What is the Probability of a DNA Match in Crime Scene Forensics?

  • Context: Graduate 
  • Thread starter Thread starter Dooga Blackrazor
  • Start date Start date
  • Tags Tags
    Dna
Click For Summary

Discussion Overview

The discussion revolves around calculating the probability of a DNA match in the context of crime scene forensics, specifically using Bayes' theorem. Participants explore the implications of prior probabilities, true positive rates, and false positive rates in determining the likelihood of a suspect's DNA matching crime scene evidence.

Discussion Character

  • Homework-related
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • Some participants discuss the importance of prior probabilities, such as the likelihood of a suspect's DNA being found at the crime scene, which is suggested to be 0.10.
  • True positive and false positive rates are introduced, with values of 0.999 and 0.005 respectively, but participants express uncertainty about how to apply these in calculations.
  • There is a request for symbolic representation of the problem, indicating a need for clarity in translating the scenario into mathematical terms.
  • Some participants express confusion about the relationship between the given probabilities and how they should be applied in Bayes' theorem.
  • A participant mentions successfully answering similar questions on a quiz, indicating some level of understanding was achieved through the discussion.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best way to express the problem symbolically or how to apply Bayes' theorem effectively. There are multiple competing views on the interpretation of the probabilities and their application.

Contextual Notes

Some participants struggle with converting the word problem into an algebraic format, and there is uncertainty regarding the definitions and implications of true and false positive rates. The discussion reflects varying levels of familiarity with Bayesian reasoning.

Who May Find This Useful

Individuals interested in forensic science, probability theory, or those preparing for assessments involving statistical reasoning in legal contexts may find this discussion beneficial.

Dooga Blackrazor
Messages
258
Reaction score
0
What is the value of prob(conclusion C1 is true IF premise P1 is true)? It depends on the prior probability that the suspect's DNA would be found at the crime scene -- which makes it a true match for some crime scene DNA -- and the true and false positive rates for the police forsensics lab reports for DNA matches. Suppose the prior probability that the suspect's DNA would be found at the crime scene is 0.10.* Suppose the true positive rate for the police forensics lab reports for DNA matches is 0.999. Finally suppose that the false positive rate is 0.005.**

Enter answer as number in box below. Number should be rounded to two decimals (eg 1.00, 0.99, 0.50, 0.00)

*We are sure of ten suspects based on evidence other than the DNA match. There is no more evidence for one than the other. The crime is one where the suspect would inevitably leave something containing his or her DNA.
**We have information or facts from testing about the false positive rate for some labs. This includes the lab that did DNA matching in the O J Simpson murder case. These numbers are representative. We do not have information about the true positive rate. These numbers are guesses.

P1 IS: The forensic lab says "This suspect's DNA matches some of the crime scene DNA"

C1 IS: The suspect's DNA matches some of the crime seen DNA.

The answer is 0.96 or 96%. Normally I can work backwards to find out how the answer is found, but I am totally confused and lost on this one. Been going at it for an hour. This is from a practice quiz, btw, so you are only helping me prepare for a future quiz - not giving me assignment marks or anything.

Thanks
 
Physics news on Phys.org
Can you write down, symbolically, everything the problem tells you?
 
Hurkyl said:
Can you write down, symbolically, everything the problem tells you?

I wouldn't know how to write them down symbolically. As far as I can tell, I could write this:

True positive rate: 0.999
False positive rate: 0.005
Probability DNA would be found at crime scene: 0.10
# Suspects: 10

Seeking to find the probability DNA will match some of the DNA at the crime scene.

Now I assume this means that DNA was found out the crime seen and the 0.10 is no longer a variable, but I am lost really and unsure about that, too.
 
Wow. I am such a newb. I just read the "This Forum is not for Homework Questions." Technically, it isn't homework, but if it is more appropriate to move it please do so. By the way, I also have msn at dooga16@hotmail.com if anyone finds it easier to help me with it then. Basically, if someone could put the variables into the Baye's formula or solve it with all the work shown, I could easily figure out how to do identical problems. Anything, though, is helpful.
 
Dooga Blackrazor said:
I wouldn't know how to write them down symbolically.
Well, one is obvious

P(The crime scene DNA is the suspects DNA) = 1/10.

True positive rate: 0.999
False positive rate: 0.005
So, what are true and false positive rates?



Basically, if someone could put the variables into the Baye's formula or solve it with all the work shown, I could easily figure out how to do identical problems.
(Based upon what I've seen so far in this thread) I really don't think that's where your problem lies; it looks like your problem is in converting the word problem into an algebra problem... not with the solution of the algebra problem!
 
Is the 10% given in the question the same as the 1/10 figure you are coming up with or a different one? I am inclined to think different though I have been up awhile

True positive rate: 0.999 (gives a positive when it is positive)
False positive rate: 0.005 (gives a positive when it is negative)

Gives a positive when it is negative (0.001) False positive
Gives a negative when it is negative (0.995) True negative

Is that what you mean or something else? Or did you want it in percent? Like 99% and .5%?
 
Baye's Theorem Notes:

P(H1|E) = P(H1|E) x P(H1) divided by P(H1|E) x P(H1) + P(E|H2) x P(H2)

probability Match / Lab Match =

prob(lab match/probmatch) x probmatch

_________________ (division)

prob(lab match/probmatch) x probmatch + prob(labmatch/nomatch) x prob(nomatch)

Trying to put these into terms I can understand better:

The probability there is a match (in reality) and a match (according to test) is what I am solving for. I will call this X.

X = true positive rate x probability there is a match divided by itself + false positive rate x probability there is no match.

So, now I am trying to find out what the probability of there being a match or no match is. There, I am a bit lost. I will go with 1/10 and 9/10th to see how that works for me.

Well, that is troubling. I seem to have the answer. If I did it incorrectly and got the right answer, that would be helpful to know as well.

Thanks for the help.
 
Is the 10% given in the question the same as the 1/10 figure you are coming up with or a different one?
Same one.


Is that what you mean or something else?
Something else. For example, I want you to write down what it means for the false positive rate to be 0.005 in the form:

The probability of ________ is ____​

or maybe

The probability of _______ given that __________ is _____​

(or, preferably, the equivalent statement using P notation)
 
Well, I got it and was able to answer all the questions of that kind on a quiz. Thanks
 

Similar threads

Replies
3
Views
3K
  • · Replies 28 ·
Replies
28
Views
4K
Replies
1
Views
4K
  • · Replies 8 ·
Replies
8
Views
2K
Replies
2
Views
3K
  • · Replies 10 ·
Replies
10
Views
3K
  • · Replies 19 ·
Replies
19
Views
8K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 13 ·
Replies
13
Views
10K
  • · Replies 32 ·
2
Replies
32
Views
7K