Skip to main content

Evidence explained

From scam reports to enforcement

Platforms can remove scam accounts, researchers can trace transactions, and police can pursue offenders across borders. But removing a profile, identifying a money trail, and securing a conviction are different achievements. The research reviewed here documents enforcement activity without establishing which responses reliably reduce romance-scam losses (Button et al., 2025; Gądzik, 2024; Reiter & Team, 2024).

Romance Scam Research CenterEvidence snapshot: September 13, 20266 min read

Explanation updated . This can reflect an editorial correction without a new evidence snapshot.

AI-assisted explanation · Automated editorial and source checks. Preparation and review details

What different responses can show

Interpretation. For teams responsible for platform safety or fraud response, the useful comparison is not simply which tool performs well. Each response addresses a different problem, and its evidence stops at a different point (Botha et al., 2025; Button et al., 2025; Gądzik, 2024; Wang, 2024).

What different responses can show
FocusWhat the studies showWhat remains uncertain
Account removal Legal review and descriptions of scam tactics

Platform rules can lead to temporary or permanent account blocking when users misuse another person’s identity or impersonate someone else (Gądzik, 2024).

Removing a dating profile may not end contact: offenders move conversations to messaging or email so they can continue independently of the platform (Thiel, 2020).

Transaction tracing Blockchain analysis and documentary research

Researchers combined blockchain records—digital ledgers of cryptocurrency transactions—with documentary research to identify service providers and connections between scam proceeds (Reiter & Team, 2024).

A separate case needed exchange-held information about account holders and outgoing transactions to investigate beyond the money trail; no response had arrived (Botha et al., 2025).

Joint enforcement Review of cross-border policing and operation reports

An international operation covering several fraud types, including romance scams, reported arrests, frozen accounts, raids, and intercepted funds (Button et al., 2025).

Interpretation. These combined operation totals do not isolate romance-scam outcomes or show how much subsequent fraud was prevented (Button et al., 2025).

Sentencing patterns Statistical analysis of court documents

In selected Chinese judgments on romance-linked investment fraud, group leadership, overseas offending, and larger defrauded amounts were associated with longer fixed-term sentences (Wang, 2023).

Missing information about courts, judges, representation, and defendants means these associations cannot fully explain why sentences differed (Wang, 2023).

A removed profile may leave the relationship intact

Offenders who make contact through dating sites may move conversations to direct messaging or email soon afterward. This lets them maintain contact even if the dating platform exposes them and blocks their profile (Thiel, 2020).

Conviction records offer a different view: which platforms appeared in prosecuted cases. A Nigerian study examined 50 romance-scam conviction cases from 2023–2024 at one regional enforcement command. Facebook appeared in 46% of cases (23 offenders). This documents platform use, not migration. The single-command, conviction-only sample excludes unreported scams and investigations without convictions; it cannot represent Nigeria as a whole (Soares & Lazarus, 2024).

A money trail is not yet an identified offender

Blockchain records can reveal connections without resolving every question about responsibility. One tracing study found that scam proceeds passed through the same parties at similar times and in similar ways. The authors assessed the flows as likely belonging to the same overall group, rather than claiming complete proof of its activities (Reiter & Team, 2024).

A 2025 cryptocurrency investment-scam case shows where technical analysis can stop. Researchers sought personal information and outgoing transaction records from Binance to investigate account holders and transfers beyond the blockchain trail. They planned further analysis once those records arrived. At publication, the request needed action by a UK law-enforcement entity, and no response had been received (Botha et al., 2025).

Large transaction totals require similar care. Another study estimated that an average of $27.8 billion a year entered likely scammer-controlled deposit accounts through the Ethereum cryptocurrency network during 2021–2023. Some funds could reflect other activity by the same or related networks, the authors cautioned. The estimate is not a clean total of romance-scam losses (Griffin & Mei, 2024).

International raids show action, not lasting prevention

Operation First Light 2022 involved 76 countries and targeted telecommunications and scamming fraud, including romance scams. Reported results included 1,770 locations raided, 4,000 bank accounts frozen, 2,000 arrests, and $50 million in illicit funds intercepted. These figures describe a broad operation, not romance-scam-only enforcement (Button et al., 2025).

The distinction matters because disruption can be temporary. Research from Vietnam’s perspective describes operations relocating after enforcement, few arrests or prosecutions of senior leaders, and limited international legal assistance. It also acknowledges that the specific activities of the highest-ranking organizers could not be legally confirmed (Luong & Ngo, 2024).

Interpretation. People found working in scam operations also require different assessments. Research on trafficking into cybercrime calls for reliable ways to distinguish coerced participants from willing offenders. Counting everyone encountered as an equivalent perpetrator would miss the possibility that some were themselves exploited (Sarkar & Shukla, 2024).

A legal basis does not guarantee a workable case

A 2025 Indonesian legal review identified a basis for prosecuting love scams, but also difficulties tracing identities, reaching offenders abroad, and obtaining reports from victims fearing shame or judgment. Missing reports also affect what official records reveal. In Korea, a 2019 study reported that absent official statistics and high underreporting obscured the scale of romance scams—a limitation specific to that country and time (Adnan, 2025; Kim & Seo, 2019).

A 2024 Ecuadorian legal analysis argued that the law against impersonating another person did not specify its application online. It recommended explicitly including catfishing—deceiving people through a false online identity—in the criminal code. This was a legal argument and reform proposal for Ecuador, not evidence that a new offense would improve enforcement (Coronel-Rivera et al., 2024).

Interpretation. For policy planning, legal coverage and investigative capacity should therefore be examined separately. The Indonesian review also described inadequate technical capability and weak international cooperation. Changing an offense definition would not, by itself, answer how investigators obtain records or identify people operating abroad (Adnan, 2025).

Court studies describe the cases that reached sentencing

Court records offer a view of cases that reached sentencing. A study of Chinese romance-linked investment fraud analyzed 172 verdicts dated May 2019 to March 2022. After screening for relevant fraud convictions involving romantic swindling, the sample contained 531 offenders. Their average fixed-term imprisonment was 39 months, and 6% received probation (Wang, 2023).

Group leaders and offenders linked to larger defrauded amounts received longer sentences in the study’s statistical analysis. But the documents did not show each person’s share of money obtained collectively, making individual responsibility harder to assess. Selective uploading or deletion of judgments could also bias the sample, so the available records may not reflect all relevant cases (Wang, 2023).

A U.S. study analyzed 148 sentencing memorandums obtained from an initial pool of 344 offenders. In the author’s analysis, guilty pleas were associated with shorter sentences. The outcome concerned length of imprisonment, not whether imprisonment occurred. Missing case and court information left other explanations for the association unresolved (Wang, 2024).

What response teams can change

Interpretation. These findings support more precise planning and evaluation, rather than a ranking of responses. Teams can distinguish what they directly control from later outcomes that depend on other organizations, jurisdictions, or evidence (Botha et al., 2025; Button et al., 2025; Thiel, 2020).

  1. Measure beyond account removal

    Platform teams should distinguish accounts removed from contact actually interrupted. When evaluating a response, ask whether communication continued elsewhere rather than assuming a blocked profile ended the scam (Gądzik, 2024; Thiel, 2020).

  2. Locate the stalled handoff

    Investigation managers should record where a case stops: identifying a service provider, obtaining account records, identifying a person, or pursuing action abroad. This would distinguish an analytical lead from usable evidence and make delays more visible (Botha et al., 2025; Reiter & Team, 2024; Shapiro, 2022).

  3. Separate activity from benefit

    Report raids, arrests and frozen accounts separately from money returned or losses prevented. State which fraud types and locations the figures cover. One Nigerian study covered 50 convictions from a single regional enforcement command; its authors recommended studying more commands before drawing broader conclusions (Button et al., 2025; Soares & Lazarus, 2024).

Which obstacle would this investment actually remove?

Before funding a new tool, legal reform, or partnership, ask which documented obstacle it addresses and what evidence would show progress beyond that point. Would success mean a removed account, an answered records request, an identified offender, or less harm to victims? (Adnan, 2025; Botha et al., 2025; Button et al., 2025; Thiel, 2020).

About the evidence

This explanation draws on selected sources from RSRC’s full synthesis. The full research record, its source list and version history remain available below.

Explore this topic’s methods and groups studied · Follow topic updates

Scope, preparation and limitations

Codex edited and reviewed this explanation from synthesis version 61. Changed passages were checked against preserved source excerpts; passing source checks were reused for unchanged passages. The editorial judgments are recorded as Codex review, not a new API grading run. They assess communication, not reader-test results or scientific certainty. No human or expert content review of this explanation is recorded.

The full synthesis preserves 30 publication records and 189 evidence statements. These counts do not establish the number of independent studies or the strength of a finding. Practical implications are editorial interpretations, with limitations explained alongside the evidence.

RSRC methodology
Sources cited in this explanation (15)

Links open the publication record or original work. Journal access may vary.

  1. Adnan, I. (2025). TINJAUAN YURIDIS TERHADAP KEJAHATAN LOVE SCAM DALAM PERSPEKTIF HUKUM PIDANA DAN UNDANG UNDANG INFORMASI DAN TRANSAKSI ELEKTRONIK DI INDONESIA [A Juridical Review of Love Scam Crimes from the Perspective of Criminal Law and the Electronic Information and Transactions Law in Indonesia]. AL-BALAD : Jurnal Hukum Tata Negara dan Politik Islam, 5(1), 55–71. https://doi.org/10.59259/ab.v5i1.246
    Publication recordOriginal work · DOIReturn to first citation ↑
  2. Botha, J. G., Singh, K., & Leenen, L. (2025). Analysis of a Cryptocurrency Investment Scam: Pig Butchering. European Conference on Cyber Warfare and Security, 24(1), 61–70. https://doi.org/10.34190/eccws.24.1.3558
    Publication recordOriginal work · DOIReturn to first citation ↑
  3. Button, M., Hock, B., Suh, J. B., & Koh, C. S. (2025). Policing cross-border fraud ‘Above and below the surface’: mapping actions and developing a more effective global response. Crime, Law and Social Change, 83(1). https://doi.org/10.1007/s10611-024-10186-2
    Publication recordOriginal work · DOIReturn to first citation ↑
  4. Coronel-Rivera, K. A., Arévalo-Vásquez, C. E., & Villavicencio, M. (2024). Análisis Jurídico del Catfishing en Transacciones de Comercio Telemático en Ecuador: Desafíos y Soluciones en el Contexto de la Seguridad Cibernética. MQRInvestigar, 8(4), 1311–1337. https://doi.org/10.56048/mqr20225.8.4.2024.1311-1337
    Publication recordOriginal work · DOIReturn to first citation ↑
  5. Griffin, J. M., & Mei, K. (2024). How Do Crypto Flows Finance Slavery? The Economics of Pig Butchering. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4742235
    Publication recordOriginal work · DOIReturn to first citation ↑
  6. Gądzik, Z. (2024). Catfishing – jako przejaw przestępstwa kradzieży tożsamości. Krytyka Prawa, 16(4). https://doi.org/10.7206/kp.2080-1084.735
    Publication recordOriginal work · DOIReturn to first citation ↑
  7. Kim, H.-S., & Seo, J.-B. (2019). 로맨스 스캠(Romance Scam) 범죄 현황 및 대응방안에 대한 고찰 [A Study on Romance Scam: The Current Situation and Effective Countermeasures]. The Police Science Journal, 14(3), 117–146. https://doi.org/10.16961/polips.2019.14.3.117
    Publication recordOriginal work · DOIReturn to first citation ↑
  8. Luong, H. T., & Ngo, H. M. (2024). Understanding the Nature of the Transnational Scam-Related Fraud: Challenges and Solutions from Vietnam’s Perspective. Laws, 13(6), 70. https://doi.org/10.3390/laws13060070
    Publication recordOriginal work · DOIReturn to first citation ↑
  9. Reiter, J., & Team, B. (2024). Connecting Chinese and American Scam Victims. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4687064
    Publication recordOriginal work · DOIReturn to first citation ↑
  10. Sarkar, G., & Shukla, S. K. (2024). Bi-Directional Exploitation of Human Trafficking Victims: Both Targets and Perpetrators in Cybercrime. Journal of Human Trafficking, 1–22. https://doi.org/10.1080/23322705.2024.2353015
    Publication recordOriginal work · DOIReturn to first citation ↑
  11. Shapiro, L. R. (2022). Online Romance Scammers. Cyberpredators and Their Prey, 17–41. https://doi.org/10.4324/9781003092292-2
    Publication recordOriginal work · DOIReturn to first citation ↑
  12. Soares, A. B., & Lazarus, S. (2024). Examining fifty cases of convicted online romance fraud offenders. Criminal Justice Studies, 37(4), 328–351. https://doi.org/10.1080/1478601x.2024.2429088
    Publication recordOriginal work · DOIReturn to first citation ↑
  13. Thiel, C. (2020). Liebesschwindel im Cyberspace [Love scam in cyberspace]. Cyberkriminologie, 241–267. https://doi.org/10.1007/978-3-658-28507-4_10
    Publication recordOriginal work · DOIReturn to first citation ↑
  14. Wang, F. (2023). Sentencing Disparity and Focal Concern: An Assessment of Judicial Decisions on Sha Zhu Pan Cases Collected From China Judgements Online. Crime & Delinquency, 71(6-7), 1845–1877. https://doi.org/10.1177/00111287231158571
    Publication recordOriginal work · DOIReturn to first citation ↑
  15. Wang, F. (2024). A Quantitative Analysis on the Sentencing Disparity in Online Romance Scam Cases in United States: A Focal Concern Theory Perspective. Crime & Delinquency, 71(11), 3478–3506. https://doi.org/10.1177/00111287241266561
    Publication recordOriginal work · DOIReturn to first citation ↑
Full synthesis and revision history

Download the full synthesis PDF

Reviewed studies describe a recurring operational pattern in relationship-based online fraud: offenders commonly initiate contact on dating or social platforms and then shift conversations to less-moderated messaging applications to avoid profile removal and tracking. Dating applications assessed in the corpus consistently requested phone numbers, email addresses, birth dates, and photos. Cryptocurrency use complicates investigations because transactions are harder to monitor and attribute to individuals, and at least one attempted subpoena to a major exchange yielded no response, preventing analysis based on account-holder information and outgoing transactions. Cross-border investigations encounter friction at many stages, and there is no specific global sanctions or fine regime for cross-border fraud with little evidence of existing regimes being used against such fraudsters. Investigations are further hindered when usernames or IP addresses are invalid or when the scam and bank are outside the investigating country. The corpus employs heterogeneous methods, including expert two-hour interviews, datasets of victim–scammer conversations, multi-year analyses of reported romance fraud, and coded court-judgment analyses using regression; automated methods such as large-language-model URL extraction in the studies achieved high F1 performance (0.911–0.918). Some work identifies gaps and proposes future directions, including detecting offender communication characteristics and implementing special address-verification methods. Estimates of aggregate financial impact in the literature may capture funds associated with other activities by the same or related criminal networks (Annadorai et al., 2020; Botha et al., 2025; Buil-Gil & Zeng, 2021; Button et al., 2025; Cross, 2023; Griffin & Mei, 2024; Ko et al., 2026; Noah et al., 2024; Shaari et al., 2019; Shapiro, 2022; C. Wang, 2022; F. Wang, 2023; Whitty, 2013).

Offender pathways, platform migration, and evidence capture

Multiple qualitative and case‑based studies describe a common operational sequence in relationship‑based fraud: offenders often initiate contact on dating or social platforms and then move conversations to less‑moderated channels (instant messaging, email, third‑party apps) shortly after first contact. Several preserved accounts explicitly state this relocation to direct messaging or email; such moves can make platform‑hosted traces harder to recover and, in some cases, may complicate investigators’ efforts to preserve provenance for prosecutions. A recent Nigerian case series analyzed 50 convicted romance‑scam cases secured between 2023 and 2024 from a single EFCC Zonal Command and found Facebook cited in 46% of those cases (23 offenders). That case series provides a concrete platform‑use example but is limited to convictions from one Zonal Command and does not represent all Nigerian offenders, unreported scams, or investigations that did not lead to conviction; the study’s authors recommend larger samples drawn from multiple EFCC Zonal Commands for broader inference (Annadorai et al., 2020; Soares & Lazarus, 2024; Thiel, 2020).

Empirical platform and conversation‑level analyses show that scammers use staged grooming and progressive persuasion to create investment or money‑transfer opportunities; studies using authentic victim‑scammer chat corpora document local accomplices, telephone interactions, and evidence loss when victims do not retain call records, which further constrains attribution and prosecution (Shaari et al., 2019; F. Wang & Zhou, 2022).

Platform governance, registration, reporting, and technical detection

App assessments in the corpus report that all assessed dating applications required phone numbers, email addresses, birthdates, and photos, and that some collected substantially more intimate information. The authors of the dating-site review described their survey as quick and informal, which limits broad generalization (Noah et al., 2024; Sorell & Whitty, 2019).

Technical detection work included large-scale measurement using a dataset of 510,503 scam accounts collected over an 11-month period (2012–2013), and analysis of monthly romance-fraud reports known to UK Action Fraud from April 2014 through December 2020. These studies note a technical limitation: moderator IP- and proxy-based signals can be obscured through unregistered proxies or compromised residential hosts, which can reduce detection reliability and produce false negatives (Buil-Gil & Zeng, 2021; Huang et al., 2015; Suarez-Tangil et al., 2019).

Law‑enforcement investigation, prosecution, and sentencing realities

A consistent theme across police‑perspective, court‑judgment and convict‑case research is the impact of anonymity, overseas operations, and jurisdictional barriers on investigation and prosecution: invalid or obscured usernames/IPs, offenders operating from other countries, slow mutual legal assistance, and limited specialist resources impede attribution and case progression (Button et al., 2025; Luong & Ngo, 2024; Shapiro, 2022; Thumboo & Mukherjee, 2024).

Sentencing and conviction studies provide jurisdiction-specific evidence that sentencing outcomes vary with role, plea, and offense characteristics; for example, U.S. sentencing analysis found guilty pleas associated with shorter imprisonment lengths, not a lower likelihood of imprisonment itself. Older or female defendants also had reduced imprisonment time. These associations do not establish causal effects. Comparable focal-concern analyses of pig-butchering judgments in China report associations between principal operational roles, overseas offending, defrauded amounts and longer fixed-term sentences. The Chinese study cautions that its relationships may be correlational because legal documents omitted potentially relevant variables, including attorney presence, court level, judges’ attitudes, and defendants’ background and health (F. Wang, 2023, 2024).

Blockchain forensics, OSINT, and dependencies on off‑chain cooperation

On‑chain tracing studies and forensic case reports show that blockchain analysis can identify deposit address clusters, intermediary exchanges, and role‑based transaction patterns that provide operational leads; several case studies estimate large flows through likely scammer‑controlled addresses but caution that seed‑address incompleteness and indirect attribution limit definitive conclusions about total funds stolen without off‑chain corroboration (Griffin & Mei, 2024; Reiter & Team, 2024).

Multiple case studies document that subpoenas and exchange cooperation are often necessary to convert on‑chain seeds into account‑level attribution, and that law‑enforcement requests to exchanges can be slow or unresponsive—delays that hinder freezing, seizure, or recovery of victim funds and constrain investigators’ operational options (Botha et al., 2025).

Policy‑mapping and country‑focused legal reviews repeatedly describe cross‑border cooperation as essential but fragmented: coordinated international operations have achieved concrete disruptions (raids, arrests, account freezes), yet systematic gaps exist in global intelligence, accredited training, coordinated forums, sanctions targeting cross‑border fraudsters, and well‑resourced investigative capacity (Button et al., 2025; Sarkar & Shukla, 2024).

Doctrinal studies from multiple jurisdictions document inconsistent legal characterizations of catfishing and romance‑initiated investment fraud, the absence of dedicated statutory offences in some countries, and practical limits on mutual legal assistance—leading many authors to propose harmonized offences, clearer prosecutorial guidance, improved cross‑border MLAT capacity, and formal public–private cooperation mechanisms for rapid exchange of evidence and intelligence (Button et al., 2025; Coronel-Rivera et al., 2024; Gądzik, 2024; Hasibuan & Syam, 2023; “The Crime of "Pig-butchering Scams" in the Securities Market and...”, 2023).

Methodological patterns, limitations, evidence gaps, and priority directions

The reviewed evidence includes a conversation dataset of 60 victim–scammer exchanges (50 volunteered via a Romance Scam Research Malaysia Facebook page and 10 via collaboration with the Royal Malaysia Police CCID) and an analysis of monthly romance‑fraud reports to Action Fraud from April 2014 through December 2020. One paper recommends an integrated approach combining conventional criminal law and cyber law for love‑scam enforcement. Identified limitations in the reviewed work include that a $27.8 billion annual estimate may capture funds from other activity by the same or closely related criminal networks, and that observed relationships in legal‑document analyses may be correlational because those documents lacked variables such as attorney presence, court level, judges’ attitudes, and defendants’ background and health. In the Korean context, a 2019 study reported absent official statistics and substantial hidden crime as obstacles to determining the true scale of romance scams; this is a dated, country-specific limitation, not evidence that official statistics are universally unavailable today (Adnan, 2025; Buil-Gil & Zeng, 2021; Griffin & Mei, 2024; Kim & Seo, 2019; Shaari et al., 2019; F. Wang, 2023).

Across the corpus, authors identified these priorities supported by the attached evidence: validating platform technical capabilities—LLM-based URL extraction achieved F1 = 0.911–0.918; documenting friction in cross-border investigations that hampers criminal prosecution efforts; noting that no specific global sanctions or fines regime exists for cross-border frauds and that there is little evidence existing regimes are used against such fraudsters; and flagging that the $27.8 billion annual estimate may include funds from other activity by the same or closely related criminal networks. The Nigerian conviction-case study recommends larger samples drawn from multiple EFCC Zonal Commands to extend its limited geographic coverage (Button et al., 2025; Griffin & Mei, 2024; Ko et al., 2026; Soares & Lazarus, 2024).

References

This AI-assisted synthesis is based on reviewed evidence records. It may contain errors or omissions. Follow the publication and DOI links, consult the original works, and make your own judgment about the evidence.

Revision history

  1. Version 61

    Current public version

    Incremental evidence update. A limited revision is warranted. Add the Nigerian platform-use finding with its single-command, conviction-only sampling limits, and incorporate the dated Korean measurement limitation without generalizing it globally. Correct synthesis claim 7 to describe the U.S. outcome as imprisonment length, retaining the accurate reader explanation. The central distinction between enforcement activity and demonstrated prevention remains intact. Several corrected Korean excerpts...

  2. Version 60

    Incremental evidence update. A limited revision is warranted. Add the Nigerian platform-use finding with its single-command, conviction-only sampling limits, and incorporate the dated Korean measurement limitation without generalizing it globally. Correct synthesis claim 7 to describe the U.S. outcome as imprisonment length, retaining the accurate reader explanation. The central distinction between enforcement activity and demonstrated prevention remains intact. Several corrected Korean excerpts...

    • Publications: 30 (+2)
    • Evidence statements: 189 (+8)
    • Cited sources: 30 (+2)

    Sources added: A Study on Romance Scam: The Current Situation and Effective Countermeasures; Examining fifty cases of convicted online romance fraud offenders.

    The approved evidence set changed while the cited-source list remained the same.

  3. Version 59

    Automated claim-level support repair round 2, correcting 1 passage from synthesis version 58.

View full revision history

Looking for reporting or support information? Visit the RSRC resource list. RSRC does not provide direct support.

Request a correction

Back to top ↑