R Meets Google

Always on the lookout for innovative and free ways to implement data analysis, I was intrigued by a post on Andrew Gelman's blog yesterday, titled "More Data Tools Worth Using From Google". It referred to a 2009 post, "How to Use A Google Spreadsheet as Data in R" (recently updated here) on the Revolutions blog. The title of the post tells it all.

Definitely worth knowing about! Free spreadsheet; free statistical software. What more could you want?

But wait ...... there's more!

In a comment from Zach on the Gelman blog this morning, I learned about Google Motion Charts With R. As Zach notes, it allows you to take data from R and graph it directly using the Google Motion charts API. You may find that it's fussy over which browser you use, but that's O.K. It's recent release, and the R package you need is on the CRAN site here.

Have fun!

© 2011, David E. Giles

Micronumerosity

In a much earlier post I took a jab at the excessive attention paid to the concept of "multicollinearity", historically, in econometrics text books.

Art Goldberger (1930-2009) made numerous important contributions to econometrics, and modelling in the social sciences in general. He wrote several great texts, the earliest of which (Goldberger, 1964) was one of the very first to use the matrix notation that we now take as standard for the linear regression model.

In one of  his text books, Art also poked fun at the attention given to multicollinearity, and I'm going to share his parody with you here in full. In a couple of places I've had to replace formulae with words. What follows is from Chapter 23.3. of Goldberger (1991):

P-Values of Differences of P-Values of Differences of....

An interesting post on Andrew Gelman's blog last week reminded us that we have to be careful to distinguish between the statistical significance of the difference between two "effects", and the difference between the significances of two separate effects. They're not the same thing, and it's the first of these that's usually relevant.

Let's re-phrase this, and put it in baby econometrics terms.

Econometrics and One-Way Streets

It's a nice sunny day out there, so you decide to get on your bike and pedal down the street a couple of blocks to your favourite ice cream parlour. Great idea! Except that it's a one-way street and you're pedalling against the traffic. Fortunately, on this particular day, most people have already headed for the beach, there are very few cars that have to avoid you, and somehow you make it to your destination in one piece. Whew!

Now, maybe your ears are suffering from some of the abuse you received along the way - maybe not. I guess it would depend on exactly who you encountered during your little trip. In any event, you savour your well-earned ice cream and feel pretty good about yourself, and life in general. You could have travelled the longer route, around the block, to avoid the one-way street, but gee, the end result was the same, so that's all that matters. Right?

A Tale of Two Tests

Here's a puzzle for you. It relates to two very standard tests that you usually encounter in a first (proper) course in econometrics. One is the Chow (1960) test for a structural break in the regression model's coefficient vector; and the other is the Goldfeld and Quandt (1965) test for homoskedasticity, against a particularly simple form of heteroskedasticity.

What's the puzzle, exactly?

Where are You Now?

One of the great thrills of this job is working together with students as they learn about econometrics. I've been most fortunate to have been associated with quite a few students who have had enough interest in the subject to subsequently go on to undertake graduate research with me.

Along the way, I've worked with some great people. They've been talented, dedicated, and a lot of fun to be around. If they learned anything from me, then I'm grateful - because I certainly learned from them, at least ten-fold.

I thought it was time to mention these students and to acknowledge their achievements. So, I've added a new Former Students page to this blog.

My fear is that I may have omitted someone from this Graduate Students page. My hope is that I'll hear from those of you who want to update me on your career progress and current position.

© 2011, David E. Giles

Still Searching for the Number of Weeks in a Year

When I put up a post titled "How Many Weeks Are There in a Year" back in April, little did I know how many hits it would get. For a while I was intrigued to see that visitors kept arriving. They still do - every day, without fail.

Then I realized the reason why. It wasn't because the readers of this post were in search of econometric enlightenment. Oh no! There was a much more obvious reason. They genuinely want to know the answer to the question posed in the title of that post!

A quick look at the blog "stats" revealed that there are some popular web search strings that lead these poor souls, no doubt kicking and screaming, to this site.

Beware of Econometricians Bearing Spreadsheets

"Let's not kid ourselves: the most widely used piece of

software for statistics is Excel"

(B. D. Ripley, RSS Conference, 2002)

What a sad state of affairs! Sad, but true when you think of all of the number crunching going on in those corporate towers.

With the billions of dollars that are at stake when some of those spreadsheets are being used by the uninitiated, you'd think (and hope) that the calculations are squeaky clean in terms of reliability. Unfortunately, you'd be wrong!

A huge number of reputable studies over the years - ranging from McCullough (1998, 1999), to the special section in Computational Statistics & Data Analysis in 2008 - have pointed out some of the numerical inaccuracies in various releases of some widely used spreadsheets. With reputations at stake, and the potential for litigation, you'd again think (and hope) that by now the purveyors of such software would be on the ball. Not so, it seems!

An Overly Confident (Future) Nobel Laureate

For some reason, students often have trouble interpreting confidence intervals correctly. Suppose they're presented with an OLS estimate of 1.1 for a regression coefficient, and an associated 95% confidence interval of [0.9,1.3]. Unfortunately, you sometimes see interpretations along the following lines: 

• There's a 95% probability that the true value of the regression coefficient lies in the interval [0.9,1.3].
• This interval includes the true value of the regression coefficient 95% of the time.

So, what's wrong with these statements?

Missing Keys and Econometrics

There couldn't possibly be any connection between conducting econometric analysis and looking for your lost keys, could there? Or, maybe there could!

Jeff Racine (McMaster U.) put a nice little piece up on his web page at the start of this month. It's titled Find Your Keys Yet?, and has the sub-title "Some Thoughts on Parametric Model Misspecification". Jeff rightly points out some of the difficulties associated with the concept of "the true model" in econometrics, and the importance of specification testing in the games we play.

BTW, this ties in with "Darren's" comments on my earlier post, Cookbook Econometrics.

Students of econometrics - please read Jeff's piece. Teachers of econometrics - ditto!

© 2011, David E. Giles