The R-Project is a free software environment for statistical computing and graphics(from website). It is a complex command line program and is usually used in conjunction with a graphical interface. Two possible GUIs are RKWard and R Commander.
This lesson uses freely available statistics software to help analyze the outcome of a scientific experiment. Students will often believe that if one plant grows two inches higher than another that it is definitive proof of something. Statisticians know that it's always a question of probability. "Is there less than a 5% chance that this observation is the result of random chance?"
This activity has students collect experimental data and then use statistical software to help determine if the observations are likely due to the treatment or random chance. This is an important concept and should be discussed with students. If you do not have some experience with statistics, be sure to do a little research before beginning. The software will only tell you what you ask. It doesn't tell you want the data means. Interpreting the results is up to you and your students.
- This activity is appropriate for basic scientific data analysis in upper elementary or middle school (could also be adapted to high school). It is tied to a data collection activity in science but can be easily adapted to any type of data activity. The only limitation is your imagination.
- The activity will take several weeks to complete. The majority of the time is simply the data collection of growing plants. The actual data analysis should take no longer that an hour and even less if you input the data. However, I would encourage you to let students input the data and run the analysis. It will be fun for the students and more meaningful. If you want to use a different experiment as the source of the data that is certainly possible. Again, feel free to adapt it to whatever activity works for you.
Computer lab with R software and at least one GUI front-end installed
- A journal for collecting daily observations
- Any reference materials you feel are appropriate
TIPS AND SUGGESTIONS
- Go through a sample data analysis at least once yourself prior to having students do it.
- Pick a time when you will have about a month without vacations and other interruptions.
- This will probably work best as a class activity with various small groups responsible for different tasks (e.g. one group plants all plants, one group waters all plants, one group conducts the actual measurements, etc). However, have each student keep their own record of observations. You could also have each student measure all plants. It would be more time consuming be each student would have their own data to analyze. You could then run an analysis on each separate student's data and compare. Play with this as you see fit.
- At some point, encourage students to share information with another student or group.
- Download and modify the handout to fit your needs. It is in both Open Office and MS Word formats.
- Students will conduct a scientific experiment and collect data
- Students will understand about independent and dependent variables
- Students will learn the importance of accurate observation and data collection
- Students will be able to keep an accurate log of observations and data
- Students will input their data into the statistical software and run the appropriate analysis.
- Students will draw conclusions about their experiment and use the statistical output to support those conclusions.
The first part of this lesson involves growing plants and making daily observations. The second part is the actual data analysis.
- You will begin by planing some sort of plant. Beans are common. You might have access to a FOSS (Full Option Science System) kit and that would work perfectly.
- In order to conduct research, you have to have a research question. Here are a couple to get you started. Work with one of these or use one that ties into your curriculum.
- How will different amounts of sunlight affect plant growth?
- What is the optimal amount of water?
- Is one fertilizer better than another or better than none at all?
- I'll use the third one for this example. Pose the question to students, develop a hypothesis and plan for testing it.
- In this case, it would make sense to plant about 30 plants. (You can think roughly one plant for each student in a class but don't tell students they each get a plant. That won't work in a controlled experiment.) 10 plants will be the control and will receive no fertilizer. 10 plants will get brand x (pick one, it doesn't matter. Ask the kids) and 10 plants will get brand y.
- Be sure to talk with students about the importance of holding all variables constant except the one in question (fertilizer). The depth of the seed, amount of water, sun, air, dirt, type of dirt used, type of water and on and on can effect the growth of the plants. In order to determine the effect of the fertilizer, all these other variable must be identified and controlled. Discuss this with the class.
- While each student will probably want control over their own plant, that is also a variable. The class should choose several students who will be in charge of planing, watering and so on. You could assign a small group to plant, another group to water, another group to measure and so on but the same people should do the work for all plants. Avoid the temptation to let each student do their own planting and watering. Even if you take measure to control their various differences, it is important that they learn the importance of controlling variables.
- Develop a system for planing, watering and general care. Begin the process.
Use the attached handout (Open Office version) (MS Word version) or something similar to begin data collection. Do not wait for the plants to sprout. Begin data collection for all plants at the same time. To avoid a bunch of zero data you could wait for the first one to sprout or simple begin after 5 days or so. You should also make plans for collecting data over the weekends. Collect data for a good three weeks or more. More data means better results. 30 days might be a good goal.
- Measure the height of the main branch in cm each day and count the number of leaves on all branches. You could adjust this to meet other curricular objectives but hight and number of leaves are different types of measurements and will help the analysis and conclusions.
- Once all data is collected, begin part 2.
note: This section is in progress. Feel free to add to it if you'd like.
This part involves the actual data analysis. Assess your students abilities and solicit outside help if you think they will need it (e.g. parents). Inputing the data could be the most time consuming and the potential for error is high. However, this is where the fun starts so try to let students do as much as possible. Use parents to monitor rather than to do the work.
- Use a spreadsheet or similar program to input the data. Use two columns. One column for height and one for number of leaves. Put in one day of data on each row. Use headers if you think students need them but it's best to remove them at the end. You want nothing but numbers for the software to read in.