dR. sEOL'S hOMEPAGe
R을 GUI 방식으로 구현한 오픈소스 통계 프로그램: jamovi
Just click on the cloud version of jamovi
jamovi Korea :Facebook Group에서 jamovi프로그램과 관련된 다양한 내용및 Q&A를 운영하고 있습니다.
"통계의 이해" 책을 "jamovi 통계프로그램의 이해와 활용" 의 보조교재로 함께 활용하세요.
지은림,설현수,김성숙(2005). 통계의 이해 (160page, download free)
1. Seol, H.(2020). seolmatrix: Correlations suite for jamovi [jamovi module]. Retrieved from https://github.com/hyunsooseol/seolmatrix
seolmatrix module has been developed by jmvtools for the first time in Korea.
This module is a tool for calculating correlations, such as Pearson, Partial, Point-Biserial, Tetrachoric correlation, rater reliability including Light's Kappa, Intraclass Correlation Coefficient(ICC), Bootstrap agreement, Multilevel correlation, and Concordance correlation.
It also allows users to produce Gaussian Graphical Model.
2. Seol, H.(2020). snowIRT: Item Response Theory for jamovi [jamovi module]. Retrieved from https://github.com/hyunsooseol/snowIRT
snowIRT module is useful for developing psychological testing with dichotomous and polytomous data.
This module provides the following statistics with MMLE algorithm in the Rasch unidimensional model.
Item and Person Statistics (Measure, S.E.measure, Infit and Outfit Mnsq).
Model fit(person reliability, MADaQ3 with p value, and Q3 correlation matrix for the assumption of local independence.
Partial credit measure and thresholds for polytomous data.
Plots including Wright Map and Item Characteristic Curve of each item based on Master's Partial Credit Model.
Raju's area method for Differential Item Functioning(DIF) with plots of z statistic and characteristic curve of each item
The rationale of snowIRT module is described in the book; https://bookdown.org/dkatz/Rasch_Biome
3. Seol, H.(2020). snowRMM: Rasch Mixture Model for jamovi [jamovi module]. Retrieved from https://github.com/hyunsooseol/snowRMM
Latent Class Analysis was implemented within snowRMM module.
Rasch Mixture model or Mixed Rasch Models (MRMs; Rost, 1989, 1990, 1996; Rost & von Davier, 1995) formally integrate latent class analysis (cf. Formann, 1984; Rost, 1988, 1996) and the Rasch model.
The basic idea is the following: If the Rasch model does not hold for a given data set although it could be expected to hold from theoretical considerations, the reason may be that the sample consists of more than one subgroups (classes) of subjects, in the sense of a mixture distribution model.
Within each of these classes, the Rasch model holds. This means that the items are measuring unidimensionally within each class, but the classes differ in the rank order and in the differences of item difficulties.
This module provides the following statistics with JMLE;
Item and person statistics and Model fit(AIC, BIC, CAIC).
Person statistics(average ability, person membership).
Bootstrap individual item fit with 95% confidence interval.
snowRMM(Youtube without voice)
This module allows users to analyze k-means and hierarchical clustering. Furthermore, it also provides various visualization results related to Principal Component, Correspondence Analysis, Multiple Factor Analysis, Multidimensional scaling. and Discriminant Analysis.
Multiple factor analysis (MFA) (J. Pagès 2002) is a multivariate data analysis method for summarizing and visualizing a complex data table in which individuals are described by several sets of variables (quantitative and /or qualitative) structured into groups.