So, much more than finding rules that explain the relation between two variables.
Southampton, seminar june 2013
Supporting model-based reasoning usingdrawings and simulationWouter van JoolingenUniversity of Twente
•Science is both experimental andtheoretical.•For learning science one should DOscience•Science is in the process as much as itis in the productView on sciencelearning
Number of people in this roomTime from startPeople coming inPeople going outFame of speakerQuality of presentationQuality of Weather+-+-+Models
Number of people in this roomTime of dayPeople coming inPeople going outFame of speakerQuality of presentationQuality of Weather+-+-• Creating a model of a system that is:• Explicit and external• A reasoning tool• Executable• How does this fit in the learning process of scientificinquiry?Modeling !?
• Modeling is a scientific activity• Learn about science and its nature• Models to make sense of the world• Modeling is learning activity• Understanding complex systems• Understand scientific content• Facilitate reasoning processesWhy?
Model-based reasoning(Nersessian, Giere)• Inquiry and discovery rely on theconstruction of models• Visual and analogical modelling• Mental simulations• Requires creative and constructive activities
KekuléBuilding structure forBenzeneProblem with the bonds &spatial structureIt came in a dream
Kekule• Using visual/spatial representations• With well defined constraints• Valence• External properties of benzene
Constructing models• Adding/changing/removing elements• Check interactions• Check internal consistency• Check correspondence• Do we still predict what we already could?• Validation• Check by new experiments or observation• Mental simulation
Changing perspectiveQuickTime™ en eenH.264-decompressorzijn vereist om deze afbeelding weer te geven.
Emergence• Phenomena may become untraceable whenthere are many objects.• New phenomena may emerge from thecollective interactions
Traffic jamsQuickTime™ en een-decompressorzijn vereist om deze afbeelding weer te geven.
An experimental study among children in the SimSketch learning environmentwith Annika Aukes, & Hannie GijlersThe Use of a Drawing-based Simulation for Modelingthe Solar System
Participants• 288 participants recruited in Science Center NEMO• 39 omitted from data analysis• total of 249 participants• 128 girls and 121 boys• 7 to 18 year old childrenMethod
The Use of a Drawing-based Simulation forModeling the Solar System31
• Domain knowledge test• Pre- and posttest design• Modeling assignment• First task: Draw Solar System (Sun, Earth, Moon, one otherplanet)• Second task: Draw Solar Eclipse• Feedback• Motivational items (Likert scale)• Software attitude (Semantic differential)Method
N Max. score Mean (SD)Score Total Model 235 14 7.25 (3.317)Score Solar System 235 9 5.38 (2.303)Score Solar Eclipse 235 5 1.87 (1.447)Results
Total Model:Differences between 7-9 years (M=6.72) and >12 years (M=8.21)Differences between 10-11 years (M= 6.93) and >12 years (M=8.21)Solar Eclipse:Differences between 7-9 years (M=1.67) and >12 years (M=2.31)Differences between 10-11 years (M=1.69) and >12 years (M=2.31) Children >12 years scored better on their solar eclipse and theirtotal modelResults
Knowledge acquisition• Pre- & posttest differed significantly• The knowledge concerning the solar system was higher inthe posttest.• Greatest differences in 7-9 years and girlsPartial correlation:• The score of modeling had a moderated correlation withthe posttest (r (222) = .286, p = .005) after correction forpretestResults
•Motivation•Perceived competence:• Difference between 7-9 years (M= 3.05) and >12 years (M= 2.71)•Valuing:• No significant differences•Attitude software:• Generally positive rates (M=3.23)• For 85% it was the first time working with SimSketch• For 72% rated SimSketch as a valuable programResults
• Drawing static elements is easier than displaying dynamic processes• Older children (>12 years) are better in modeling• Knowledge acquisition is higher in young children (7-9 years) and girls• However, ceiling effect in older children• Model score relates to knowledge acquisition• Young children have a higher perceived competence• Tendency that young children have higher valuingConclusion & discussion
Overall conclusion• Drawing-based modeling• Accessible to young kids• Based on interpretation of drawings• Work to do!• Integration• Extending the tools• Studies into processes and effects of modelling
Open Questions• How to measure modelling performance?• Result and/or process• Lasting results of modelling with SimSketch?• E.g. visual reasoning,• How to integrate inquiry/modelling in thescience curriculum?