HOW CAN WE UNDERSTAND OUR WATER …

1 HOW CAN WE U N D E R S TA N D OUR W AT E R R E S O U R C E S ? analyzing experimental data T. he information in this chapter is a following format. The layout for the data table short summary of some topics that is based on the variables for the experiment. are covered in depth in the book The first column lists the different values or Students and Research written by Cothron, levels of the independent variable, and the Giese, and Rezba. See the end of this chapter remaining columns list the corresponding for full information on obtaining this book. observed values of the dependent variable.

2 The book is an invaluable resource for any Values for each repeated trial are listed in middle or high school science teacher. separate columns, and then in the last column the average value. The average value will usual- S ETTING U P S IMPLE D ATA TABLES ly be a mean value, but there are some situa- AND G RAPHS tions where a median value or even mode value would be more appropriate. The labels for the Students should be encouraged to set up independent and dependent variable columns data tables and graphs on a computer should both include the measuring unit if whenever possible.

3 This can be done by applicable, and the table as a whole should be first entering data into a spreadsheet given a title. The example below is a data table program such as Microsoft Excel. for a simple experiment to measure the effect of Making simple data the length of a pendulum string on the number of pendulum swings per minute. A simple data table can be set up in the THE EFFECT OF LENGTH OF PENDULUM STRING ON THE NUMBER. OF SWING CYCLES PER MINUTE. Length of Number of Swing Cycles per Minute Pendulum String (cm) Trial 1 Trial 2 Trial 3 Average 30. 50. 70. 90.

4 A N A L Y Z I N G E X P E R I M E N TA L D ATA 6 / 1 3. HOW CAN WE U N D E R S TA N D OUR W AT E R R E S O U R C E S ? Displaying data with in presenting results is in graphical form, and the highest level is when the results can Displaying data from an experiment with be summarized in the form of a mathematical a graph makes it easier to see trends and equation. patterns in the data . The graph presents the data in a visual format that often brings out Depending on the exact nature of an experi- the significance of the data much more clearly ment, students may need to display their data than a data table.

5 There is a hierarchy of ways using different types of graphs. As usual, the to present results from an experiment that best way to make the appropriate decision allows increasing levels of understanding of the about the experiment is to base the decision on meaning of the results. The lowest level is a the nature of the independent and dependent simple qualitative description of the results variables for the experiment in question. The using words. Quantitative data , with measure- table below summarizes the different graphs ments of numerical data , allows for greater that can be drawn in different situations.

6 Understanding. The third level of complexity D ECIDING ON A G RAPH FOR D IFFERENT E XPERIMENTS. Independent Variable Dependent Variable Type of Graph Standard Scale with Equal Intervals Quantitative Bar Graph or Line Graph Examples: Temperature, Qualitative Frequency Distribution Height, Weight, Volume Graph Discrete Categories Quantitative Bar Graph Examples: Day of the Week, Gender, Qualitative Frequency Distribution Color, Brand of Detergent Graph One advantage of drawing a line graph over a A frequency distribution graph is a type of bar bar graph is that extra information can then graph.

7 In a frequency distribution graph, there be obtained from the graph. The line can be will be several bars for each value of the used to infer values of the dependent variable independent variable. The bars represent the corresponding to given values of the independ- number of counts for each possible category ent variable that were not measured in the of the dependent variable. For example, an experiment. A line graph is not appropriate experiment could be carried out to record the when the independent variable is represented leaf quality of plants grown with different by discrete categories.

8 In these cases, intervals amounts of a fertilizer. If the leaf quality is between categories have no significance and a recorded using a rating scale of 1-4, the bar graph is a more appropriate choice. frequency distribution graph would have bars A N A L Y Z I N G E X P E R I M E N TA L D ATA 6 / 1 4. HOW CAN WE U N D E R S TA N D OUR W AT E R R E S O U R C E S ? showing the number of plants with each leaf These sentences can be written directly on quality rating, and sets of bars would be drawn the piece of graph paper. for each different level of fertilizer. There are some important things to require of S UMMARIZING E XPERIMENTAL D ATA : students when they draw graphs.

9 First of all D ESCRIPTIVE S TATISTICS. the values for the independent variable should Making summary data be plotted on the horizontal x-axis and the A simple data table shows all of the measure- values for the dependent variable on the ments collected in an experiment. It can be vertical y-axis. Setting up numerical scales will better to simplify the display of data by usually be the hardest part of the graph for reducing the amount of information with the student to accomplish, and you may need a summary data table, especially if the to help students with this.

10 Certainly, you will experiment is a big or long one. A summary need to check them carefully as they work on data table reduces the information by summa- this. If students are drawing a graph with two rizing the raw data using measures of central numerical scales, they should usually begin the tendency and variation. A measure of central scales of both axes from zero. The reason for tendency is a typical or average value for a set this is that the (0,0) data point often has real of data . This can be a mean, median, or mode significance and should be the starting point value.