Conference on Social Science Statistics

19 September 2006 at 9:48 am Peter G. Klein 7 comments

| Peter Klein |

Attention, _______ometricians (econ, psych, soci, cli . . . ). The University of Missouri is hosting the Winemiller 2006 Conference on Methodological Developments of Statistics in the Social Sciences, October 11-14. Topics include structural equations modeling, multilevel models, cluster analysis, social networks, measurement theory, Bayesian methods, survey data analysis, computational issues, and missing data.

If you don’t do quantitative research come anyway and join me for a protest rally outside the conference facility, where we’ll chant “Case Studies Forever!” and burn a stack of SAS manuals. (Not really.)

NB: On structural equations modeling see this and this.

Entry filed under: - Klein -, Methods/Methodology/Theory of Science.

New Online Journal — COPE More on Elite Universities

7 Comments Add your own

1. Bo | 20 September 2006 at 10:28 am

Very interesting! I am currently struggling with a conceptual issue, which I am having difficulty solving via SEM. I am attempting to estimate a latent growth (curve) model (Yes Peter, SEM can handle longitudinal data) where certain of my predictor variables are time-variant – since most of the more advanced studies using SEM comes from educational psychology etc where most predictor variables (such as gender) are time-invariant, I am looking for a good source of information on how to model time-variance in predictor variables in a change model?
2. Peter Klein | 20 September 2006 at 11:36 pm

Bo, I ran your question by my student Jianhong Xue, a specialist in SEM, and he recommends “Latent Growth Curve Analysis: A Gentle Introduction” by Alan C. Acock:

Click to access lgcgeneral.pdf
3. Bo | 21 September 2006 at 10:14 am

Thank you! And also thank you for alerting me to where I could find my SEM book (still on Nicolajs table…):)
Bo
4. Mark E Hoffer | 26 September 2006 at 12:45 pm

For us young Grasshoppers about, could posters be so courteous to unpack, at least once, their acronyms?

Thank you very much, in advance~
5. Peter Klein | 27 September 2006 at 1:08 pm

Mark, which acronyms did you have in mind? SEM is structural equations modeling.
6. Mark E Hoffer | 27 September 2006 at 8:08 pm

Good Dr. Klein,

SEM was certainly one. And, thank you.

SAS might be the other. Though, it was a general wish, as well :)
7. Peter Klein | 27 September 2006 at 9:39 pm

SAS is — hmmm, now that you mention it, what does SAS stand for? It’s become rather like ISBN or SAT, a word unto itself. I seem to recall the name Statistical Analysis Software, but nothing at http://www.sas.com confirms this.