Professional Documents
Culture Documents
6/12/2005
X-Y Plot
Summary
The X-Y Plot plots the data in two numeric columns to illustrate the relationship between them.
The points may be coded according to the value of a third variable or connected by a line if
sequential. If the points represent averages, standard error bars may be added to illustrate the
uncertainty in X and/or Y.
Sample Data:
The file 93cars.sf3 contains information on 26 variables for n = 93 makes and models of
automobiles, taken from Lock (1993). The table below shows a partial list of 5 columns from
that file:
Data Input
The data to be analyzed consist of two numeric columns containing n = 2 or more pairs of
observations.
Analysis Summary
The Analysis Summary shows the number of pairs of observations plotted.
93 values
Scatterplot
45
40
35
30
25
20
1600 2100 2600 3100 3600 4100 4600
Weight
Pane Options
The Point Codes option is used to illustrate groupings of the data, such as the number of
passenger seats in each vehicle:
45 4
40 5
6
35 7
8
30
25
20
1600 2100 2600 3100 3600 4100 4600
Weight
The Std. Error fields are used to plot error bars around each point, often used when the points
represent averages. For example, the table below shows the sample means and standard errors
calculated by the Oneway ANOVA procedure for the 93 automobiles grouped by number of
passengers:
# Cylinders Count Mean MPG Stnd error Mean weight Stnd error
Highway MPG Highway Weight
3 3 43.3333 3.75648 1901.67 105.883
4 49 31.449 0.587614 2710 53.2269
5 2 24.5 3.5 3602.5 357.5
6 31 25.3226 0.445738 3559.19 47.5723
8 7 25.0 0.534522 3835.71 92.2886
To display the data, create an X-Y Plot of Mean MPG Highway versus Mean Weight and
complete the Pane Options dialog box as shown below:
The resulting plot shows the means for each group, together with an indication of the sampling
error in each direction:
51 # cylinders
3
46 4
41 5
6
36 8
31
26
21
1700 2100 2500 2900 3300 3700 4100
Mean Weight
The smaller samples, such as the 3 and 5 cylinder automobiles, have more uncertainty associated
with them as indicated by the long standard error bars.