Posts by Stephen Judd
Last ←Newer Page 1 2 3 4 5 Older→ First
-
I sleep well when I get exercise during the day and don’t drink too much.
I've also come to realise that I have a "window" between about 9:30 and 10:15 where I find it easy to get to sleep, then I start to become alert again. The next window is some time after 1 am.
Almost two years ago I discovered to my horror that if I have enough stuff to think about – stress – I can actually get into a state so agitated that I stay awake all night. I had always thought that if I just calmly waited, I would fall asleep, and discovering this isn’t so led to a worry that sits behind the other worries – what if I can’t get to sleep at all?
Luckily, I discovered that this is what zopiclone is for. On the rare occasions since where I have got into a stress tangle and just cannot stop thinking, it does the job.
-
Hard News: Cannabis: The Experiment is Real, in reply to Russell Brown,
I think our senior pols in all parties are convinced this is a vote loser. None of them want to lead on this issue.
As much as anything else, the press and the kind of stories we see in the press about the impact of drug prohibition are to blame. It is fear of being labelled loony and lambasted as irresponsible druggie-lovers who hate our children that prevents our biggest parties from leading.
-
Hard News: Narcissists and bullies, in reply to Ana Simkiss,
interested in the PAS view, particularly Labour Party members:
I don't speak for anyone, and I have no special knowledge, but I think Tamihere's political future in any party died this week, or it should have.
Pagani seems to imply that there are no working class, Maori men who aren't rape apologists, and that lack of empathy with your daughters and sisters is something you get with working class cred as a package deal. I think that's problematic.
-
Hard News: Cannabis: The Experiment is Real, in reply to Chris Waugh,
Probably much of Labour frothing at the mouth
The question came up at the Labour leadership forum in Christchurch. Jones was a very firm no, not at all, and Robertson and Cunliffe were slightly more open but said not a priority, too much of a policy distraction when there are more important things to do.
-
Hard News: Cannabis: The Experiment is Real, in reply to Jeremy Andrew,
you risk all the temptations of increasing income at the expense of the vulnerable, either deliberately, or by omission.
But at the moment, the vulnerable are the best market. At an appropriate price point and with appropriate regulation, solid middle-aged punters like me become appealing customers and importantly, we are actually in the market. Right now I have much more to lose trying to hook up with a reputable supplier than a spotty youth with no career and no responsibilities.
-
Speaker: Gender quotas (and helping…, in reply to Joe Wylie,
I dunno, tricky with his head already up his arse.
-
At times like this I wish PAS had a +1 or favourite button, because I would mash it on Andrew's comment so hard.
-
Speaker: Gender quotas (and helping…, in reply to Graeme Edgeler,
Organisations in the political domain?
-
Speaker: Gender quotas (and helping…, in reply to flash,
so to counter this you introduce a policy to select on merit only AFTER you have been selected on the basis of your gender?
My remark about how organisations that claim to select purely on merit don't shouldn't be taken as mere snark. It is very hard to do in this context. Women succeeded in modern symphony orchestras once audition set up was changed to make it impossible for the judges to see the player. I don't see any tests for selecting candidates for a party that could be applied objectively. I mean it would be great to select purely on merit, whatever the hell that is, but attempts to do that leave us stalled at less than parity.
My model has the assumption that an organisation intends to select on merit, but isn't good at it, on account of structural sexism.
Another thing I should say is that I can certainly imagine that if our bar for meritoriousness is very very high, so that even in a big pool of people only a very few meet the bar, that sometimes there will be more men in the top n selections than women. But in a fair environment, that should just as often be the other way around. And it never is. Which strongly suggests to me that maleness is an unconscious part of how we judge "merit". A quota isn't as good as pure merit, but evidence is that we can't do pure merit, and a quota beats sexism.
-
Speaker: Gender quotas (and helping…, in reply to Graeme Edgeler,
Another argument I considered including, but left out for length, is this: please show me an organisation that claims to select on merit and actually does. I have yet to see one. Sexism is systemic and structural. So to suggest "just select on merit" elides the very real difficulties in achieving an objective measure of merit that is unaffected by sexism.
My thought experiment is intended only to show that a quota, all other things being equal, is better than an existing sexist selection. There will be cases where that isn't so, but over time, and with any reasonable sized set of numbers, quotas are better than sexism, and this illustrates the principle nicely.
I actually wrote a little simulation to exercise this, and very consistently, while a "pure merit" scenario is best, an even quota is close to it and better than a sexist selection -- see http://pastebin.com/CfFnMjDT if you have a Python installation to play with.
If you have 100 men, and 100 women, with an average merit score of 100 and a standard deviation of 15 (which I chose because that's the classic IQ model), and you do a bunch of runs, you end up with something like the below. "pure" is the aggregate score of the top 10 people; "equal" is the aggregate of the top 5 men and top 5 women; "sexist" is the top 7 men and the top 3 women.
Averages - pure:1307, equal:1304, sexist:1298
Averages - pure:1307, equal:1303, sexist:1298
Averages - pure:1307, equal:1304, sexist:1297
Averages - pure:1305, equal:1302, sexist:1296
Averages - pure:1305, equal:1301, sexist:1296
Averages - pure:1305, equal:1302, sexist:1296
Averages - pure:1305, equal:1302, sexist:1296
Averages - pure:1306, equal:1302, sexist:1297
Averages - pure:1305, equal:1302, sexist:1296
Averages - pure:1306, equal:1303, sexist:1297