The Dangers of Data. Recognising the Limitations of Crime Statistics

This research paper critiques the reliability of non-police crime statistics in South Africa, specifically those from the Medical Research Council (MRC) and the National Injury Mortality Surveillance System (NIMSS). The author demonstrates how urban sampling biases and flawed statistical assumptions led to an overestimation of murder rates in rural areas (such as Limpopo) and inaccurate conclusions regarding the vulnerability of the 'Coloured' community. The paper concludes that while police data has limitations, the alternative datasets used in these cases were stretched beyond their limits, and SAPS figures should be preferred unless proven erroneous.

• The Medical Research Council (MRC) likely overstated the number of murders in South Africa for 2000/01 by using an outdated average for non-natural deaths. The MRC used a 12% average for non-natural deaths from 1997-2001, whereas the actual proportion in 2000 was 10%. Correcting this figure alone could reduce the estimated number of murders by nearly 5,000, potentially aligning the MRC's figures with those of the South African Police Service (SAPS).
• The National Injury Mortality Surveillance System (NIMSS) data is heavily biased toward urban areas, which leads to inaccurate murder rate estimates in rural provinces. In 2001, 62% of bodies surveyed by NIMSS were from Gauteng and Western Cape mortuaries, despite these provinces housing only 38% of the population. This urban bias resulted in the MRC estimating the number of murders in Limpopo to be nearly three times higher than the figures reported by the SAPS.
• Claims that the homicide rate in the 'Coloured' community is significantly higher than the national average are based on a misuse of NIMSS raw data. Because NIMSS does not extrapolate its findings to the general population and its sample is biased toward urban areas where Coloured populations are concentrated, the data cannot be used to establish risk distribution. A test using provincial population proportions suggested that Coloured victims were actually under-represented in the 2001 NIMSS sample.
• There is a significant disparity between the murder rates reported by the SAPS and those estimated by the MRC for 2000/01. The MRC's age-standardised rate was 66.6 murders per 100,000 people (30,069 murders), while the SAPS reported a rate of 49.8 per 100,000 (21,785 murders). The author argues that the MRC's use of age-standardised rates further exaggerates this difference.
• The author identifies specific areas of criminal justice policy where empirical data is often not decisive because decisions are driven by moral or value-based stances rather than evidence. Examples include the death penalty, prostitution, and narcotics, where prohibition may cause greater problems than the activities themselves, yet enforcement is demanded based on moral positions.

APA

Altbeker, A. (2005). The Dangers of Data. Recognising the Limitations of Crime Statistics. Institute for Security Studies. https://issafrica.org/topics/crime-and-criminal-justice/01-dec-2005-the-dangers-of-data.-recognising-the-limitations-of-crime-statistics-antony-altbeker

Chicago

Altbeker, Antony. The Dangers of Data. Recognising the Limitations of Crime Statistics. Institute for Security Studies, 2005. https://issafrica.org/topics/crime-and-criminal-justice/01-dec-2005-the-dangers-of-data.-recognising-the-limitations-of-crime-statistics-antony-altbeker.

Wikipedia

{{cite report |last1=Altbeker |first1=Antony |title=The Dangers of Data. Recognising the Limitations of Crime Statistics |publisher=Institute for Security Studies |date=1 December 2005 |url=https://issafrica.org/topics/crime-and-criminal-justice/01-dec-2005-the-dangers-of-data.-recognising-the-limitations-of-crime-statistics-antony-altbeker |access-date=17 August 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{altbeker2005dangers, author = {Altbeker, Antony}, title = {{The Dangers of Data. Recognising the Limitations of Crime Statistics}}, institution = {Institute for Security Studies}, year = {2005}, month = dec, url = {https://issafrica.org/topics/crime-and-criminal-justice/01-dec-2005-the-dangers-of-data.-recognising-the-limitations-of-crime-statistics-antony-altbeker}, urldate = {2026-08-17}, note = {Indexed by Climate Insights Directory} }