We all love to watch those CSI programs where seemingly no evidence is tested scientifically and results in a conviction of someone no one would otherwise have suspected. The real world is different and duller. Some forms of scientific evidence may be part of the prosecutor's case but not all are reliable and most can be rebutted.
Here you'll find a short summary of the kinds of scientific evidence commonly used in a prosecutions of violent crimes and the reliability of each.
Blood Spatter Analysis
This form of analysis (yes spatter, not splatter) is really more an art than a science. It examines the pattern blood drops make after a traumatic injury on nearby objects and forms hypotheses based on the form of the spatter about what happened. This analysis is only as good as the person who makes it. Plus blood spatter patterns on a wall can be interpreted differently. In short, this is not really a science.
This form of analysis (actually called Firearms and Tool Marking Analysis) studies the markings left on a bullet after it is fired and in some cases can connect a bullet to a specific gun. However, the older the gun is, the more often it has been fired, the less the chance that identifiable markings will be made. Scientific but generally not conclusive.
Gunshot Residue Analysis
When a bullet is fired explosive material besides the bullet exit the gun. In a close range shooting, this material will leave burn marks on the skin and clothing of the victim. This marking is called stipling. If a gun is placed against someone's head and fired there will be no stippling at all. Similarly a gun fired from 10 feet away will not leave stipling on the victim although any gunshot may leave residue on the shooter. This form of analysis is questioned more and more by criminalists.
A match is determined by the number of points in common between the fingerprints of a suspect and some crime scene object. Different technicians require different numbers of common points for a match. The more common points the greater the possibility that the match is good. Results are reliable depending on a high number of match points.
This is currently the hottest technique in forensics but it is not fool proof. False matches are possible among identical twins. How the DNA was taken and stored may affect its usefulness. A full analysis of DNA is beyond the scope of this article. In many states, anyone convicted of a felony must give a DNA sample. From this data pool, cold hits can be made in crimes where all other leads have long since disappeared.