What abstract, broad ideas interest us?
Examples: self-esteem, aggression, prejudice, gender roles, dissonance
How can we measure them?
Operational Definitions
We measure variables by coming up with operational definitions
The more specific, the better
Multiple ways to operationally define a conceptual variable is ideal
- converging operations
Conceptual Operational
Attraction
Do P's ask for female confederate's phone #?
Self-Esteem
P's response time to positive self-descriptors
Prejudice
Do P's mail "lost" letters addressed to NAACP?
Pro-Social Behavior
P's reported donations to charity in the past year
Intelligence
P's scores on a standardized IQ test
Variable Measurement
Nominal variables are categorical
I want to know what type of high school Pís attended:
1 = public
2 = private, secular
3 = private, religious
The number assignments are arbitrary and nonmathmetical
Quantitative Variables
Numbers are not arbitrary
Three types:
Ordinal
Interval
Ratio
Mnemonic Tip: NOIR
* Each one is more mathematical/sophisticated than
the previous one
Ordinal Variables
Includes most scales used in social psych
* There is meaning to the numbers, as 2 is greater
than 1, 3 is greater than 2, etc.
* But the interval between numbers is not standardized
Ordinal variables tell us order, but not relative distance
Interval Variables
Again, these tell us about order
Scales have standardized differences between numbers
Example: IQ
- The difference between
100 and 110 represents the same difference as between 140 and 150
- But the midpoint of 100
is arbitrary
- There is no absolute zero;
an IQ of 140 does not mean you are twice as smart as an IQ of 70
Ratio Variables
These scales do have an absolute zero
Example: Money
- The difference between
$6 and $5 is the same as between $113 and $112
- Also, unlike an interval
scale, you can say someone with $140 has twice as much as $70
- $0 is an absolute lack
of money
Why do we care?
Different statistical techniques appropriate for different types of scales
Determines what we can do to the data
- Ratio data can be multiplied, divided, averaged,
etc.
- Makes no sense to multiply or average nominal
data!
Back to Overheads Main Page
Back to 381-002 Main Page