Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.
Display:
Dimension reduction - find a conveniently small subset of linear combinations of the mass of observed variables which captures a suitably large fraction of the total variance.

As hypothesis-testing goes, that strikes me as putting the cart before the horse. If they have a hypothesis for how the data should behave, then why are they not just fitting directly to their hypothesis? And if they don't have a hypothesis, then it sounds suspiciously like a fishing expedition. Which, like all fishing expeditions, runs a large risk of bycatch. Are they modeling signal, or are they modeling noise here?

Because linear algebra is the only thing we humans know how to work with...

And it is, after all, better to be very precisely wrong than to be roughly right...

- Jake

Friends come and go. Enemies accumulate.

by JakeS (JangoSierra 'at' gmail 'dot' com) on Mon Apr 20th, 2009 at 04:08:21 AM EST
[ Parent ]
JakeS:
As hypothesis-testing goes, that strikes me as putting the cart before the horse. If they have a hypothesis for how the data should behave, then why are they not just fitting directly to their hypothesis?
Because they don't have a hypothesis. Factor analysis should be seen as an exploratory data analysis technique.

If you then colour-code the points and find that the former Soviet Republics and the English-speaking countries form two distinct clusters in the sense that the within-group variance in the two-factor space is smaller than the between-group variance, then you have discovered something. Assuming that the two factors, together, explain a sufficiently high amount of the overall variance to begin with.

Most economists teach a theoretical framework that has been shown to be fundamentally useless. -- James K. Galbraith

by Migeru (migeru at eurotrib dot com) on Mon Apr 20th, 2009 at 04:24:23 AM EST
[ Parent ]
Nothing wrong with trying to reduce complex phenomena to more simply defined ones - or at least to try to explain complex phenomena with respect to a few variables and see how well those variables can explain observed data.  But you then have to understand that you are working with your model of reality, and not with reality itself.  And the model you use may well help to determine the results you achieve.  The observer (and his/her mindset/model) is part of the processes he observes.

There is no doubt that the "modernisation" model appeals to (even if it is not confined to) the "advanced" societal mindset which implicitly sees a progressive evolution from feudal/agricultural to urban/industrial and post industrial "advanced" "knowledge" based societies.  It's about understanding (and helping) underdeveloped societies to develop and become more like us.

Its a sucker argument if you want to pull in the research funding dollars...

Equally, it can be argued, that even Marxist class based analyses also contain deterministic evolutionary elements (feudal, capitalistic, socialistic) or that a religious model - which defines everything in terms of a theological understanding of God/Allah's will - is also as much about shaping how societies can develop as it is about describing how they have/can develop.  Your research dollars come from different sources if these are the models you want to apply.

The "success" of the western modernisation model - as defined by its post WW2 - economic/political/technological/military success - has radically shaped hearts and minds - and it is not surprising that many people, in both the "developed" and "undeveloped" world subscribe to it - even if only unconsciously.  But is that a measure of the theory's innate explanatory power, or of the political/economic forces that have made it the dominant mode of popular (and elite) consciousness?

notes from no w here

by Frank Schnittger (mail Frankschnittger at hot male dotty communists) on Mon Apr 20th, 2009 at 07:20:12 AM EST
[ Parent ]
Frank Schnittger:
But is that a measure of the theory's innate explanatory power, or of the political/economic forces that have made it the dominant mode of popular (and elite) consciousness?

great question.

i think it's a case of overshoot, too much of a good thing, what starts off healthy move then morphs into a sick one.

for example,  'economic development' is a no-brainer Good Thing for the very poor, and at the other end of the scale, behold the corruption and headgames of the US financial usury machine/Ponzi scams industry.

(of course madison avenue and sheer novelty are the hook, the barb comes later, when the food aid displaces local businesses, the GM seeds supplant the traditional, IMF loans cripple the economy with debt etc etc...)

 

'The history of public debt is full of irony. It rarely follows our ideas of order and justice.' Thomas Piketty

by melo (melometa4(at)gmail.com) on Mon Apr 20th, 2009 at 08:55:34 AM EST
[ Parent ]
Frank Schnittger:
Nothing wrong with trying to reduce complex phenomena to more simply defined ones - or at least to try to explain complex phenomena with respect to a few variables and see how well those variables can explain observed data.  But you then have to understand that you are working with your model of reality, and not with reality itself.  And the model you use may well help to determine the results you achieve.  The observer (and his/her mindset/model) is part of the processes he observes.

There is no doubt that the "modernisation" model appeals to (even if it is not confined to) the "advanced" societal mindset which implicitly sees a progressive evolution from feudal/agricultural to urban/industrial and post industrial "advanced" "knowledge" based societies.

But the validity of factor analysis for "dimension reduction" is independent of the "modernisation" model or even the "modernisation agenda" you summarise as
It's about understanding (and helping) underdeveloped societies to develop and become more like us.


Most economists teach a theoretical framework that has been shown to be fundamentally useless. -- James K. Galbraith
by Migeru (migeru at eurotrib dot com) on Mon Apr 20th, 2009 at 09:24:05 AM EST
[ Parent ]
I'm not querying the math - see first para. -  but saying that how/when/where it is applied may be a political choice or based on available funding.  Neither am I dismissing the modernisation theory entirely - it is a model that can be applied - just saying that it is not the only, or necessarilly the "best" theory, but that it is very congruent with the dominant western conceptions of development and underdevelopment.

notes from no w here
by Frank Schnittger (mail Frankschnittger at hot male dotty communists) on Mon Apr 20th, 2009 at 11:48:04 AM EST
[ Parent ]
I can buy that. But then you'd need a new data set to test your hypothesis on. Preferably a reasonably independent data set... Presenting a hypothesis based on a data set, and then using the same data set to make predictions... that strikes me as falling dangerously close to a fishing expedition.

But of course it's possible that I'm applying the standards of physics (where you can almost always just take another data series) to a discipline to which it is not applicable (because we only have one data set - the Earth).

- Jake

Friends come and go. Enemies accumulate.

by JakeS (JangoSierra 'at' gmail 'dot' com) on Mon Apr 20th, 2009 at 06:40:23 PM EST
[ Parent ]

Display:

Occasional Series