Background: Sandra Guerra Thompson, a Professor of Law and Criminal Justice and Institute Director at the University of Houston Law Center, wrote Beyond a Reasonable Doubt? Reconsidering Uncorroborated Eyewitness Identification Testimony, in 2008 in the U. C. Davis Law Review.
This article reviews the overwhelming scientific evidence that establishes that eyewitnesses are notoriously inaccurate in identifying strangers, especially under the conditions that exist in many serious offenses such as robbery. Many of the factors that tend to decrease the accuracy of an identification are intrinsic to a witness’ abilities, and not the product of inappropriate suggestion by the police.
We know, for example, that eyewitnesses identify a known wrong person (a "filler" or "foil") in approximately 20% of all real criminal lineups.
Using the data that 20% of eyewitnesses identify a known wrong person, we wish to create a probability distribution and answer some probability questions.
Suppose we have three randomly selected lineups. Let I represent an incorrect identification.
Note: The complement of I is that an incorrect identification hat not occurred. This means that in any randomly selected lineup either the correct person or nobody was identified. In either case nothing incorrect has happened. So, for simplicity, let C represent the complement of I.
a. What is the sample space for this experiment?
b. Find the probability for each element in the sample space using the sample space you found above.
c. Let X represent a discrete random variable that counts the number of incorrect identifications. Create the probability distribution using the probabilities found above.