CCT: Mentored Teaching Project

Mentored Teaching Project: program evaluation of teaching techniques in the MSU AI for Advanced Semiconductor Manufacturing workshop (2026).

← CCT portfolio

Project: Program Evaluation of Teaching Techniques in the MSU AI for Advanced Semiconductor Manufacturing Workshop
Setting: 5-day intensive summer workshop at Michigan State University (Summer 2026) for adult learners (technicians and pre-engineering / technical-program students), in small groups of up to 5 participants
Faculty mentor: Andrew Krause, Specialist – Teacher-Continuing, Department of Mathematics, Michigan State University
My role: Program evaluator and observer
Date completed: May 29, 2026

Project Description Outline (6 steps)

  1. Teaching and learning goal: Help adult learners in a technical workshop move from instructor-guided learning to confident, independent problem-solving with real data and AI tools.
  2. Teaching question: How do participants in the MSU AI workshop perceive the effectiveness of the teaching techniques used, especially the balance between instructor guidance and independent hands-on learning?
    • Which teaching methods do participants find most helpful?
    • Do participants feel more confident learning independently by the end?
    • How do participants experience the shift from guided instruction (Days 1–2) to applied work (Days 3–5)?
  3. Assessment technique: Anonymous pre-workshop (Day 1) and post-workshop (Day 5) surveys with Likert-scale (1–5), multiple-choice, and open-ended questions on background, preferred teaching methods, expected and experienced barriers, confidence, guidance versus independence, engagement, and overall satisfaction.
  4. Classroom practice: The workshop’s built-in progression from teacher-led to independent work. Days 1–2 combine guided instruction with hands-on semiconductor wafer fabrication and capture of images of defective substrates (a “ground-truth” dataset). Days 3–5 move to the lab, where participants clean and structure their own image data and perform AI-based classification with industry tools.
  5. Summary of results: Descriptive statistics and charts for the survey items, and thematic analysis of open-ended responses (themes: clarity, hands-on learning, independence, support, engagement). See findings below.
  6. Conclusion: Effective teaching for adult learners in technical training balances structured early guidance with gradually increasing independence, supported by hands-on work with real data, and needs extra scaffolding for learners with limited background.

Project Summary

Teaching and Learning Goal

Adult learners come to technical training with varied backgrounds, clear career motivations, and often some anxiety about whether they are prepared. The goal was to understand how the workshop’s teaching techniques help these learners engage, understand technical material, and become confident learning independently, and to use that evidence to improve future offerings of the workshop.

Teaching Question

How do participants perceive the effectiveness of the workshop’s teaching techniques, particularly the balance between instructor guidance and independent hands-on learning?

Classroom Practice (artifact)

The workshop is structured as a gradual release of responsibility: guided instruction and demonstrations during fabrication and data capture on Days 1–2, then increasingly independent data cleaning, analysis, and machine-learning classification on Days 3–5. Small groups (at most 5 participants) allow close mentoring.

My role was evaluator and observer. I did not teach the workshop sessions; I designed and administered the surveys, observed how instruction moved from guided to independent work, analyzed the data, and wrote the evaluation report with recommendations for the instructors.

Assessment Instrument (artifact)

Pre- and post-workshop surveys (evaluation plan and full survey instruments, PDF). The post-survey Likert items covered teaching effectiveness (clear explanations, organization, pace), hands-on learning (activities and real data helped), guidance versus independence (enough guidance, chances to work independently, confidence solving problems alone), engagement and relevance, and overall support for learning. Open-ended questions asked which approach helped most, when participants wanted more guidance or more independence, what to improve, and what adult learners need from instructors in technical training.

Because the project was designed as a program evaluation to improve this specific workshop, I submitted an MSU IRB determination request (HRP-512) explaining that the activity is not human subject research. Participation was voluntary and anonymous, and no identifying or sensitive information was collected.

Data and Summary of Findings (artifacts: tables and charts)

Pre-workshop survey: concept familiarity, confidence in each concept, preferred learning methods, and anticipated barriers.
Post-workshop survey: understanding ratings, confidence growth, independent problem-solving, what helped most, and challenges experienced.

Before the workshop

  • Most participants were pre-engineering students; others came from occupational or technical programs.
  • Participants were more familiar and confident with algebraic equations, graphs, and linear relationships (average confidence 4.2–4.4 out of 5) than with vectors (3.0), statistics (2.7), and matrices (2.5), the mathematics behind machine learning.
  • Every participant chose hands-on practice as a preferred way to learn and 90% chose demonstrations, while only 20% chose lecture and 10% group work.
  • The most anticipated barrier was lack of prior knowledge, followed by time constraints and confidence.
  • Participants were motivated by career-relevant skills in robotics, automation, manufacturing, and engineering.

After the workshop

  • Participants reported stronger understanding of linear relationships, data analysis, machine-learning concepts, and applied AI, with the highest confidence in concepts reinforced through hands-on work.
  • 70% reported higher confidence after the workshop (30% much higher, 40% slightly higher), and 60% rated their independent problem-solving ability higher than before.
  • Average post-workshop understanding ratings were 4.4/5 for linear relationships and for working with data patterns, 3.9 for matrix-based thinking, and 3.7 for dimensionality reduction (PCA).
  • 9 of 10 respondents said the balance of guidance and independence was right; 1 wanted more guidance; none wanted more independence.
  • 9 of 10 rated the workshop “Excellent” (1 “Good”), and 9 of 10 would “definitely” recommend it (1 “probably”).
  • Hands-on activities, lab work, demonstrations, and instructor explanations were named most helpful; instructors were described as knowledgeable, approachable, and supportive.
  • The challenges participants actually experienced were lack of prior knowledge (4 of 10), difficulty with tools or technology (2), and time or fatigue (2); 2 reported no major challenges, and none said the pace was too fast. Suggestions included more breaks, better computer performance, more setup guidance for tools, and preparatory materials for advanced mathematics.

Conclusion

The workshop’s teaching worked well for adult learners: hands-on work with real, career-relevant data, together with clear early guidance that gradually gave way to independent application, raised participants’ confidence and satisfaction. This matches Knowles’ principles of andragogy: adults learn best when learning is relevant, problem-centered, experiential, and increasingly self-directed.

The evaluation also showed where to improve. Learners with limited mathematical or programming background need more support, so I recommended:

  • optional pre-workshop readings on foundational math, statistics, and programming;
  • more reflection and discussion periods to consolidate difficult concepts;
  • breaking long lecture blocks into shorter segments;
  • more examples and scaffolded exercises for advanced ideas such as dimensionality reduction and matrix-based thinking.

Limitations. The sample was small (10 post-survey respondents in sessions capped at 5 participants), and all data were self-reported perceptions rather than direct measures of learning. As an observer rather than the instructor, I also could not test changes to the teaching directly. Next time I would add a short skills check before and after the workshop, use the anonymous participant IDs to report individual pre/post changes rather than only group summaries, and collect data across more sessions.

For my own teaching, the project confirmed the importance of balancing guidance with independent exploration, building in experiential learning, and designing instruction that responds to the needs of diverse adult learners. It also showed me how a simple pre/post survey can turn impressions about teaching into evidence I can act on.

Mentored Teaching Project Document