The third Japanese Stata Users Group meeting was Saturday, 16 September 2017 at Kyoto Research Park, but you can view the program and presentation slides below.
Proceedings
10:10–11:00 |
Abstract:
Regarding agricultural productivity in developing countries, even
in the same natural environment, it is common for a large disparity
to arise among households of each farm. It is certain that
these differences are derived from farm household individuals,
characteristics of households, etc., however, in recent research,
connection with neighbors and relatives, that is, that difference in
relationships within social networks, is one factor of disparity that
has been pointed out. Social networks in rural developing countries are
thought to affect agricultural production through three paths: (1)
provision of new technical information, (2) mutual insurance function, and
(3) relaxation of labor market imperfections and credit constraints.
However, because of the difficulty of accumulation of research in the field,
this is also not fully done. Therefore, based on household
survey data of a rural Madagascar village collected independently,
we focus on the provision of new technical information and how the
social network of rural communities influences the adoption of agricultural
new technology.
Additional information: japan17_Kurita.pdf
Masao Kurita
Kwansei Gakuin University
|
11:00–11:50 |
Abstract:
Item response theory (IRT) simultaneously estimates potential
characteristics (ability, psychological characteristics, etc.)
of each respondent and difficulty and discrimination of each item
from respondents' answer patterns for each item of the scale. It
is a test theory that can be done. IRT has many advantages in
estimating reliability and validity for classical test theory, such
as being able to study the measurement accuracy of each item in
detail without depending on the characteristics of the sample group,
such as psychological characteristics, as well as in the
development of a scale for measuring the size of a sample. In this
study, I used Stata to develop a scale to measure school atmosphere
(school culture) and to examine the reliability and validity of the
scale by IRT. Because the answer was multivalued data of five methods, a
multilevel response model (graded response model) was used.
Additional information: japan17_Nishimura.pptx
Rinko Nishimura
Hamamatsu University School of Medicine Research Center for Child Mental Development
|
11:50–1:00 |
Abstract:
The propensity to estimate the treatment effect introduces the basic idea of score analysis
and explains the functional difference between the Stata command teffects psmatch
and the conventional ado-file psmatch2, pscore command.
Using these commands, you can avoid unnecessary confusion when referring
to previous research.
Additional information: japan17_Lightstone.pdf
Lightstone Co., Ltd.
|
1:00–1:50 |
Abstract:
In the field of social science, analysis by survey observation data is
the majority and, unlike experimental data, it is not easy to measure
the effect of a specific treatment. For example, in the
field of labor economics, the effects of treatments such as university
admission and vocational training participation on subsequent wages will
be analyzed, but becausee such treatments are not randomly assigned, bias
on wages is biased. Several analytical methods have been developed to
deal with these problems, one of which is treatment estimation.
I will introduce analysis examples, especially when there are multiple treatments.
Additional information: japan17_Mizuochi.pptx
Masaaki Mizuochi
Nanzan University, Faculty of Policy Studies
|
1:50–2:40 |
Abstract:
A programming language called Mata is available in Stata.
In this presentation, I will show that flexible program development
can be performed according to the purpose of each researcher
using Mata. I will introduce
examples of spatial statistics and spatial econometric
applications that have been actively performed in recent years at
Stata. The advantage of Mata is that it handles
matrix operations intuitively. Because geographical space is
treated as a matrix, one can perform flexible spatial
analysis in Stata by linking it with Mata. The range
of statistical analysis broadens widely by using the spgen command
to calculate the space lag variables developed by the author; the
getisord command can perform hotspot analysis and mutually
uses the advantages of Stata and Mata. I will also introduce
examples of using the Stata command related to the geographic
information system, such as the method of creating maps in Stata.
Additional information: japan17_Kondo.pdf
Keisuke Kondo
Research Institute of Economy, Trade and Industry
|
3:00–3:50 |
Abstract:
Cancer is a disease whose cause is clear and preventable: it
can be overcome by early diagnosis and treatment.
While Japan is under the national health insurance system, it has
become apparent in recent years that disparities in health condition and
life span are occurring because of socioeconomic reasons. In this presentation,
I analyzed the survival rate, morbidity rate (incidence), and social economic
disparity of mortality rates of cancer patients by using all public survey
data such as cancer registration data and demographic statistics. Using
socioeconomic indicators based on the residential area of cancer patients
using Osaka prefecture cancer registration data, we found that there is a
disparity in the survival rate of cancer patients. Also, looking at the
prevalence rate by degree of progress, men living in rich areas
suffered from early cancer and the prevalence rate of advanced cancer was
low. Social and economic disparities also occurred in the mortality rate of
cancer in the statistics of population dynamics throughout the country. I
introduce statistical analysis and graphical expression by Stata, including
survival analysis and various regression analysis used in this research.
Additional information: japan17_Ito.pdf
Yuri Ito
Osaka International Cancer Institute Cancer Control Center
|
3:50–4:50 |
Abstract:
Bayesian analysis has become a popular tool for many statistical
applications. Yet many statisticians have little training in the theory
of Bayesian analysis and software used to fit Bayesian models. This
talk will provide an intuitive introduction to the concepts of Bayesian
analysis and demonstrate how to fit Bayesian models using Stata. No
prior knowledge of Bayesian analysis is necessary and specific topics
will include the relationship between likelihood functions, prior, and
posterior distributions, Markov Chain Monte Carlo (MCMC) using the
Metropolis–Hastings algorithm, and how to use the new Bayes prefix in
Stata 15 to fit Bayesian models.
Additional information: japan17_Huber.pptx
Chuck Huber
StataCorp
|
Organizers
Logistics organizer
The logistics organizer for the 2017 Japanese Stata Users Group meeting is LightStone Corp., the distributor of Stata in Japan.
View the proceedings of previous Stata Users Group meetings.