This study investigates the development and evaluation of a Retrieval-Augmented Generation (RAG)-based statistics tutor designed to assist students with quantitative analysis methods. The RAG approach was employed to address the well-documented issue of hallucination in Large Language Models (LLMs). A computer tutor was developed that utilizes ChatGPT to understand student questions in natural language and retrieves relevant information from validated course notes to generate responses. The tutor was evaluated using twenty questions curated from the course notes. Two instructors who taught the course using the course notes assessed the quality of the tutor's response using a pre-defined rubric. The evaluation results indicate that the RAG-based tutor was able to provide correct and relevant answers to all twenty questions and to generate R code snippets upon request. This study highlights the potential of RAG-based tutors to provide accurate and personalized learning experiences for students while mitigating the risk of providing fabricated information often associated with LLMs.