Journal of the Royal Statistical Society: Series A (Statistics in Society)

Cover image for Vol. 180 Issue 1

Edited By: H. Goldstein and L. Sharples

Impact Factor: 1.702

ISI Journal Citation Reports © Ranking: 2015: 13/49 (Social Sciences Mathematical Methods); 24/123 (Statistics & Probability)

Online ISSN: 1467-985X

Associated Title(s): Journal of the Royal Statistical Society: Series B (Statistical Methodology), Journal of the Royal Statistical Society: Series C (Applied Statistics), Significance

175:2


Joint modelling of longitudinal outcome and interval-censored competing risk dropout in a schizophrenia clinical trial, by R. Gueorguieva, R. Rosenheck and H. Lin, Journal of the Royal Statistical Society, Series A, Statistics in Society, Volume 175, (2012), part 2, pages 417 - 433

The "Simulation_code_Gueorguieva_et_al.txt" file contains SAS code for performing the simulation study in the paper. The program generates 200 data sets using PROC IML, fits the three models described in the paper (joint model with separate dropouts, joint model with common dropout and separate model) using PROC NLMIXED and summarizes the results. The program can be run directly in SAS by just changing the path used in the libname statement.

Ralitza Gueorguieva
Division of Biostatistics
School of Public Health and School of Medicine
Yale University
Room 215B
60 College St
New Haven
CT 06520-8034
USA

E-mail: ralitza.gueorguieva@yale.edu

Dataset (4kb)

A Bayesian non-linear model for forecasting insurance loss payments, by Y. Zhang, V. Dukic and J. Guszcza, Journal of the Royal Statistical Society, Series A, Statistics in Society, Volume 175, (2012), part 2, pages 637–656.

Data set:
The data used in this paper are available in the file "wc10InsData.txt". This data set is extracted from the statutory annual statements that insurance companies are required to report to the National Association of Insurance Commissioners each year. The line of business used is workers' compensation insurance. Of the 1,070 companies represented in the available data, we select data for 10 large insurers whose combined premium volume accounts for approximately 36% to the industry's total premium. This data set is kindly made available by the National Association of Insurance Commissioners and the Casualty Actuarial Society. The full data set is available at http://www.casact.org/research/index.cfm?fa=loss_reserves_data

Explanation of the variables in "wc10InsData.txt":
# comp - company id, from 1 to 10
# year - accident year of the claims, from 1988-1997
# ay - computed as (year-1988), to be used for indexing
# lag - development lag, from 1 to 10
# t - time elapsed since the origin of the accident year, computed as (lag*12)
# cum - cumulative paid loss
# incre- incremental paid loss
# p - accident year earned premium

Program:
The Bayesian nonlinear model is estimated using "WinBUGS" called within "R". The R code used is provided in the file "wc10InsCode.R", which also contains the WinBUGS code for the model specification. The R code should be directly executable if WinBUGS has been set up correctly and correct path of the data set is supplied.

Yanwei Zhang
Statistical Research
CNA Insurance Company
30th Floor
333 South Wabash Avenue
Chicago
IL 60604
USA

E-mail: actuary_zhang@hotmail.com

Dataset (11kb)

Evaluating continuous training programmes by using the generalized propensity score, by J. Kluve, H. Schneider, A. Uhlendorff and Z. Zhao, Journal of the Royal Statistical Society, Series A, Statistics in Society, Volume 175, (2012), part 2, pages 587 - 617.

The stata do file named "JRSSA kluve et al gps code" is the program to create Fig. 4(a). The outcome variable of this figure is employment probability two years after programme start; the main independent variable is actual duration of training. This program:

  1. Estimates the generalized propensity score (GPS);
  2. Estimates treatment effect every 7 days based on the GPS;
  3. Estimates standard errors through bootstrapping the whole process, including estimation of the GPS.

Other main results, such as Fig. 5, were estimated using similar Stata do files. Researchers interested in additional Stata code used in the paper are welcome to contact the authors.

Jochen Kluve
School of Business and Economics
Humboldt Universität zu Berlin
Spandauer Straase 1
10178 Berlin
Germany

E-mail: jochen.kluve@hu-berlin.de

Dataset (4KB)

SEARCH

SEARCH BY CITATION