[sci.stat.consult added, in case anyone still reads ssc]
On Wed, 11 Aug 2021 13:00:28 -0400, Quinn C
<
lispa...@crommatograph.info> wrote:
>* Rich Ulrich:
>
>> On Wed, 11 Aug 2021 06:25:25 +0100, Richard Heathfield
>> <
r...@cpax.org.uk> wrote:
>>
>>>On 11/08/2021 02:44, Peter Moylan wrote:
>>>> On 11/08/21 12:25, Peter T. Daniels wrote:
>>>>> On Tuesday, August 10, 2021 at 6:50:46 PM UTC-4, Richard Heathfield
>>>>> wrote:
>>>>>
>>>>>> "Transgender prisoners are five times more likely to carry out sex
>>>>>> attacks on inmates at women’s jails than other prisoners are, official
>>>>>> figures show.
>>>>>>
>>>>>> Male prisoners who were transferred to women’s jails during gender
>>>>>> reassignment and women inmates who are transitioning committed seven of
>>>>>> the 124 sex attacks recorded between 2010 and 2018. They occurred at HMP
>>>>>> Low Newton in Co Durham, Foston Hall in Derbyshire, Peterborough,
>>>>>> Bronzefield in Middlesex and New Hall, West Yorkshire."
>>>>>
>>>>> Let's see. Seven versus 117. Pretty heinous!
>>>>>
>>>>>>
https://www.thetimes.co.uk/article/seven-sex-attacks-in-womens-jails-by-transgender-convicts-cx9m8zqpg
>>>>>>
>>>>
>>>> To evaluate that, you'd have to know what proportion of the prison
>>>> population is transgender.
>>>
>>>Numeracy might be an ask too far.
>>
>> Being a statistician and data analyst, I feel compelled to point
>> out that the events are not "independent" and that "Mr White"
>> in a previous post accounted for 5 assaults. Proper statistical
>> testing of the most common sort assumes independence of
>> events.
>>
>> So, yeah, a few people need to be watched with suspicion.
>> And that suggests to me the possibility that fair and accurate
>> "counting" may be hard to achieve, if some people are watched
>> more closely.
>
>Wow. I thought there might be more to this if one were to look closely,
>but I didn't feel I have the time and energy, so thanks.
>
>This makes it almost a textbook case of manipulating the message by
>choosing the way of presenting statistics.
I suppose it could have been intentional misrepresentation, but
I still figure Hanlon's Razor applies - "never attribute to malice
that which is adequately explained by stupidity".
--
Rich Ulrich