Bloodstain Pattern Analysis: What Can the Pattern Support?

Standards do not all address accuracy. The ASB terminology document and its SWGSTAIN predecessor define labels such as impact and cast-off patterns. Hook et al. reported in 2024 that there was still no national or international classification method, but ten schemes across fourteen sources. Different analysts may therefore classify the same pattern differently. The authors also note that mechanistic labels name a cause before the sorting task is complete.

Different mechanisms may produce similar or indistinguishable patterns. Dror made that point in a 2023 keynote to IABPA. Coughing and impact, or dripping and impact, may leave overlapping appearances. A finding that a pattern is consistent with a beating has little value unless other mechanisms producing the same observations are evaluated. The former low-, medium- and high-velocity scheme was abandoned because stain-size ranges overlapped across mechanisms.

Taylor et al. examined surface effects. Experienced court-qualified analysts were wrong in 13.1% of classifications on rigid smooth surfaces and 23.4% on fabric. More than half of fabric errors involved satellite stains from dripping being interpreted as blunt-force impact; dripped blood was classified as an attack pattern in 85% of those examples. That distinction may affect whether blood on clothing is attributed to assault or assistance.

Expirated blood can also resemble fine impact or gunshot spatter. Geoghegan et al. found bubbles or beading in only 20% of expirated patterns and concluded that stain shape alone could not distinguish expiration from gunshot or impact. On fabric swipe patterns, Yuen et al. found analysts identified direction correctly in only 3% of cases.

The physics defines a useful but limited scope. Connolly et al. found that careful selection and averaging across stains reduced random angle error, and described straight-line directional analysis as valid and reliable within that scope. Kabaliuk et al. found minimal drag effect for a passively falling drop within about 1.5 m, a common indoor distance. Error increases with distance, height and velocity.

Classic stringing can therefore support broad regions and coarse distinctions rather than a precise point. It tends to overestimate height at greater distances. Stain size also cannot establish impact velocity because the former low-, medium- and high-velocity size categories overlapped across mechanisms.

Experience and awareness do not remove contextual effects. Edmond et al. call the downstream problem the biasing snowball: pathology information influences a bloodstain conclusion, which is then presented as independent corroboration of the pathology. The same information is effectively counted twice.

Osborne and Taylor trialled blind independent review in a laboratory. The second analyst worked without case information or the first conclusion. The two agreed in nine of ten cases. In the tenth, review identified a decision the first analyst accepted had been driven by case circumstances rather than the pattern. The procedure was inexpensive and well received. In court, the relevant evidence is whether blind review occurred, not an assurance that context was put aside.

Meijrink et al. demonstrated a likelihood-ratio approach in a case comparing beating with expirated blood during CPR. Their overall LR was about 1,710 in favour of beating. Sensitivity analysis produced values from about 120 to 9,405 depending on judgement-based probabilities without calibration data. The authors warned that the point value implied unwarranted exactness and should indicate only an order of magnitude.

The OSAC methodology standard says opinions must not use unsupported data. It gives examples: inferring that an attacker was within two metres, a minimum volume of blood, or a time since bleeding from clotting. When stains are ambiguous, insufficient, altered or overlapping, the standard permits a conclusion that the pattern cannot be classified.

Morgan's 2023 registry study identified 33 wrongful convictions involving bloodstain examinations, with case error in nineteen. Thirteen involved testimony error, ten involved an officer of the court and nine involved classification error. Morgan also found that BPA cases reflected the contextual-bias concerns observed experimentally. He describes one conclusion by Dr Henry Lee as speculative and overstating the probative value of the findings.

Admissibility does not itself establish validity. One study found prosecution experts survived challenge about 92% of the time and defence experts 33%. The 2023 amendment to Federal Rule of Evidence 702 directs courts to examine reliable application and overstatement. The recurring lesson is to evaluate alternative mechanisms, state uncertainty and avoid making a categorical event conclusion from a small number of ambiguous stains.