Proposal Title

Changing students into experimenters with simulation-based feedback

Presenter Information

Eli Meir, SimBioFollow

Session Type

Presentation

Room

P&A 34

Start Date

6-7-2017 10:30 AM

Keywords

virtual lab, scientific process, experimental design, biology, simulation

Primary Threads

Education Technologies and Innovative Resources

Abstract

A challenge of using student-centered approaches in large classes is providing useful feedback from open-ended exercises. Simulation-based labs allow students to practice skills that are often too time­ or resource ­intensive for classrooms, such as designing experiments, but there is a tradeoff between simulation freedom and providing detailed feedback: Low­ constraint labs give students freedom, but this open­-endedness defeats attempts to provide specific automated feedback. Conversely, highly constrained labs limit student exploration, but allow automated feedback.

I'll present a new simulation-based lab we wrote called Understanding Experimental Design, available from SimBiotic Software (free trial available to conference participants) where biology students hypothesize and then experimentally determine the cause of a mysterious illness. With the lab, we asked: how does lightly constraining an experimental design task (such as presence / absence of causal factors rather adding any amount) and thereby enabling student-specific feedback, affect student learning?

Our preliminary data shows that learning was higher when the lab had constraints and feedback ("medium-constraint") than when we removed some constraints and the feedback ("low-constraint"). The percent of students who designed a “good” experiment (systematic variation of variables; valid controls) was 79% in the medium-constraint lab as compared with 36% in the low-constraint version (N=14 per treatment), and good replication showed a similar pattern (79% vs 43%, respectively). In addition to taking our lab home for a classroom trial, participants may get ideas for constraining other open-ended activities they do in large classrooms.

Bring your laptops to play along.

Elements of Engagement

Will bring usb drives with the Understanding Experimental Design lab so all participants can try it for themselves during the presentation - including designing their own experiments to deduce the likely cause of the mystery illness. Will also generate some discussion of what features participants feel are responsible for the learning we see with the lab, and compare that to the data we have from our study.

This document is currently not available here.


Share

COinS
 
Jul 6th, 10:30 AM

Changing students into experimenters with simulation-based feedback

P&A 34

A challenge of using student-centered approaches in large classes is providing useful feedback from open-ended exercises. Simulation-based labs allow students to practice skills that are often too time­ or resource ­intensive for classrooms, such as designing experiments, but there is a tradeoff between simulation freedom and providing detailed feedback: Low­ constraint labs give students freedom, but this open­-endedness defeats attempts to provide specific automated feedback. Conversely, highly constrained labs limit student exploration, but allow automated feedback.

I'll present a new simulation-based lab we wrote called Understanding Experimental Design, available from SimBiotic Software (free trial available to conference participants) where biology students hypothesize and then experimentally determine the cause of a mysterious illness. With the lab, we asked: how does lightly constraining an experimental design task (such as presence / absence of causal factors rather adding any amount) and thereby enabling student-specific feedback, affect student learning?

Our preliminary data shows that learning was higher when the lab had constraints and feedback ("medium-constraint") than when we removed some constraints and the feedback ("low-constraint"). The percent of students who designed a “good” experiment (systematic variation of variables; valid controls) was 79% in the medium-constraint lab as compared with 36% in the low-constraint version (N=14 per treatment), and good replication showed a similar pattern (79% vs 43%, respectively). In addition to taking our lab home for a classroom trial, participants may get ideas for constraining other open-ended activities they do in large classrooms.

Bring your laptops to play along.