Intelligent (Language) Tutoring Systems: A Second-order Meta-analytic Review

 Intelligent Tutoring Systems (ITSs) are referred to as computerized learning environments that incorporated pedagogical, cognitive learning, knowledge representation theories into computational models of tutor, learner, and domain knowledge respectively in order to provide individualized instruction in diverse subject areas. Since they take advantages of many technological artifacts of artificial intelligence (AI), knowledge representation models, and computational linguistics, they have been both object and means of research into AI, pedagogy, psychology, applied linguistics etc. Under the circumstances, the bulk of research findings is expectable and unavoidable. Sometimes researchers are badly in need of conducting first- or second-order meta-analytic reviews. This study first introduces the components of a typical ITS; then provides a descriptive account of the effectiveness, potential, and requirements of ITSs for foreign language teaching and learning (FLTL); and finally and most importantly, as a second-order review, synthesizes and describes the results of some previous meta-analytic reviews in reference to a few prevailing themes in the ITS research area such as the effects of the forms of instruction, subject matter variation, ITSs’ pedagogical features etc. Toward the third end, exclusion and inclusion procedures were applied to select seven target reviews using keyphrases scheme. The thematic meta-synthesis showed that the findings of almost all these reviews are consistent and congruent with some slight variations due to the study features. It is implied that meta-analyses like this can, by implication, inform and advance the science of ITS design, i.e. a theory of ITS design.