DOI
https://doi.org/10.1007/s10459-007-9081-3

Summary
The article focuses on the form and level of guidance natural language intelligent tutoring system (ITS) should be used for training pathologists to analyze and report on melanoma cases. Firstly, the authors present the architecture and design of developed natural language interface (NLI) for ITS in diagnostic pathology. Secondly, the experimental environment was described by the means of participants, activities, survey methodology. The crossover 2x2 factorial design was used in the study of twenty students from four academic programs. Thirdly, the authors provide the performance metrics of the system, results of ANOVA analysis of pre-test and post-test self-scores, self-grading analysis, and attitude analysis towards the system were presented. Finally, a discussion regarding research questions is provided. The authors consider the improvement of report writing skills, the effect of feedback timing on the performance gains and correlation between the self-assessment and actual performance.

Relevance for Complex Systems Knowledge
The authors present natural language intelligent tutoring system for training pathologists. This system enables students to interact with the system and learn through conversational interfaces. One of its features is that it corresponds human tutoring by engaging in conversation, offering guidance if help is needed and encouraging autonomy decisions. The study showed that students who used this system showed similar results to those who did not use it after examining a smaller number of examples. It was concluded that usage of such systems could be particularly useful in cases where it is a limited number of case examples. The relevance to the complex system is the teaching / learning approach where students make decisions by evaluating a complex provided information on the analyzed case and have guidance from a natural language interface.

Point of Strength
The point of strength of this article is a provided discussion on how natural language interface can change learning / teaching outcome.