The purpose of this study is to present the validity issues raised in task engagement analysis, a class analysis tool, and to explore ways to solve these problems. For this purpose, data such as task engagement analysis results, class videos, and the researcher's reflective journal were collected. These data were collected based on the researcher's many years of experience in instructional consultation. The collected data were analyzed according to an analysis of the degree of concentration on the task. The validity problems of the task engagement analysis derived through this study were revealed as ambiguity in judgment about learner behavior, reliability problems due to single observer analysis, limitations in analysis due to flat observation, and lack of evidence for the causes of results. To solve these problems, the following four methods of securing validity are suggested: First, clarification of the action in parallel with class video recording. Second, securing reliability through multiple observers. Third, three-dimensional analysis using other instructional analysis tools. Fourth, the collection of rich data through interviews and surveys. The following conclusions were drawn. First, use of the task engagement analysis tool can contribute to strengthening the competency of observers who perform class analysis. Second, an instructional analyst or observer can derive highly valid instructional analysis results by considering the method for securing validity suggested in this study. Third, contributing to the practical improvement of the instructor's class can contribute to the improvement of the instructor's class expertise.