Press [2nd STAT] or [2nd LIST], then select "EDIT" from the menu. If you have old data in the lists xStat and yStat, clear the lists by doing the following: put the cursor on the name of the list xStat (up at the top of the screen), press the CLEAR button, then ENTER or the down arrow; do the same thing for yStat.
Enter your data, putting the x-coordinates in the list xStat, and the y-coordinates in yStat. Use the up/down arrows to move up or down the list, and the right/left arrows to move between xStat and yStat. If you make a mistake, you can correct an entry by typing over it. Use the DEL key or 2nd INS to delete or insert an entry.
In the list editor, to the right of yStat is another list called fStat. This is the frequency list, which indicates how many copies there are of each point. For example, if your first point has coordinates (0,55) and for some reason you needed three points with those coordinates, you would enter 0 in xStat(1), 55 in yStat(1), and 3 in fStat(1).
For our purposes, all of the frequencies need to be 1, that is, we will always have just one of each point. You can either type in ones in all of the entries of fStat (but only as far down as the end of your xStat and yStat lists, since all three lists need to be the same length), or, you can exit and run the program ONES, which will automatically put in the ones for you (exit from the list editor, press PRGM, then select "NAMES", "ONES", and Enter).
Go to GRAPH, y(x)=, and turn on Plot1. Put the cursor on "Plot1", up at the top of the screen, and use "SELCT" (F5) to turn it on. It is on if there is a dark box around the name "Plot1"; otherwise it is off. If there are functions in your y(x) list that you do not want graphed, use "SELCT" (F5) to turn them off (or CLEAR to get rid of them entirely). Set an appropriate window for your data (set xMin a little less than your lowest x value, xMax a little higher than your highest x value, and similarly for y). Then select "Graph" from the menu to display your scatter plot. If the menus are obscuring some of the points, press the CLEAR button to get rid of them (and EXIT to get them back).
If you can't get your scatter plot to work, and you know that it is turned on, your data is entered correctly, and the window is set appropriately, check the plot settings. Ask your instructor if you need help with this.
With your data entered in xStat and yStat, and ones in fStat, go to [2nd STAT] (if you are in the list editor, Exit first), and from the menu select "CALC" (calculate), select the type of regression (best fit function) that is appropriate for your data; don't press ENTER just yet. Use:
LinR for a linear regression, y=a+bx
ExpR for an exponential regression, y=a*(bx)
P2Reg for a quadratic (degree 2 polynomial), y=ax2+bx+c
P3Reg for a cubic (degree 3 polynomial),
y=ax3+bx2+cx+d
LgstR for a logistic regression, y=a/(1+b ecx)+d
There are other types of regressions on the STAT CALC menu which we won't use in this course. Be sure not to confuse LinR, which stands for Linear Regression (one we will often use), with LnR, which is a logarithmic regression (one we will not use).
Most of the time we will want to graph our best fit function. The easiest way to do this is to have the calculator save the function to y1 (or y2 or wherever) when it computes it. To do this, simply type in y1 (or y2 or whatever) immediately after selecting the type of regression, before pressing Enter. For example, if you are finding a linear regression, it would look like this:
LinR y1
(Use 2nd alpha to get the lower case y.) Then when you press Enter, the calculator will find the best fit line (or whatever function you select) and save it directly to y1. You can check that it is there by going to GRAPH, y(x)=. When you display the graph, the calculator will graph the best fit function along with the scatter plot of the data (assuming that you left Plot1 turned on).
Once you have your best fit function graphed, you can use it to interpolate (find a value between data points) or extrapolate (predict a value outside of the data--of course this is much less reliable than interpolating). As usual, if you know the x-value and need to find y, use TRACE (press the down arrow to switch from Plot1 to y1 before you type in the x-value); if you know y and need to find x, graph the horizontal line and find the intersection (MORE MATH MORE ISECT).
Another way to find x or y values from your best fit function is to go to 2nd STAT FCST. ("FCST" stands for forecast.) This works sort of like the Solver, but there is no equation to type in; it automatically uses the most recently computed regression equation. Note that FCST will always work if you need to plug in a value for x and find y, but if you know y and need to solve for x, FCST will only work if the function is one-to-one.
When you are through with your data, it's a good idea to turn Plot1 off. (To do this, go to the y(x) list and use SELCT.)
Page written and maintained by
Helen Read.