How is conducting experiments similar/different when using computer models?

Posted April 17, 2017 by ilee

Reflect on the Science practices you teach with regards to conducting experiments in class. How is it similar/different from the experimental design enabled with computer models?

Post your thoughts to the forum below, and come back later to read other entries and comment on another teacher's response.

Comments

Submitted by eso on Fri, 06/02/2017 - 23:08 · Permalink

My immediate thought is the need to teach lab safety and proper handling of and use of equipment. In a hands-on experiment, I'm always worried about measuring out materials, replenishing supplies for each class, and making sure that the scholars are handling things the right way. 

With computer models, those safety skills would only have to be taught once with proper handling and care of the laptops/ computers. Materials don't need to be physically set up, instead the scholars are responsible for creating or manipulating the agents and environment virtually.

While the computer models aren't physically hands-on, they are similarly engaging and provide a different kind of experience for the scholars to learn from.

Submitted by Shaw'sae Dodson on Sun, 06/04/2017 - 08:37 · Permalink

How is conducting experiments similar/different when using computer models?

When using computer models you can thoroughly explore what you have modeled. If you run your model and find that you have left something out, you can modify your model to include it. Once the model is ran it doesn't cost anything but time. This is more efficient in some ways rather than conducting an experiment in class.  

Submitted by Bright_Eyed_Science on Thu, 06/08/2017 - 17:28 · Permalink

Reflect on the Science practices you teach with regards to conducting experiments in class. How is it similar/different from the experimental design enabled with computer models?

There are many similarities between physical experimentation and experimental design enabled with computer models. These similarities include:

  • the need to have a well-developed experimental question centered on a impact of single independent variable on an observable phenomena, 
  • a thorough procedure,
  • collection and analysis of data,
  • reflection on limitations, 
  • and the importance of synthesising.

Computer model enabled experimentation differs from traditional physical experimentation in that there is:

  • a need to understand the "black box",
  • a decreased reliance on physical skills (e.g. fine motor skills required for measuring, operating specialized tools) and increased accessibility for students with special needs, 
  • the ability to manipulate spatiotemporal scales to fit within the confines of a classroom and class period, 
  • a more tangible view of complex interactions by being able to manipulate agents and interactions in a way that is either not possible or feasible in the physical world or classroom, 
  • and far less safety concerns.

Submitted by jsimpson on Fri, 06/09/2017 - 13:02 · Permalink

Reflect on the Science practices you teach with regards to conducting experiments in class. How is it similar/different from the experimental design enabled with computer models?

The differences that I see right away is around the use of lab tools (beakers, rulers, scales, etc.).  I don't see that students will not need this kind of training for this particular area.  This will need to take place when doing controlled experiments in class.  However, I can see not using canned science experiments and instead only doing controlled experiments when the student decides they need more information and needs to do an actual controlled experiment.  There are also more parts to understand using this model.

Something similar is identifying variables, asking questions, etc.  It looks like there is a format that we can teach to students and I really like the thought process.  

Submitted by llegault on Tue, 06/13/2017 - 14:57 · Permalink

When my class does an experiment, we follow the traditional scientific method of starting with a question and forming a hypothesis, and from that designing the experiment, changing the variables and producing a result that proves or disproves the hypothesis.  Because I teach three science classes with a total of 80 students in homerooms, I have limited space for the products of the experiment.  I could do the same types of experiments with computer models but I could expand the variables and run similar experiments in less time and take up less space, with more hypotheses being formed and experimented on.  I do think my fifth grade students enjoy the actual physical experiments so a combination of real world and computer simulation would be ideal. 

Submitted by hjoyal on Tue, 06/27/2017 - 16:51 · Permalink

I think there is always a time and place for real world experiments, but as a science teacher with limited funds and computer generation kids I see that the use of computer simulated experiments can lead to the testing of very abstract concepts and the sky is the limit, not the bank account for what we can do in the classroom. I also agree with the idea that the construction of the model can lead to the formation of a hypothesis. We, as teachers, frequently skim over the hypothesis stage but in many cases it is the most important and where the background learning has to be strong to make assertions about the data and draw conclusions that are knowledgeable. 

Submitted by bonitagirl on Mon, 07/17/2017 - 21:22 · Permalink

In using a real life model experiment, I often run into not having enough materials, or it's too costly. In a computer model experiment, I can have plenty of materials (maybe ones that I hadn't thought of) and my students can run them over and over without worrying about using up too much materials or not being able to do something because of safety in the classroom. 

Submitted by jfretz on Fri, 07/28/2017 - 19:10 · Permalink

Traditional classroom experiments and experiments enabled with computational models both need to be controlled by isolating a variable to be tested and controlling all other variables.  However, using computational models makes it much easier to control the other variables, because one can simplify the model to include only those variables one thinks are relevant to the problem.  Additionally, in computational science experiments, one can actually change the behavior of the components (which is usually impossible with real-life experiments) so if one discovers an error or another way to approach the problem, changes to the experimental design can be made without having to scrap the entire project.  Replication of the experiment is also much easier with computational science experiments, since automation mostly removes the possibility of human error.  Overall, it seems that experimental design using computational science is much more flexible and allows one to study problems without such a firm commitment to one particular hypothesis.

Submitted by Amy_Myers on Sun, 07/30/2017 - 15:26 · Permalink

Reflect on the Science practices you teach with regards to conducting experiments in class. How is it similar/different from the experimental design enabled with computer models?

Currently, conducting experiments in class requires much preparation on part of the instructor. Before any lab can be initiated, students must first past a safety test and be informed of all measures of safety within the classroom. Then, I need to be sure that all materials are available. This may mean store bought items, prepared solutions, etc. Most of the time, students work in pairs or in a group due to resource limitations. Data is gathered as a class so that students can make their conclusions based off of many "repetitions" of the experiment. The hypothesis is analyzed through this experimentation and inferences are announced during class discussions and lab reports. 

Experimental design enabled with computer models allows for very little safety instruction. Students must be simply monitored for appropriate use of the computer. Available materials are irrelevant, due to our computational model having almost infinite options. Also - students can work alone and therefore take responsibility for their work. Students can also very easily repeat a simulation or change the variables in order to create a hypothesis OR make a conclusion. Students can change the environment and simulate factors that would be unavailable in a classroom setting. Although it is not as "hands-on" as actual lab experimentation, it requires students to think differently, and therefore is a different form of learning. 

Submitted by person11d on Fri, 08/04/2017 - 23:20 · Permalink

A lot of the science concepts I had in mind to teach in my classroom involved me trying to find ways to make abstract ideas that are hard to implement in the classroom into a hands on project that my students could relate to. I was worried that some of the content would be challenging to create a physical example for to display to my students. Computer science provides that exact means I've been looking for. By teaching my students how to code, they can create experiments that mimic the real world to answer their scientific inquiries through their own observation and data collecting. 

Submitted by jstblue on Sun, 08/27/2017 - 14:35 · Permalink

Experimental design with computer models seems to give more flexibility (can change behavior and control variables) and speed in running experiments.  It also seems more cost effective and that the teacher will save time and energy on safety procedures and organizing materials.  Similarities are that the scientific method remains in place.  Students still need to be able to ask good questions, analyze and synthesize data, etc.