Anything That Interests Me! :)





Thursday, November 13, 2008

Anything that Interests Me! - How to learn Econometrics 2

Anything that Interests Me! - How to learn Econometrics 2

My friend DF replied to my earlier post on How to Learn Econometrics, and he came up with very good arguments and analysis on how I did not do a good job. Well, as that was my first analysis on how to learn Econometrics I must admit that I was sloppy with my language and pandering to my readers somewhat. I will now take a more formal and more academic tone, and I shall try to give some of the dry wit and humour that accompanied his thinly veiled rebuttals, or more accurately, criticisms.

How to learn Econometrics, part 2

His article has two main premises -

1. I do not separate between performance and actual learning.

2. My list of the relevant fields in econometrics is wrong.

These two premises lead him to the conclusion that I morph into a statistic geek, and this leads one to the inexorable conclusion that the only way to be good at Econometrics is to basically be good at statistics, and in particular, for higher levels, matrix representation.

There are basically three counter criticisms that I will make, and if he is reasonable, he will no doubt agree with what I say. However, if he insists upon having his erroneous views, it has more to do with the fact that we are both intelligent people and differ on the approaches to learning. Have no doubt, dear reader, that I am challenging someone of a high intellectual standing.

1. I don't separate between performance and actual learning. This is true, and I acknowledge that what he says is true. However, is there a real difference between performance and actual learning?

Allow me to state categorically that "how to learn Econometrics" means different things to different people. My opponent is assuming that I mean learn in his own way, however, learning for the purposes of my career and education, and for the purposes for many other readers who want to learn econometrics, has nothing to do with understanding Euclidean spaces, or even mathematical proofs, or the like. It has to do with performance.

Why then is it called learning? I chose the word learning because it appears higher in the search engines. Having made that callous joking remark, I now go on to state:

the only way in which you can tell if someone has learnt something is by performance, even though it does not follow that someone who performs has learnt something.

That is, in other words, the behaviour is more important than the intention, because we can see performance and then infer that learning has taken place, although this might not be true at all. The key is that my opponent believes in real learning, whatever that might be - and I am not prepared to venture as to what he means by that. Can one read another's mind? As a mentalist I can say the answer is no, but we can infer that another person has a mind by the remarks that the person makes. This is a complex philosophical issue, so I will just leave it as that - performance matters more, so that's why I wrote that.

2. The list that I offer is wrong - and this I can categorically say, that remains to be seen. I can in fact suggest that all he says is similar to mine:

"probability, statistics, matrices, linear regression and other regression models"

"philosophy of statistical inference
validity and accuracy of statistical inference procedures, in this case linear regression
limits and tradeoffs of statistical inference, of each of the assumptions, in a few different situations
limits and tradeoffs of cost of information, e.g. sample size issues, experimental control issues
matrix representation of linear models, datasets, and procedures"

Being of an argumentative bent, and also since I could have very well been a lawyer, what he says is not actually excluded from the general terms that I use. Yes, a technical distinction - but is not philosophy of statistical inference inferred from "statistics", and so on? Matrix representation seems plausible for linear models, and in fact it is. I have used general terms; he hasn't.

In fact, I will go on to argue that he is wrong that my list has no common purpose. They are all part of econometrics. From the outset, econometrics is not a single subject where the field is clearly defined. There are no single definitions for this, and the history is quite complex. Econometrics is but a tool in the hands of economists; how you define the tool is another matter.

3. The greatest joke of all is that I am a statistics geek. In particular, an econometrics geek, where our statistics is different from others (e.g. Psychologists and their ANOVA). He is totally wrong. The problem is that I am unable to convert from the type of statistics that we have trained for into the matrix representation, whereas mathematicians and statisticians have an easier time.

Having said that, the post was not to strategise, although that was indeed a possibility. The post had its ultimate purpose in SEO and other fiduciary considerations that I shall not delve into for fear of ruffling feathers.

How to learn Econometrics / how to do well in EC3304, part 2
All these are learnt from my detractor who posted those comments...
1. Complete proofs.

2. Learn the individual parts and then bring them all together.

3. Practise the problem sets.

4. Understanding truly, and real learning is better than getting the form right - i.e. don't do my performance method; use his real learning method.

5. Know your lecturer. In this case, he got this one right.

The fact of the matter remains that mathematics has taken hold of the social sciences and will not let go, and that is something that we have to deal with. At the same time, I will not hesitate to point out that my other humanities and social sciences skills far surpass my opponents (and this particular opponent in particular).

In terms of studying, academics: I am an Economist.

I consider myself a historian as well, and in fact, an educated person in general.

And as a former arts person, and a student from the Faculty of Arts and Social Sciences

LET ME REPEAT, ARTS AND SOCIAL SCIENCES

I feel a certain academic distaste when people outside my discipline attempt to reform it in ways that they see fit.

Economics always has the human factor. As Akerloff said, I believe and paraphrase: if in real life people are irrational, stupid and silly, then why do we stick to our assumptions that they are rational beings?

Because it makes the math easier. I say. Because we are slaves to the math as psychological and sociological answers are not mathematical enough, especially to my detractor DF.

I feel that it is far more productive in arts terms if we can consider Shakespeare's "neither a borrower or lender be" and ask why an economist would disagree.

Not what the stupid proof for some formulae is, when I can look that up in a book and besides I was among the best statisticians in NUS until that upstart, false economist, real mathematician, friend and good colleague DF popped along. And, also, until the mathematicians changed the rules beyond the scope of economics, for what is a Euclidean norm but a mathematical concept for distance, whereas linear regression regresses, as it were, to demand and supply and other beautiful academic things where I reside. That's the point.

Anything that interests me!