Let's try some great things together with me ; )

• # Stan
• # ggpolt2
• # R

• # CAT demo
• # Estimation
• # Ph.D. application

Stan

Estimating parameters of a model is always a thrilling moment, especially when that is your own model! One popular procedure for realizing this goal is the Markov chain Monte Carlo (MCMC) method. It has a long tradition of success in constructing desired posterior distributions and facilitating solutions. For MCMCers, Stan offers a really fast sampler and an easy-to-learn modeling language. With its spectacular features, it has amassed a considerable user base. Although most of our interactions with Stan are joyful, there are still potential holes that might trouble us for several weeks or even months. In this article, I will share four tips to help you avoid pitfalls and save your valuable time! [Read more]

ggpolt2

This article may be useful to you if you want to have similar features on your plots. Here you can figure out how to include additional text and lines, how to organize the point according to a specific order, how to change the transparency, etc. [Read more]

R

When we first come up with a plan to learn something new, it is more likely for us to keep working toward that goal if we get positive feedback immediately. Now! The prompt feedback you are looking for when learning R is here! After reading this article, you surely will be equipped with the ability to start your own project and the paintbrush to draw a bigger picture. Let's start! [Read more]

CAT demo

Process data tell a lot. It is a natural thought to utilize response time in computer-based testing since such data is readily available. Herein, I will illustrate how to incorporate response time in item selection methods for computerized adaptive testing. In this demo, you can see (a) the generation process for response time based on the lognormal model, (b) a complete routine for computerized adaptive testing, and (c) the performance of different item selection methods. [Read more]

Estimation

Obtaining the examinee's ability estimate is always the very first step for a psychometrician to move further, for example, servicing for the selection of the next item during the process of online calibration or adaptive testing. In this article, a detailed description of R codes for maximum likelihood estimation using Newton-Raphson accompanied by the implementation of the downhill algorithm is presented. With these codes, you can easily use them to estimate your own theta with your item parameters! [Read more]

Ph.D. Application