Talk:Ceorana Department of State Human Development Index
There is an error in the life expectancy index calculation - having a low life expectancy produces a lower (ie, better) score. Also, the article doesn't say whether the average of the three indices is taken to be the CHDI. In case my edits are controversial, I will put it here first. I suggest that the life expectancy index calculation look like this:
(85 - 25)/(life expectancy-25)
I'm playing around with the rest of the formulae... but I suspect that they might all need to be inverted too... --Pantocratoria 13:31, 30 January 2006 (GMT)
The next problem that I note with the life expectancy calculation is that it penalises a good GDP index and education index (both of which will be below 0). I suggest that the first multiplication be turned into a division. I'm not in any position to comment as to the accuracy of the number it produces, but I will say that it seems to make more sense than the previous way, because a good education or good GDP index now increase life expectancy, as opposed to decrease it.
The GDP index also seems to be a backwards division, as does the education index... so that makes all the indices. Note that I still don't really know about the numbers, I am just trying to get it so that it works in the proper way - doing it up in a spreadsheet and putting a few countries side by side exposes that numbers which individually seem reasonable don't work in a comparative system.
Sorry to bitch so much, it's a very helpful article, I'm just playing around with it now with the aim of making a calculator... --Pantocratoria 13:46, 30 January 2006 (GMT)
- Sorry to be unclear, I'll adjust the article, but in this index, as in the real HDI, a higher score is better, not a lower one. So your suggestions are valid, except that they'll yield a number on a different scale. → Ceo \ squawk 13:53, 30 January 2006 (GMT)
- Uhhh... :) I'm an idiot. For some reason I thought it worked the other way around... *adjusts his spreadsheet back to the original formulae* Hmm... yes, now that I have it the right way around, your formulae are much more satisfactory than my adjustments! Pardon my misunderstanding. I've uploaded an Excel spreadsheet with the formulae, and examples calculated using numbers from NSeconomy. I will however suggest that the GDP per capita should have a slightly higher impact on life expectancy, as well as some sort of impact on the education index calculation, which is at this moment entirely based on publicly funded education only. I suspect that such adjustments would mean that in the examples in my spreadsheet, Knootoss and Pacitalia in particular would perform a good deal better - right now their scores are fairly mediocre. Anyway, the spreadsheet is here. --Pantocratoria 14:13, 30 January 2006 (GMT)
- Do have a look at that spreadsheet, if you have Excel or OpenOffice, because I'm curious what I've done wrong so as to produce a poorer CHDI for Ceorana than the one in your article - if I am going to do a calculator, I want to get it right! --Pantocratoria 14:27, 30 January 2006 (GMT)
I'm going to adjust the formula so it is possible to use an RP'd literacy rate, or one from Sunset's calculator, [1]. That will make it more realistic, and account for vast differences in education spending that either cause some nations to have it way too high and others to have it way too low. I'm also going to throw in an IF statement to account for privatised education and healthcare. → Ceo \ squawk 02:24, 31 January 2006 (GMT)
Quite a bit better now, except now it actually overly advantages countries with healthy economies but no public education whatsoever, and doesn't model nations with a combination of public and private education. --Pantocratoria 07:51, 2 February 2006 (GMT)
Here is my latest spreadsheet. The formula is better now, although I still get a minor disparity between Knootoss' number on this page and the number in my spreadsheet, and a major disparity in the case of Pacitalia. Ceorana's rating has improved a couple of fractional points in my spreadsheet but I assume that's just the natural passing of time :). Have a look, play around, and tell me what you think we can do here... the problem seems to me to be that there is no real recognition of a combination of private and public education (Pacitalia's UNHDI would be much higher in my spreadsheet if it didn't have any education spending at all!), and that there is no recognition of private healthcare. I don't think substituting the GDP per capita index for these scores is an answer. I also think that in states where there is obviously private contribution towards healthcare and education, the GDP per capita index needs to be reduced slightly. Sorry I have nothing approach hard figures for any of these suggestions, but do have a look at the spreadsheet and have a play anyway. --Pantocratoria 06:16, 3 February 2006 (GMT)
- Yeah, that's odd. Putting 0 for education and healthcare gives me an HDI of 0.987 (using my RPed stats). (( Pacitalkia )) Time sent: 18:44, 3 February 2006 (GMT)
- And putting in my estimates for private-sector education and healthcare spending gives me an HDI of 1.03. So, yeah... I don't know. Then I actually used 61466.21 * population to get the GDP, entered the GDP in the spreadsheet and got an automatic figure of 0.989. It's really odd. But the spreadsheet works, so it's just funny that Pacitalia would be, in this respect, "perfect". (( Pacitalkia )) Time sent: 18:49, 3 February 2006 (GMT)
- Yeah, the spreadsheet works, in that it accurately reflects the formulae, but I think the formulae are still a little off, for health and education anyway. --Pantocratoria 05:22, 4 February 2006 (GMT)
Education formula
I’d use this formula, except that because of CH’s one percent education spending, the education index drops from .991 to .334, causing an overall change in the HDI from .962 (very high) to .597 (reasonable). I know government spending in areas like education is destructive, but it’s not that destructive. :P —Commerce Heights (talk) 23:02, 5 February 2006 (GMT)
