4.2 Managing conflicting findings
One of the greatest challenges and opportunities in mixed methods research is integration. But what happens when your quantitative data tells one story and your qualitative data tells another?
Rigour requires that you do not simply ignore the contradiction or assume one method is ‘better’ than the other. Instead these thematic conflicts can lead to deeper insights and a re-evaluation of your core concepts.
Activity 4.2: Resolving the AI tutor conflict
Read the following research scenario:
You are conducting a concurrent mixed methods study (QUAN + QUAL taking place at the same time) to answer whether the introduction of an AI personal tutor improves student outcomes in a first-year Computer Science course.
- QUAN: A randomised controlled trial with 400 students shows that the AI group scored 15% higher on final exams and completed assignments 30% faster than the human-tutor group.
- QUAL: Semi-structured interviews with 12 students from the AI group reveal that students felt ‘robbed of critical thinking’. They reported that the AI often gave them the answer rather than teaching the logic, leading to what they called ‘hollow learning’.
- Thematic conflict: While grades and speed were higher, the development of critical thinking and problem-solving skills was reportedly much lower.
Which data is more important? Do you trust the quantitative outcomes or the qualitative experiences? Make some notes in the space below.
Discussion
Neither dataset is inherently ‘more important’ and a rigorous mixed methods researcher would not discard one in favour of the other. The QUAN data accurately measures performance metrics (grades and speed) while the QUAL data accurately captures the learning process and cognitive depth.
To integrate these findings you must redefine what ‘student outcomes’ means in this context. The contradiction reveals that efficiency does not equal deep learning, highlighting an important point that either method alone would have missed.
4.1 Designing for rigour: a starting scenario


