torture getting reliable information out of people

knocking someone out to harmlessly incapacitate them for like an hour

jumping into water from staggering heights and surviving the fall completely intact

calling the police to deescalate a situation

rafting your way off a desert island

correctly profiling total strangers based on vibes

effectively operating every computer by typing and nothing else

ripping an IV out of your arm without consequences

"as someone doing a fake thing, I'm here to tell you other stuff is fake too"

Single source DNA is real.

Toxicology and drug testing are real.

Probabilistic genotyping accurately suggests sources are present but we have no way of knowing if the likelihood ratios are correct because the software has not been (and cannot be fully) independently verified and validated.

Firearm toolmark examiners disagree with themselves 60% of the time when they view the same casings or bullets a second time.

Fingerprints are unique but fingerprint analysis does not accurately identify fingerprints to the level suggested by analysts, being wrong roughly 1 in 20 times.

Facial recognition has huge racial biases and is not accurate.

Polygraphs, hair analysis, bite marks, and profiling are all comedically discredited. For some ungodly reason the odontology people are trying to bring bitemark back.

Bloodspatter evidence vastly overstates the amount of accurate information it can provide.

I was just at a conference where someone was developing the technology to "identify" people using nail beds and knuckle prints. When we asked them to show us the studies that indicate there is any level of uniqueness/semi-uniqueness to fingernail beds or knuckle prints they could not do so.

torture getting reliable information out of people

knocking someone out to harmlessly incapacitate them for like an hour

jumping into water from staggering heights and surviving the fall completely intact

calling the police to deescalate a situation

rafting your way off a desert island

correctly profiling total strangers based on vibes

effectively operating every computer by typing and nothing else

ripping an IV out of your arm without consequences

"as someone doing a fake thing, I'm here to tell you other stuff is fake too"

Single source DNA is real.

Toxicology and drug testing are real.

Probabilistic genotyping accurately suggests sources are present but we have no way of knowing if the likelihood ratios are correct because the software has not been (and cannot be fully) independently verified and validated.

Firearm toolmark examiners disagree with themselves 60% of the time when they view the same casings or bullets a second time.

Fingerprints are unique but fingerprint analysis does not accurately identify fingerprints to the level suggested by analysts, being wrong roughly 1 in 20 times.

Facial recognition has huge racial biases and is not accurate.

Polygraphs, hair analysis, bite marks, and profiling are all comedically discredited. For some ungodly reason the odontology people are trying to bring bitemark back.

Bloodspatter evidence vastly overstates the amount of accurate information it can provide.

I was just at a conference where someone was developing the technology to "identify" people using nail beds and knuckle prints. When we asked them to show us the studies that indicate there is any level of uniqueness/semi-uniqueness to fingernail beds or knuckle prints they could not do so.

torture getting reliable information out of people

knocking someone out to harmlessly incapacitate them for like an hour

jumping into water from staggering heights and surviving the fall completely intact

calling the police to deescalate a situation

rafting your way off a desert island

correctly profiling total strangers based on vibes

effectively operating every computer by typing and nothing else

ripping an IV out of your arm without consequences

"as someone doing a fake thing, I'm here to tell you other stuff is fake too"

Single source DNA is real.

Toxicology and drug testing are real.

Probabilistic genotyping accurately suggests sources are present but we have no way of knowing if the likelihood ratios are correct because the software has not been (and cannot be fully) independently verified and validated.

Firearm toolmark examiners disagree with themselves 60% of the time when they view the same casings or bullets a second time.

Fingerprints are unique but fingerprint analysis does not accurately identify fingerprints to the level suggested by analysts, being wrong roughly 1 in 20 times.

Facial recognition has huge racial biases and is not accurate.

Polygraphs, hair analysis, bite marks, and profiling are all comedically discredited. For some ungodly reason the odontology people are trying to bring bitemark back.

Bloodspatter evidence vastly overstates the amount of accurate information it can provide.

I was just at a conference where someone was developing the technology to "identify" people using nail beds and knuckle prints. When we asked them to show us the studies that indicate there is any level of uniqueness/semi-uniqueness to fingernail beds or knuckle prints they could not do so.

the problem with forensics (even the real parts!) is that tv shows just make up a bunch of shit and have no idea what’s real and what’s pseudoscience and then jurors/prosecutors/general society roll with it because people assume tv is real life. i even saw OSU promoting their new forensic anthropology program by saying something like “have you ever wanted to be in criminal minds? bones?”. OSU!!! forensics programs should be promptly teaching undergrads that their career will NOT be like CSI, because that isn’t how real life works. mine did.