80 Section 6: Statistics Functions
Example: Using the accumulated statistics from the preceding problem,
estimate the amount of sales delivered by a new salesperson working 48 hours
per week.
The reliability of a linear estimate depends upon how closely the data pairs
would, if plotted on a graph, lie in a straight line. The usual measure of this
reliability is the correlation coefficient, r. This quantity is automatically
calculated whenever or is calculated; to display it, press ~. A correlation
coefficient close to 1 or –1 indicates that the data pairs lie very close to a straight
line. On the other hand, a correlation coefficient close to 0 indicates that the data
pairs do not lie closely to a straight line; and a linear estimate using this data
would not be very reliable.
Example: Check the reliability of the linear estimate in the preceding example
by displaying the correlation coefficient.
To graph the regression line, calculate the coefficients of the linear equation
y = A + Bx.
1. Press 0gR to compute the y-intercept (A).
2. Press 1gR~d~- to compute the slope of the line (B).
Example: Compute the slope and intercept of the regression line in the
preceding example.
The equation that describes the regression line is:
y = 15.55 + 0.001x
Keystrokes Display
48gQ
28,818.93
Estimated sales for a 48 hour
workweek.
Keystrokes Display
~
0.90
The correlation coefficient is close
to 1, so the sales calculated in the
preceding example is a good
estimate.
Keystrokes (RPN mode) Display
0gR
15.55
y-intercept (A); projected value for
x = 0.
1 gR~d~-
0.001
Slope of the line (B); indicates the
change in the projected values
caused by an incremental change in
the x value.
y
x
