Critical thinking and the ability to make connections between knowledge from different areas are for the students crucial to master. This requires the implementation of modern educational approaches such as integrated teaching. In this contribution, integrated teaching is perceived as teaching that connects subjects matters, expected outcomes, key competencies, and aims of integrated subjects, and simultaneously forms a new aim as a result of an integrated whole. But, at least in the Czech Republic, integrated teaching is not common in schools. To find out why this is and what the teachers see as obstacles for implementation of integrated teaching was the aim of our research. In this contribution data analysis and results obtained by traditional approach are compared with the approach using artificial intelligence (hereinafter referred to as AI) namely ChatGPT.

For data collection, the Q methodology was chosen because it combines qualitative and quantitative techniques to study subjectivity. Q methodology consists of five steps (Q types creation, choice of respondents, Q sort, data analysis, data interpretation). For each step of the Q methodology, a comparison of the use of AI and the traditional approach has been made, but this contribution focusses on the fourth and the fifth step of the Q methodology.

The fourth step consists of data analysis that includes factor analysis. Collected data was first processed using IBM SPSS Statistics program combined with PQMethod program. While ChatGPT generated identical results for Bartlett's test (including chi-square statistics and degrees of freedom), it could not calculate the Kaiser-Meyer-Olkin measure of sampling adequacy because of its inability to perform the necessary oblimin rotation. The correlation matrix showed the same values using IBM SPSS Statistics program, PQMethod program and ChatGPT. Principal Component Analysis with varimax rotation was performed using the PQMethod program as the traditional approach in contrast to ChatGPT. The PQMethod program showed eight factors, two of which were significant. ChatGPT showed six factors, four of which were significant. Despite the differences, there were similarities between the two approaches, particularly in one overlapping factor. This factor showed some similarities in z-scores and factor Q-Sort Values.

The fifth step consists of data interpretation. Data interpretation differed depending on the use of ChatGPT versus the traditional approach that consists of researchers’ conclusions. Obtained different results caused the different data interpretation. Because of that, ChatGPT was used to interpret the data collected by the traditional approach. Both approaches concluded for the Factor 2 respondent group that successful implementation of integrated teaching requires, among other things, overcoming teachers' fears that students will know more than they do or addressing teacher workload issues.

In conclusion, the contribution will present comparisons of the use of AI and the traditional approach for all five steps of the Q methodology, focussing on the fourth and fifth steps, which consists of data analysis (including factor analysis) and data interpretation.

Keywords: Q methodology, AI, ChatGPT, Integrated teaching.