings. A more methodologically robust approach would involve point pattern analysis (a framework designed to account for underlying hidden spatial processes),Footnote ¹⁵ incorporating multiple scales of analysis to provide a more comprehensive and consistent interpretation of data to triangulate findings.

Lastly, and most importantly, the article does not actually confirm “the utility of data-driven research methods in collecting and analyzing crime data related to antiquities trafficking” (p. 557), as this necessitates data on the movement of the objects and, ideally, the people involved,Footnote ¹⁶ which are both missing from this article. Instead, the measure of trafficking is constructed of counts of “all locations in Syria where antiquities looting, theft, or destruction was happening” by district (p. 536). While the removal of objects via theft or looting – if confirmed (you cannot guarantee anything was removed from a picture of a looting pit) – may be part of trafficking, that is not always true. Here, the best the authors can say is that they can speak to patterns of looting attempts, theft, or destruction. It is even more of an overstatement to assert that “findings of this study confirm literature arguing that war, just like poverty, motivates people and organized groups to loot heritage sites either to guarantee individual/familiar survival or to fund insurgency” (p. 557). There is limited empirical data available that confirm organized criminal involvement in antiquities trafficking, especially related to terrorism,Footnote ¹⁷ none of which are critically assessed in this article.

Despite the urgent need for expanded empirical research to address the looting and trafficking of cultural objects, it is crucial that the quality of scholarly work and publications is not compromised. Open-source data – meaning data that are openly accessible, editable, and freely distributable – can provide critical insights that can assist to strengthen efforts against antiquities looting and trafficking, but its effectiveness depends on ensuring research is both accurate and actionable. While the authors assert that any rigorous research on antiquities trafficking and heritage loss in politically and economically unstable regions represents a meaningful safeguarding effort (p. 559), the actual methodological rigor and empirical accuracy of such work remain paramount to its impact.