People v. Collins

Supreme Court of California · 1968 · Evidence

68 Cal. 2d 319 (1968)

Updated August 22, 2026

Evidencemathematical probabilityexpert testimonyfoundationstatistical independenceproduct ruleidentity evidencecircumstantial evidence

Facts

The prosecution's identification case was weak and circumstantial: the victim could not identify Janet and had never seen Malcolm, while another witness identified Malcolm as the driver of a yellow car and described the fleeing woman as a blonde Caucasian with a ponytail and dark clothing. To bolster identity, the prosecutor called a mathematics instructor and, without statistical evidence supporting the numbers used, assigned probability factors to characteristics such as a partly yellow car, a mustache, a beard, a blonde ponytail, and an interracial couple. Using the product rule, the prosecutor argued there was only one chance in 12 million that any other couple shared those characteristics, and then told the jury the real odds were more like one in a billion. The trial court admitted the testimony over objection and denied a motion to strike it.

Issue

Whether the prosecution properly introduced and used mathematical probability evidence to identify the defendants as the robbers. More specifically, whether such probability testimony was admissible when based on unsupported numerical assumptions and no showing that the multiplied factors were statistically independent.

Rule

Mathematical probability evidence in a criminal case is erroneous and prejudicial when the prosecution fails to establish an adequate evidentiary foundation for the underlying probability factors or fails to prove that the factors are mutually statistically independent as required for use of the product rule. Even apart from those defects, probability calculations about how rare a combination of traits may be cannot by themselves guide the jury on the crucial question whether this defendant committed the crime and may not be used to displace the jury's duty to decide guilt beyond a reasonable doubt.

See the holding & full analysis

Create a free KwikCourt account to unlock the rest of this brief — and practice the case.

The court's holding and reasoning
Doctrine tests, pitfalls & exam hypotheticals
10 practice questions + 4 AI-graded essays on this case

Free sample · practice this case

Test yourself

One of 10 multiple-choice questions for this case. Pick an answer to see why.

In a purse-snatching trial in Phoenix, the prosecution's identification evidence is shaky. To strengthen it, the prosecutor calls a college statistics lecturer, supplies his own estimates that 1 in 8 getaway cars are green, 1 in 5 male drivers have goatees, and 1 in 12 female passengers have red braids, and asks the lecturer to multiply those figures to show how unlikely it is that anyone other than Devin Ortiz and Marla Kent committed the crime.

Should the trial court admit the probability testimony?

Explanation. The testimony should be excluded. The majority held that probability evidence identifying a defendant is erroneous when the prosecution offers no evidentiary support for the individual probability factors and instead has counsel supply the numbers. The problem is not arithmetic but the absence of valid foundational data. The case does not create a blanket ban on mathematics in criminal trials.