Scam data needs rigour, not racial narratives

Framing scam trends along ethnic lines risks flawed conclusions and harmful perceptions. Crime data must be analysed with scientific rigour and communicated with care, writes Datuk Dr P. Sundramoorthy.

Recent reports highlighting scam victims by ethnicity raise an important yet sensitive issue: how crime data is interpreted, framed and communicated to the public.

A recent MySinchew report, quoting Penang Commercial Crime Investigation Department deputy chief DSP Pang Meng Tuck, suggested Indians were less likely to fall victim, mainly because they ask too many questions for scammers to handle.

Such observations may be anecdotal, but broader conclusions must be grounded in scientific, methodologically sound analysis.

In a diverse society like Malaysia, where ethnicity, socio-economic status and geography intersect, the risk of misinterpretation is high.

First, crime data must be assessed within a rigorous empirical framework. The science of crime analysis, rooted in criminology, statistics and data science, requires more than surface-level observations. Claims such as scammers “giving up” on particular groups remain speculative unless supported by systematic evidence.

Reliable conclusions require clear sampling strategies, representative datasets and appropriate statistical techniques such as regression analysis and hypothesis testing. Longitudinal data is equally important, as crime patterns evolve due to policy changes, technological shifts and enforcement strategies. A single snapshot risks capturing noise rather than meaningful trends.

Equally important is triangulation, a cornerstone of scientific inquiry. Crime data should not rely solely on police reports, which often underrepresent incidents due to underreporting. Instead, it should be cross-validated with victimisation surveys, financial fraud databases and behavioural research.

Victimisation surveys help uncover the “dark figure” of crime, while behavioural studies explain how individuals respond to scam attempts. Without such triangulation, analysis risks being incomplete or biased.

Second, descriptive statistics must present a holistic, context-sensitive picture. Malaysia’s multi-ethnic composition, including Malays, Chinese, Indians and indigenous groups, comes with distinct demographic and socio-economic profiles.

For instance, Penang has a higher proportion of ethnic Chinese than the national average. This alone can influence the apparent distribution of scam victims. Higher numbers within a group may reflect population proportions rather than targeted victimisation.

This is where base rates matter. The base rate fallacy occurs when underlying population proportions are ignored. Proper analysis should standardise victimisation rates, such as victims per 100,000 individuals within each group, before making comparisons.

Confounding variables must also be controlled. Factors such as age, income, urbanisation, education and digital literacy significantly influence vulnerability. Multivariate models help isolate these effects, ensuring ethnicity is not mistaken for causation when it may only correlate with other factors.

Third, the risk of reinforcing stereotypes must be managed. In criminology, the social construction of crime is as important as its measurement. Narratives that loosely link crime patterns to ethnicity can stigmatise communities or create complacency.

Suggesting one group is “harder to scam” may reduce vigilance, while portraying another as more vulnerable may lead to victim-blaming.

From a behavioural science perspective, scams are opportunistic and adaptive. Offenders target individuals based on opportunity, accessibility and potential gain, not fixed characteristics.

As technology evolves, so do scam tactics, from phishing emails to complex social engineering schemes. Vulnerability is shaped by dynamic factors such as digital exposure, trust heuristics and situational awareness, not ethnicity alone.

Fourth, the communication of crime data carries ethical responsibility. Media organisations shape public understanding. Risk communication emphasises clarity, context and caution.

Reports should state limitations clearly, avoid overgeneralisation and provide context. Sensational or reductive framing, especially along ethnic lines, can distort perception and undermine social cohesion. More useful reporting would shift focus from “who is more likely to be scammed” to “what increases or reduces vulnerability”. Research consistently highlights digital literacy, scepticism towards unsolicited communications and awareness of scam tactics.

Community education, multilingual campaigns and targeted outreach to higher-risk groups offer more meaningful outcomes than ethnicity-based narratives.

Finally, a scientific approach must prioritise prevention over anecdote. Evidence-based policymaking depends on actionable insights from robust data.

In the context of scams, this includes strengthening cybersecurity, enhancing cross-border cooperation, improving reporting mechanisms and using data analytics to detect emerging fraud patterns. Predictive modelling, when used responsibly, can support proactive enforcement.

At the same time, inclusivity must remain central. Policies should benefit all segments of society without reinforcing divisions.

In a plural society like Malaysia, trust and cohesion are essential for effective crime prevention. Misinterpreting data risks flawed policy and social division.

While the MySinchew report provides a starting point, it underscores the need for disciplined, data-driven interpretation of crime statistics. Crime data shapes narratives, informs policy and influences behaviour.

A scientifically grounded, context-aware and responsible approach is essential to building a safer and more cohesive society.

The views expressed here are the personal opinion of the writer and do not represent that of Twentytwo13.