Skip to main content

Fraudsters target the elderly: Behavioural evidence from randomised controlled scam-baiting experiments

Robinson, Jemima ; Edwards, Matthew (2024) — Security Journal

View Original Source (opens in a new tab)

DOI: 10.1057/s41284-023-00410-4 (opens in a new tab)

Type:   Journal Article
Tags:   AI-Enabled Fraud, Automated Responders, Conversation Length, Email Fraud, Measurement, Offender Tactics, Offender Targeting, Older Adults, Platform Policy, Psychology, Romance Scams, Scam Baiting, Victim Experience, Victim Vulnerability
Citation tools
Download RIS Download BibTeX 


              
              

            

          

        


      

      

    

        
Transparency

        
Evidence and review status

        
This page contains AI-generated content. No human content review or subject-matter-expert review is recorded.

    


    

                    

                
Source basis

                

                    Downloaded PDF
                

            

                    

                
AI-generated page content

                

                    Yes
                

            

                    

                
Automated checks

                

                    Passed
                

            

                    

                
Administrative approval

                

                    Yes
                

            

                    

                
Human content review

                

                    Not recorded
                

            

                    

                
Subject-matter-expert review

                

                    Not recorded
                

            

            


    

        
            How this was prepared
        
        

                            

                    
Source basis

                    

                    
RSRC downloaded and privately stored a copy of the paper for internal analysis. The PDF is not offered to viewers from this page.


                                            
    
                                                            
  • 
                                    PDF added to RSRC:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
AI-generated page content

                    

                    
AI-generated research notes displayed on this page: Synopsis, Identified gaps, Methods, Limitations, Future work. The paper itself is not described as AI-generated.


                                            
    
                                                            
  • 
                                    Document analysis recorded:
                                    
                                
    • 
                                                            
  • 
                                    Synopsis generation recorded:
                                    
                                
    • 
                                                            
  • 
                                    Page record updated:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
Automated checks

                    

                    
The current, source-bound synopsis passed the recorded versioned publication checks.


                                            
    
                                                            
  • 
                                    Checks completed:
                                    
                                
    • 
                                                    

                    
                                            

                            
                                View passed checks (3)
                            
                            
    
                                                                    
  • Length, completeness, repetition, refusal, boilerplate, and active-markup screening
    • 
                                                                    
  • Numerical claims checked against the available source text
    • 
                                                                    
  • English-source lexical grounding check
    • 
                                                            

                        

                                        

                

                            

                    
Administrative approval

                    

                    
An authenticated administrator approved the bibliographic record for public Library display. This is not a review of every research claim.


                                            
    
                                                            
  • 
                                    Approved for public display:
                                    
                                
    • 
                                                    

                    
                                        

                

                            

                    
Human content review

                    

                    
No human review is recorded for the AI-generated content displayed on this page.


                    
                                        

                

                            

                    
Subject-matter-expert review

                    

                    
RSRC has not recorded review of this content by a subject-matter or methods expert.


                    
                                        

                

                    

    


    

        Review-state definitions

        Found a possible error?
        Request a correction.
    



    



              

        
                  

            
Synopsis

            
This study investigates which purported victim presentations in email-based fraud attract online scammers, using an automated scam-baiting platform to collect direct behavioral evidence from offenders. The authors design four distinct victim personas drawn from prior literature and a small public survey, and they embed these personalities in reply systems that engage scammers across common scam formats, including transactional, non-transactional, lottery, and love scams. The project aims to shed light on which characteristic profiles may hold the scammers’ interest and how that interest translates into continued engagement. The four personalities are Doris, an elderly, trusting widow; Alex, a naive young person; Dave, a rude, middle-aged man; and Sam, a professional business figure. A control condition uses Chen et al.’s randomized, template-based classifier and template responses. Conversations with real scammers were conducted over a defined period, and the study tracks initiation rates and conversation length for each personality. The researchers report that Doris generally attracted higher engagement and produced longer conversations, while Dave tended to underperform relative to the control. The other two personas showed mixed patterns in initiation and duration. Validation steps and ethical safeguards are described, including measures to avoid exploiting or humiliating scammers and to sanitize logs. The authors discuss potential interpretations of their findings and acknowledge methodological considerations, including the artificial nature of scripted interactions and the limits of generalizing from automated replies to real-world victimization risk. They frame the results as contributing to understanding which victim presentations scammers perceive as viable, with implications for designing preventative countermeasures and for refining fraud-awareness messaging, while noting boundaries in what the evidence can conclusively establish.

          

        
                  

            
Identified Gaps

            
Prior fraud-victim research is conflicted and often relies on self-reported victimization, which may be distorted by demographic differences in reporting and cybercrime underreporting. It remains difficult to distinguish victim susceptibility from offenders' selective targeting. The study also identifies limited access to representative mid-conversation scammer communications. Existing romance-fraud victim profiles may not generalize to other email-fraud types.

          

        
                  

            
Methods

            
The authors conducted a one-month randomized automated email scam-baiting experiment. Real fraudsters were assigned to one of four scripted pseudo-victim personalities or a random-response control. The personalities were informed by prior literature, a 92-person public-perception survey, and a 23-person validation survey. After excluding likely automated scammer conversations, the analysis covered 296 scammers and 1,416 scammer replies. Outcomes were reply-initiation rate and conversation length; personality-control differences in conversation length were tested with one-tailed Mann–Whitney U tests.

          

        
                  

            
Limitations

            
Each personality combined multiple characteristics, so the study cannot attribute engagement differences to age, gender, trust, loneliness, technical literacy, or other individual cues. Personality designs may reflect the authors' cultural biases; the English-grandmother cues of Doris may not be interpreted consistently by scammers from other backgrounds. Template-based replies eventually repeat and reveal automation, and personalities may be recognized or shared among scammers. Few love-fraud emails were observed, limiting conclusions related to romance fraud.

          

        
                  

            
Future Work

            
Run larger, longer experiments with a tapering-off period to improve statistical power and separate delayed replies from non-engagement. Isolate individual personality facets while holding others constant, including comparisons among elderly presentations. Test whether scammers respond differently to religious signals or degrees of religiosity. Expand template banks or use generated text after a credible pretext is established to avoid repetitive messages and improve human realism.

          

              

    
    

        
Related Research Insights

        

            See how this publication connects to RSRC's living evidence syntheses through current citations and research-topic mapping.