Lucy Letby case - Evidence discussion thread

[IT Minion]

A new thread to allow discussing the evidence and the bugger picture of how statistics can be problems in complicated cases.

There's a critical review here: https://triedbystats.com/

 

[noseall]

Seems as though prosecutors and their expert witnesses have painted the target around the arrow, on this one.

 

[IT Minion]

Texas Sharpshooter fallacy.

The analysis I linked to was persuasive, but the analysis of the staff schedule used noise as the comparison to generate the mock shift register.

That's interesting as an example but it's deeply flawed itself. Shift patterns aren't random, you can't do four 13 hour shifts back to back, but if you use random noise for a year I'd expect that to occur multiple times. Random noise probably understates the chance of her being on shift, or having a preceding shift.

 

[ajohn]

Statistics can often be arranged to suite what someone wants. I'd be a bit dubious about it's use to find a murderer either way. It needs evidence. An expert saying that such and such could cause a death isn't really good enough on it's own. The ideal solution is a pathologist who can say such and such resulted in death.

Nurses work rotational blocks of shifts with gaps. I assume this one is the same. Statistics might show a difference from the norm when she is around. The norm though Babies do die in ICU's. My daughter in law ( if they ever get married) had that experience not all that long ago. The hospital took it's time to let that sink in for obvious reasons. It finished up in a situation where a baby can't remain in the ICU for ever so support was ended. This baby's problem was obvious. The results expected.

 

[motorbiking]

On the one hand we have the potential challenge of finding experts to support the defence and contradict the prosecution expert.

On the other, we have the specific anomaly of the drugs found, the circumstantial evidence and the diary of “confession.”

There is even the possibility that another killer framed her.

 

[ajohn]

I am assuming all of the baby deaths were in an ICU? I haven't followed the case that much so may be wrong. An ICU is not the usual place for a baby to finish up. I didn't count the number in the hospital I visited to see the baby I mentioned but it wasn't a few and there may be more than one ICU.

 

[IT Minion]

On the one hand we have the potential challenge of finding experts to support the defence and contradict the prosecution expert.

On the other, we have the specific anomaly of the drugs found, the circumstantial evidence and the diary of “confession.”

There is even the possibility that another killer framed her.

The drugs detected were using a test that shouldn't be used for forensic purposes.

The diary included 'i didn't do it' as well as 'i did it', so it's a bit contradictory.

 

[Mrs Ten Names]

Seems as though prosecutors and their expert witnesses have painted the target around the arrow, on this one.

Never heard that expression before.

 

[Trazor]

There is even the possibility that another killer framed her.

With her confessions, how would that work.

 

[Brigadier]

Never heard that expression before.

As already posted, The Texas Sharpshooter Fallacy.

 

[Notch7]

David Davis opinion:

 

[ajohn]

The guardian has pulled it all together, Afraid there are a lot of points.

Lucy Letby: killer or coincidence? Why some experts question the evidence

Exclusive: Doubts raised over safety of convictions of nurse found guilty of murdering babies

www.theguardian.com

 

[Mrs Ten Names]

David Davis opinion:

He has a track record of fighting potential miscarriages of justice, with some success.

 

[IT Minion]

David Davis opinion:

He's an idiot, but stopped clocks are right twice a day.

 

[fillyboy]

The guardian has pulled it all together, Afraid there are a lot of points.

Lucy Letby: killer or coincidence? Why some experts question the evidence

Exclusive: Doubts raised over safety of convictions of nurse found guilty of murdering babies

www.theguardian.com

Good article and good points raised by David Davis above yours. Certainly beginning to look like it could be an unsafe conviction but that is going to be desperately difficult to prove, but surely just an element of doubt is sufficient to make it unsafe.