Introduction to GLM in R: Binary, Multinomial, and Ordinal Logistic Regression (Part 1)

These videos provide a tutorial on estimating models for categorical dependent variables in R. Generalized linear models (GLM), as the name implies, generalize ordinary linear regression. Common applications include binary (two categories), multinomial (3+ unordered categories) and ordinal regression (3+ ordered categories). Part 1 focuses on binary logistic regression, from model estimation to evaluating fit (e.g., likelihood ratio test, Hosmer-Lemeshow goodness-of-fit) and classification (e.g., area under the curve). Part 2 continues with a focus on multinomial and ordinal logistic regression.

Refs/Recs:

Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic regression (3rd ed.). John Wiley & Sons.

Nelder, J. A., & Wedderburn, R. W. M. (1972). Generalized linear models. Journal of the Royal Statistical Society, 135(3), 370-384.

[yaregaltayelegesse]

Hello sir, I have attended your you tube presentation and I found it interesting and have used it. I am PhD candidate in India Punjabi University, Patiala. I am Ethiopian who came to India getting scholarship.
I have submitted my synopsis that has one dependent variable and 3 independent variables all of whose data are collected through likert scale each having multiple items. I planned to do multiple linear regressions in my synopsis but after attending different YouTube presentations and reading different books, I have come to the conclusion that I cannot because linear regression makes use of mean. But data from rating scales is non-parametric data that has to be analyzed with non-parametric tests and from ordinal data it is not possible to calculate mean. Now, I have decided to analyze my data via ordinal regression in IRT (Item response theory). However, I have not got enough information how to conduct it. All presentations of IRT graded response I have come across are about discriminations, difficulty level of items, guessing and ability of persons. Hence, how can I conduct ordinal regression analysis focused on dependent and independent variables? Please help me maam in whatever ways. Or you can make me have contact with any known statisticians you know.