國立高雄大學圖資館 |

登入

回首頁

Essays on political methodology.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Essays on political methodology.
作者:	Imai, Kosuke.
面頁冊數:	196 p.
附註:	Chair: Gary King.
附註:	Source: Dissertation Abstracts International, Volume: 64-09, Section: A, page: 3458.
Contained By:	Dissertation Abstracts International64-09A.
標題:	Political Science, General.
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3106650
ISBN:	049654215X

Essays on political methodology.
Imai, Kosuke.

Essays on political methodology. [electronic resource] - 196 p.

Chair: Gary King.

Thesis (Ph.D.)--Harvard University, 2003.

The final essay develops efficient Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Despite its theoretical appeal for the analysis of candidate choice by individual voters, few have used multinomial probit because model fitting is computationally demanding. The proposed algorithms overcome this difficulty and dominate other available methods in that they are easy to implement, quickly converge, and have a more attractive prior specification. A software is made publicly available so that applied researchers can reliably estimate the multinomial probit model.

ISBN: 049654215XSubjects--Topical Terms:

212408
Political Science, General.

Essays on political methodology.
LDR:03368nmm _2200301 _450 001 162114
005 20051017073412.5
008 230606s2003 eng d
020 $a 049654215X
035 $a 00148615
035 $a 162114
040 $a UnM $c UnM
100 0 $a Imai, Kosuke. $3 227227
245 1 0 $a Essays on political methodology. $h [electronic resource]
300 $a 196 p.
500 $a Chair: Gary King.
500 $a Source: Dissertation Abstracts International, Volume: 64-09, Section: A, page: 3458.
502 $a Thesis (Ph.D.)--Harvard University, 2003.
520 # $a The final essay develops efficient Markov chain Monte Carlo algorithms for Bayesian analysis of the multinomial probit model. Despite its theoretical appeal for the analysis of candidate choice by individual voters, few have used multinomial probit because model fitting is computationally demanding. The proposed algorithms overcome this difficulty and dominate other available methods in that they are easy to implement, quickly converge, and have a more attractive prior specification. A software is made publicly available so that applied researchers can reliably estimate the multinomial probit model.
520 # $a The first essay considers causal inferences with field experiments using a voter mobilization study as an example. It introduces statistical methods that enable us to uncover discrepancies between experimental design and actual implementation. The essay also shows that propensity score matching can be used to adjust for implementation errors. The analysis reverses the previous finding and suggests that telephone canvassing increases voter turnout.
520 # $a The second essay examines causal inferences with observational studies. It develops the theoretical properties of the propensity function, which generalizes the propensity score used in the first essay. Previous propensity score methods have been limited to binary treatments. In contrast, the proposed method can also be applied to non-binary treatment regimes that are commonly analyzed for causal inferences in observational studies.
520 # $a The third essay applies the Bayesian model averaging method to the ecological inference problem. It analyzes illegally counted overseas absentee ballots in Florida that may have determined the outcome of the 2000 US presidential election. The essay delineates the conditions under which Al Gore would have been elected president, and shows the effectiveness of the Republican effort to convince local election officials to count invalid ballots in Republican counties and not count them in Democratic counties.
520 # $a This thesis is a collection of four independent essays. Each essay is motivated by practical problems that are often encountered in empirical research, and offers solutions by applying existing statistical methods and developing new ones. In particular, the thesis explores aspects of causal inference and Bayesian inference that have recently attracted attention in the field of political methodology.
590 $a School code: 0084.
650 # 0 $a Political Science, General. $3 212408
650 # 0 $a Statistics. $3 182057
710 0 # $a Harvard University. $3 212445
773 0 # $g 64-09A. $t Dissertation Abstracts International
790 $a 0084
790 1 0 $a King, Gary, $e advisor
791 $a Ph.D.
792 $a 2003
856 4 0 $u http://libsw.nuk.edu.tw/login?url=http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3106650 $z http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3106650