Computer Introduction -- Introductory Data Analysis Using Quattro Pro Fall 1996

The following was set up to allow you to understand how to analyze data and prepare graphs for your lab write-ups in the future. It is suggested that you use Quattro Pro on the computers in the lab, but any spreadsheet will work. This handout will show you how to analyze and interpret data as well as prepare a graph representing the data.

Let’s assume that an experiment was run for the analysis of Fe in water using a flame atomic emission system and data was collected and is listed on the handout as Table #1. We are assuming that the data should be linear (usually the case, but not always). First, it’s a good idea to graph the data, as in Graph #1. Using Quattro Pro, choose Graphics..New Graph and then select the series for the corresponding X and Y axes. After that choose Graphics..Titles to label axes and give your graph a title. Make sure that you look at the type of graph you have; it should be XY in the Graphics..Type directory. Then, if the calibration should be linear, perform a standard linear regression using the Tools..Numeric Tools..Regression command in Quattro Pro and select the independent variable (concentration Fe) and dependent variable(signal). The R-squared value shows the "linearity" of your data. The closer the R-squared value is to 1.0000, the better. When reporting R-squared values, you only list the 9's and the first digit that is not a 9. So, the R-squared value of Graph #1 was calculated to be 0.98159079, as shown in Regression Output #1, and should be reported as 0.98. If you look at the data points, the point for 500 ppm Fe deviates from the line. This is due to "rollover," which will be discussed later in class. It’s not due to human error, such as not making the solution correctly. Because of this deviation, we are going to eliminate it from the plot.

When that point is eliminated, the R-squared value for the 5 remaining data points calculated in Regression Output #2 is 0.9997, which close enough to linear to be usable for a calibration plot. The regression output generates the slope and y-intercept used to draw the best-fit straight line through the data, which will probably not go through all data points. In Quattro Pro, the term "Constant" is the y-intercept and the "X Coefficient(s)" is the slope. You can use the slope and y-intercept to calculate the regression data points, which are in the column labeled regression of signal. Graph #2 shows the following: red points are actual data points (please note, as in Graph #1, the points are not connected!) and the green line is the regression line without the generated data points, along with proper labeling. If you need help figuring out how to use the graphing part of Quattro Pro, ask. You might want to include grids to see the linearity. You do not have to print the graphs in color, as the example, as this was done for clarification purposes only.

Analyze the following data and turn in a plot similar to Graph #2 with your lab write-up. If it’s not turned in, you will not get credit for the lab (the graph doesn’t have any point value, but we will look at it to make sure you did it correctly).

Conc.(ppm) signal(mV)

50 8.483

100 16.873

200 30.041

250 37.255

500 68.875 <== If it deviates, it’s not due to human error.