Posts by David Hood
Last ←Newer Page 1 2 3 4 5 Older→ First
-
So how has the NDP claim the middle been going? http://www.bbc.com/news/world-us-canada-34431009
-
Google's response to Facebook and Apple's in-gathering of news http://www.niemanlab.org/2015/10/get-ampd-heres-what-publishers-need-to-know-about-googles-new-plan-to-speed-up-your-website/
-
Without TPM restrictions, it is much easier to tell if manufacture software is lying to you (like VW)
-
Legal Beagle: The Greg King Memorial…, in reply to mpledger,
Unfortunately, that’s not really falsifiable.
It should show up in reported crime, if nothing else (unless you are positing the crimes are so crafty no one notices them, which is not likely in the case of strike offences).
-
Key and Peele's commentary on Chris Brown/ Rihanna
-
And "Life Imprisonment" for that cluster of offences. The unusual peak in around 2010 just shows how careful you are going to need to be about date ranges, as it was about double the 2005 numbers and that is quite a lot of variation.
-
More graphs. For the cluster of offences I called “violentish” above (because within them lie the really serious ones). We know from above that the reporting rate looks flat with cycles above. Here is the conviction raw numbers for those (black line graph, rising cyclical). The imprisonment raw numbers for those (rising, not as cyclical) and the proportion imprisoned of those convicted (blue steady cyclical, but we haven't talked proportions anywhere else, it really just shows the raw number peaks in convictions don't relate to imprisonments).
So flat cyclical reports, rising cyclical convictions, rising imprisonment. Take home point: Comparing times is complicated.
-
Note: I realise I only used two years, but recidivism does seem to be a decline path that is getting pretty flat by the end of year two.
Now allowing for 2004-2009 being a natural peak, and 2009-2014 a natural trough, we should probably try and estimate if the drop is that unusual if in is not a long term average vs intervention, but opposite ends of a cycle. So I would be inclined to fit the recorded cycle to the data before making a comparison (but that we need to wait until afternoon tea-time)
-
Yay in general for evidence based discussions.
That said- I'm not convinced you can do a direct past/present comparison as we know the crime rate changes over time, I think the comparison needs to account for that.
Also, If I take the recorded crime stats from http://www.stats.govt.nz/tools_and_services/nzdotstat/tables-by-subject/new-zealand-recorded-crime-tables/ASOC-offence-calendar-year-statistics.aspx
and combine the recorded categories of " Acts intended to cause injury", " Abduction, harassment and other related offences against a person", " Sexual assault and related offences", Offence == " Homicide and related offences" the proportion of absolute number of reported cases compared two years previous looks like the attached graph. Which shows that drops over a period are going to depend very much on your starting period. OTOH crime per head of population has been dropping dramatically since the late nineties so you also need to correct per capita discussions that creep in on the basis of that long term trend. -
Cherry blossoms
