Why do judges keep letting bad science into criminal trials?

Relying on past decisions made by other courts – called "precedents" – is a handy shortcut for the UK legal systems and similar systems in the USA, Canada, Australia and New Zealand. It saves running the same arguments over and over again, it helps keep court decisions consistent over time, it reduces trial running-times, and helps lawyers advise their clients about how their case might be decided by a court. Precedent means that cases already decided by higher courts 'bind' any subsequent cases heard in lower courts, or courts at the same level. So a decision of the Supreme Court must be adhered to in all similar subsequent cases in that system, unless the decision has been modified by parliament, or the lower court can successfully 'distinguish' its current case from the decision in the precedent.

Unfortunately, this means that where a court allows poor quality evidence into a trial, such evidence is likely to be used again and again in subsequent cases. There is a further way to 'game the system' of precedent though, and some prosecutors use this for emerging or controversial forensic science techniques that are very convincing for juries but might have unproven or highly contentious scientific credentials. Here the prosecution is able to wait for a case where the defence is poor (underprepared, under-resourced, or incompetent) and will present the sub-standard forensic science evidence to the court in that case. Because the defence is underprepared, under-resourced or incompetent, the poor quality forensic science is not challenged; it is accepted by the court as evidence against the accused.