Understanding how best to provide training for task analysis remains unspecified. This study examined how different training methods affect knowledge acquisition, focusing on content learned and errors made by novices in the initial phase of learning Hierarchical Task Analysis (HTA). Participants were assigned to one of three types of declarative instructions: procedures, decision/actions, and concept map, which were representative of different types of diagrams (matrix, network, hierarchy). After instructions, participants performed task analyses of five different tasks. Questionnaire data (declarative knowledge) and task analyses (procedural knowledge) were coded on five criteria: hierarchical representation, stating high-level goal, stating plan, stating subgoals, and satisfaction criteria. Results indicated that various aspects of the task analyses were influenced more by the nature of the task than the differences in training material at this early stage of learning. Participants recognized the importance of mentioning the high-level goal, redescribing goals into subgoals, and the plan. However, the hierarchical aspect and considering satisfaction criteria were not prevalent in either recall test or the task analyses produced. Participants in the concept map condition did identify a larger number of subgoals, but further studies are needed to understand this effect. This study informs research about various types of diagrams that are useful for training and adds to the literature on training HTA itself.