As has been shown by previous research, students may possess various misconceptions in the area of thermal physics. In order to help them overcome misconceptions observed prior to instruction, we implemented a one-hour lecture-based intervention in their introductory thermal physics course. The intervention was held after the conventional lectures and homework sessions, and it consisted of three phases: individual working, hinting, and peer discussion. To probe students' conceptual understanding before, during, and after the intervention, use was made of a diagnostic test related to the multiphased process of an ideal gas [D.?E. Meltzer, "Am. J. Phys." 72, 1432 (2004)]. The students' conceptions were monitored by analyzing the explanations they provided and by recording the peer discussions of five voluntary pairs. The intervention helped students to realize the flaws in their explanations and increased the proportion of their scientific explanations, the increase being statistically significant in five tasks out of seven. When the same themes were addressed in a post-test, it was shown that the level of accurate explanations remained almost constant after the intervention, and hence it could be deduced that the impact had not been short-lived. In comparison with earlier studies conducted with the same material, our intervention produced a better learning outcome, the difference being 15-20 percentage points. In addition, the number of misconceptions on the part of the students was smaller in our study, although with individual exceptions. Hence, we conclude that the intervention was successful and that similar interventions could also be designed and implemented in other areas of physics. (Contains 14 tables, 7 figures, and 7 footnotes.)