Friday, December 29, 2006

More cool stuff for Eclipse

A Eclipse package that handles LaTeX document writing called "TeXlipse" is very cool. It has at least one thing (and maybe more) that WinEdt does not: code folding.

Eclipse can also work with R, the statistical package. A plug-in called "StatET". It is pretty basic right now (especially when compared to another open source software, "Tinn-R"), but I think it has the potential to become a real deal.

For TeXlipse:
For StatET:

Monday, December 25, 2006

Eclipse + CDT + MinGW

Following instructions here:, I successfully configured Eclipse + CDT + MinGW on my Vaio SZ360 laptop. Now I have a rather comfortable C/C++ development environment.

Thursday, December 21, 2006

Cool windows utility software to boost performance

I rarely use windows utility software after several unhappy incidents many years ago. But this software changes my mind:

My new Sony SZ360 was having troubles because of too many junk software pre-installed. I was thinking about re-install the whole system. Since I am traveling and did bring many software installation disc with me, I figure maybe I try to "optimize" the system. So far it worked very well. The free utility actually boosts the performance of my laptop, and I am not sure if I want to re-install the whole system.

This is one of the rare moments that I find a software that actually delivers its promise.

Inequality might be the key

The linkage between childhood obesity and socioeconomic status is certainly very interesting. It has been studied by several scholars before, as summarized by Wang and Lobstein (2006). The results, however, raise more questions than they have answered. The most important question is this: why this relationship keeps changing from research to research, are there anything missing in these research?

I have a bold hypothesis on this: social inequality is the most important moderator of the relationship between childhood obesity and SES.

It will be very interesting to see whether the empirical evidence supports this argument.

Tuesday, December 19, 2006

Python for scientific computing

Here is a very useful tutorial (video and PDF) on how to do scientific computing using the Python programming language and its addons:

I find it very helpful.

Sunday, December 17, 2006

PAA 2007 presentations

I will present my joint research with Bengt Muthen, "Growth Mixture Modeling for Sequential Growth Processes", at PAA 2007 annual meeting at New York. I also have a poster session, "Local Variation in the One-Child Policy and Its Determinants: 1989-2000". Lots of work between now and March 2007.

Tuesday, December 12, 2006

Data can be used to study childhood obesity in China

Two data sets contains the information that can be used to study childhood obesity in China: the China Health and Nutrition Survey (CHNS) and the Food and Nutrition Surveillance System (FNSS).

While the CHNS data is freely available to researchers (, the status/availability of FNSS is unknown.

Monday, December 11, 2006

Cross-sectional data

Sarah sent me a list of cross-sectional data that can be used to study childhood obesity:

There are several other data sources that are often mentioned in the literature:
1. NHANES (National Health and Nutrition Examination Survey):
2. NHES (National Health Examination Survey):
3. NGHS (National Heart, Lung and Blood Institute's Growth and Health Survey):

Will ask Sarah to figure out their relative strengths and weaknesses and how they can be used in our research.

Mplus vs. Stata for Heavy Computation

Both Stata and Mplus have multiple processor support. Stata/MP is licensed on per CPU basis, while Mplus is licensed on per machine basis, no matte how many processors a machine has. Combining a non-MP version of Stata and Mplus seems to provide an ideal solution for researchers who have to deal with complicated statistical models, especially those involve latent variables. One can use Stata for all the data manipulation and some light duty estimation, while relying on Mplus (on a fast core duo machine) for the heavy computation. The Stata ado file "stata2mplus.ado", provided by the ATS computing at UCLA provides an ideal solution to bridge the two packages.

Sunday, December 10, 2006

Mplus 5.0

Had a pleasant talk with Linda couple of days ago (that was the first time I met her in person). She told me that version 4.2 should be called version 5.0 in terms of functionality. I personally agree, the parallel computation capability alone is a breakthrough in the development of Mplus. She also told me that the next major improvement will be on 1) making latent transition/latent Markov modeling easier, 2) extending to 3-level or even higher level models and, 3) porting to Unix/Linux . I am looking forward to seeing that coming! Good luck, Linda!