[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
"Feiveson, Alan H. (JSC-SK311)" <Alan.H.Feiveson@nasa.gov> |

To |
<statalist@hsphsun2.harvard.edu> |

Subject |
RE: st: A methodological problem |

Date |
Thu, 30 Oct 2008 09:37:48 -0500 |

Carlo - Even though you are analyzing all the games ,couldn't the results from a given year be considered a realization of random outcomes over years? So each years' data can be considered an observation of a 190-vector (as Stas points out) of results. Al Feiveson -----Original Message----- From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Stas Kolenikov Sent: Thursday, October 30, 2008 9:05 AM To: statalist@hsphsun2.harvard.edu Subject: Re: st: A methodological problem 1. You really have a problem of dyadic data analysis: your actual observations are games between the two teams (right?) so you probably have 20*19/2 = 190 observations if every team played other team. Dyadic data analysis is an emerging theme, the methods are somewhat complex as they need to take symmetry of the problem into account which produces some weird non-linear constraints and puts parameters in unwieldy places in the likelihoods. Peter Hoff from U of Washington has been working on this and proposing some Bayesian solutions. Kit Baum should know something too as he worked on international trade data that has similar structure (imports and exports between two countries). 2. I don't think there are universally recognized and well established solutions with complete populations. Certainly most statistical methods start with "Let's take an i.i.d. sample from distribution f(x,theta)". In survey statistics, there are some ideas about superpopulations, the infinite populations from which a given finite population was drawn; and some methods take those superpopulations into account and explicitly talk about estimation of parameters of those superpopulations. I am not totally sure you want to get into this talk. Alternatively, there's been some amount of work on what's called apparent populations (http://www.citeulike.org/user/ctacmo/article/333410). The solutions proposed there are mostly along Bayesian lines, again: since all the data you have appears to be fixed as a population, that paradigm may indeed be a better way to go. Of course the whole idea of fixed population with fixed characteristics runs contrary, to say the least, to the nature of sports: you can make some predictions, in some sort of large numbers sense (Manchester United is more likely to win Champions League title than any of the Russian clubs... at least so far :) ), but nobody can predict the results perfectly, let alone for any particular game. That's an interesting methodological challenge. Finally, in American Statistical Association, there is Section on Statistics in Sports, and I believe there's a journal with the same title. You can take a look and see if they have something sensible for you. On 10/30/08, Carlo Amenta <carlo.amenta@gmail.com> wrote: > Dear statalisters, > > I am studying a sport team league with 20 teams in a specific years. > I am using a 2sls estimator because of simultaneity problem with a > specific variable which was confirmed using the -ivendog- procedure. > At this stage the study is cross sectional and regards a specific > season. It is correct to say that I have not any efficiency or > consistency problem wuth the estimator considering the fact that I am > studying the entire populationa and not a sample? As a matter of fact > n=20 even if very small it is not the number of observation but all > the teams in the league so the entire population. I think I have not > to worry about any inference problem. Can someone confirm that or > indicate any specific references? > Thank you > > > Carlo Amenta > University of Palermo > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/statalist/faq > * http://www.ats.ucla.edu/stat/stata/ > -- Stas Kolenikov, also found at http://stas.kolenikov.name Small print: I use this email account for mailing lists only. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

**References**:**st: A methodological problem***From:*"Carlo Amenta" <carlo.amenta@gmail.com>

**Re: st: A methodological problem***From:*"Stas Kolenikov" <skolenik@gmail.com>

- Prev by Date:
**Re: st: Insheet whiel Ignoring Headers** - Next by Date:
**Re: st: Mincer wage function as a log function - how can stata calculate the percent values** - Previous by thread:
**Re: st: A methodological problem** - Next by thread:
**st: RE: graphing a series of vectors (latent growth curve model estimate summaries)** - Index(es):

© Copyright 1996–2014 StataCorp LP | Terms of use | Privacy | Contact us | What's new | Site index |