|
The Scatter
Diagram
<Back to Page 1>
Not
all sets of data pairs will exhibit a strong positive correlation, even
if an increase in the x-value somehow results generally in an increase
in the y-value. An example of this 'weak' type of positive
correlation is shown in the scatter diagram of Figure 2b, which is said
to exhibit just a
'possible
positive correlation.'
This
scatter diagram still shows a perceivable diagonal line going in the
upper right direction, but the points are more spread apart than in a
scatter diagram with strong positive correlation.

Figure 2.
Scatter Diagrams showing positive correlation (a, left) and
just a possible positive
correlation (b, right)
If the scatter diagram
formed also shows a perceivable diagonal line, but the line is going
in a direction opposite that of positive correlation (i.e., from the upper left to the lower right corner) as shown in Figure 3a,
then the data pairs are exhibiting
negative correlation. This
means that y decreases as x increases.
Again, the negative correlation is strong if the line formed by the data
points is narrow and very defined.
If the negative correlation
is not strong, resulting in data points that are not closely packed
together, then there is just a
'possible negative
correlation.'
An
example of a scatter diagram for such type of correlation is shown in
Figure 3b.

Figure 3.
Scatter Diagrams showing negative correlation (a, left) and
just a possible negative
correlation (b, right)
Of course, more complex
types of correlation may also be identified using a scatter diagram.
Once a type of correlation is established, the engineer may choose
to proceed with a further and more in-depth investigation of the
correlation using other analysis tools.
Determining the exact nature
of correlation between variables can lead to benefits. These
include: 1) better understanding of cause-effect relationships; 2)
reduction of data gathering requirements; 3) establishment of more
effective process controls; 4) easier development of check and balance
schemes; etc. To realize these benefits, however, the engineer has
to use other analytical tools to complement the scatter diagram, since
the latter is only used as a quick visual check for possible correlation before
a more in-depth study is undertaken.
<Back to Page 1>
See Also:
Matrix Diagram; Ishikawa Diagram
HOME
Copyright
©
2004-2005
SiliconFarEast.com.
All Rights Reserved. |