More examples in Legal/Criminal Justice
Types of Relationships in Gangs Example
SmartDraw includes thousands of professional-looking diagrams like this that you can easily edit and make your own.
Text in this example:
Types of Relationships
Non Spuriousness:
Assuming a relationship
between two variables (X and Y), we must
learn whether a third variable (Z) causes the
changes in them. We can detect spuriousness
statistically by controlling for Z. If the
relationship between X and Y disappears
when controlling for Z, the relationship is
said to be spurious. That is, our original
observation that X and Y are related is
mistaken. They seemed related because of
their mutual relationship to Z. (See Exhibit
49 for an illustration of spurious and
non spurious relationships.)
X Must Precede Y in Time:
That is obvious unless we do not rule out
the possibility of reverse causation. In program
evaluation, the program’s strategies must be
operating before effects take place if we
are to attribute cause to the program.
X and Y Must be Related:
Correlation between the two variables
must exist. That is, they must co-vary in some
way showing they are connected. If the two
variables act independently (that is, have no
relationship), they vary in ways unconnected
to each other. X Non Spurious Y Z X Y Spurious Source: ADDRESSING COMMUNITY GANG PROBLEMS:A Practical Guide
Non Spuriousness:
Assuming a relationship
between two variables (X and Y), we must
learn whether a third variable (Z) causes the
changes in them. We can detect spuriousness
statistically by controlling for Z. If the
relationship between X and Y disappears
when controlling for Z, the relationship is
said to be spurious. That is, our original
observation that X and Y are related is
mistaken. They seemed related because of
their mutual relationship to Z. (See Exhibit
49 for an illustration of spurious and
non spurious relationships.)
X Must Precede Y in Time:
That is obvious unless we do not rule out
the possibility of reverse causation. In program
evaluation, the program’s strategies must be
operating before effects take place if we
are to attribute cause to the program.
X and Y Must be Related:
Correlation between the two variables
must exist. That is, they must co-vary in some
way showing they are connected. If the two
variables act independently (that is, have no
relationship), they vary in ways unconnected
to each other. X Non Spurious Y Z X Y Spurious Source: ADDRESSING COMMUNITY GANG PROBLEMS:A Practical Guide