Probabilities and predictions: modeling the development of scientific problem-solving skills

Area of Science:

• Educational Technology
• Molecular Genetics Education
• Computational Biology in Education

Background:

• The Interactive Multi-Media Exercises (IMMEX) platform facilitates online multimedia simulations for genetics education.
• Utilizing IMMEX data for formative assessment and understanding student learning is a key next step.
• Previous research has not fully detailed the evolution of student problem-solving strategies in molecular genetics.

Purpose of the Study:

• To develop probabilistic models of undergraduate student problem-solving in molecular genetics.
• To analyze the spectrum of strategies students employ and how these evolve with experience.
• To identify factors influencing strategy selection and learning progression.

Main Methods:

• Recorded and analyzed the actions of 776 undergraduate biology majors across six molecular genetics simulations.
• Employed artificial neural network clustering to group simulation performances and derive individual measures.
• Utilized hidden Markov modeling to probabilistically model sequences of performances and assess progress.

Main Results:

• Probabilistic models revealed that students with different initial problem-solving abilities adopt distinct strategies.
• Observed variations in initial and final strategies across different course sections, with no strong correlation to other achievement measures.
• Found no significant gender differences in problem-solving strategies, contrary to some previous studies.

Conclusions:

• Student problem-solving strategies in molecular genetics are diverse and influenced by initial abilities.
• Early instructor interventions, informed by simulation performance, can guide students toward more effective strategies.
• This approach holds potential for enhancing student learning and improving the efficiency of problem-solving in molecular genetics.