Posts by linger
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Hard News: The Day After Tomorrow, in reply to Trevor Nicholls,
True that: and makes a mockery of the Commission's highly-prioritised concern that parties would need >5 MPs in order to function effectively in Parliament. (That concern was arguably misplaced by many miles: (i) it's not true if a larger party has agreed to provide logistic support; and (ii) it would not be true of any genuinely single-issue and therefore single-portfolio party, were such a thing to be made practical by removing the threshold.)
-
Speaker: The Government lost the election, in reply to Mr Mark,
That’s disturbing, and certainly supports a need for public education, though to be fair, “importance” is a value judgement with more than one dimension, and even in the intended narrow sense of “importance for determining makeup of Parliament” may depend on local circumstances. In 2014, voters in at least four electorates key to potentially including minor parties (Internet-Mana/Maori/UF/ACT) might legitimately have considered their electorate vote more important than their party vote in terms of its possible effect on makeup of Parliament.
-
Hard News: The Day After Tomorrow, in reply to Sacha,
I like that: essentially the same outcome as, but simpler mechanism than, a 1% threshold.
-
Hard News: The Day After Tomorrow, in reply to izogi,
Oh yes: the fact that what was tossed wasn't good enough doesn't make the tosser any less of a tosser.
-
Hard News: The Day After Tomorrow, in reply to izogi,
leader of exaggerated personality
Am I wrong for immediately thinking of the "rodents of unusual size" from The Princess Bride ?
-
tbh the Commission’s recommendation on thresholds (–> 4%) was too weak, and in combination with the recommendation to remove the so-called “coat-tailing” exception, would not have improved the proportionality of MMP at all — so it was a relief those got tossed, so we can try again.
As I’ve said before, the threshold should be set below a level equivalent to support that can win more than one electorate (which solves both problems at once). A plurality in an electorate can be reached with votes equivalent to 0.5% of the total, so a 1% threshold is sufficiently high. Sure, this could allow some strange minority views into Parliament; but giving everyone a chance of representation is really the point of a representatative democracy, isn’t it? (And it’s not like we haven’t already had anti-fluoridists, anti-vaccinationists, racists, sexists and bigots in Parliament under our systems to date.) -
For me there’s an almost total disjunct between conservatism (“we should keep doing what we’re doing now, because that’s working out really well so far for me and my mates personally”) and conservation (“we should preserve the environment against change that would otherwise threaten future choices and freedoms”). I don’t see any possible common ground between “conservative” lassez-faire capitalism and social or environmental sustainability. I do see considerable common ground between environmental sustainability and social sustainability, in terms of having a commitment to being pro-active rather than reactive, being future-oriented rather than past-oriented, seeking to help all others rather than just those like you, and setting quantifiable long-term goals, together with a realistic plan for achieving them, measured against shorter-term benchmarks. As long as National is characterised by short-term minimal managerialism, it’s not a suitable vehicle for Green objectives. So, “blue-greens” … well, I’m willing to believe they may exist, but on both a theoretical basis and a practical basis it’s not obvious to me how. Can you explain this position in more concrete terms?
-
Hard News: The Day After Tomorrow, in reply to Bart Janssen,
Several caveats are necessary when looking at the results of those correlational analyses. They’re interesting, but I’d regard them as indicative rather than conclusive.
(i) Based on correlation only, and data is electorate-population-level, not individual-level, so the chain between evidence and causal interpretation is very indirect. Strictly, the evidence behind a claim such asthe correlation between being aged 5-14 and voting Green increased sharply, from -0.42 in 2014 to -0.08 in 2017
is of the form “people who live in areas where there are larger numbers of young children than average were slightly less likely than average to vote Green; but they were even more strongly less likely to vote Green in 2014”. (Which perhaps means that people in such areas were less likely to abandon the Greens in 2017, and possibly indicates that parents of young children were less likely to abandon the Greens in 2017, though that is an extra interpretive leap.)
(ii) Pearson correlation coefficients are highly sensitive to outliers, and it only takes a few very oddly-behaving electorates (say, because one local candidate is personally popular and attracts more votes for their party than the national average) to create significant associations, or conversely, to mask otherwise consistent associations. (Changes in high-profile personnel may be an important confound when looking at changes in correlation between 2014 and 2017.)
(iii) 9000 correlations were analyzed. Sounds impressive, but I would actually regard it as a problem, as it becomes impossible to attach a reliable statistical significance level to any one result among this number. The analysis doesn’t seem to have corrected reported significance for the number of tests performed.
(iv) The results do not yet include the 15% of specials. If the special votes are distributed unevenly by electorate, or demographically, or by party preference, the correlation matrix may change considerably.
(v) Results for smaller parties will be especially uncertain, as the numbers of supporters per electorate are relatively small, and those supporters may not necessarily be typical of the electorate population. -
Hard News: The Day After Tomorrow, in reply to Katharine Moody,
Ah, I see: it isn’t based on correlations between individual voter characteristics and individual voting preference, but rather on correlations between electorate-wide population characteristics and electorate-wide voting preference. (And that means that for now the specials are not included, which may affect the conclusions.) Fair enough. I hope the analysis includes a Bonferroni correction or similar adjustment for the number of significance tests performed, though; otherwise there’s so many potential explanatory variables going in that there’s a risk of getting a large number of spuriously significant results.
-
Hard News: The Day After Tomorrow, in reply to simon g,
whatever they will call themselves
you kind of suggest it yourself: Astroturf Party (for fake greenery, and fake concern).