Making Systematic Observations

Question 1

Deciding what to Observe

Answer 1

Exposure
Through professors
Through participating as a subject
Through reading popular press articles
Through reading peer-reviewed research in journals

Where are the holes?
Where are the inconsistencies?

Question 2

Operational Definitions

Answer 2

Operational definitions are important
Start with a construct (a broad general “thing” like “self-esteem” or “risky sex” or “intelligence”)
Define it further- BE SPECIFIC
Risky sex could be measured by- # of partners, # of partners not use condom, # of times not use condom, % of partners not use condom, etc.
How would we measure alcohol abuse?
Quantity by frequency (how often x how much)
Alcohol dependence scale
Lifetime drinking history

Question 3

Self Esteem is a Construct

Answer 3

Self-esteem is a construct.
Operationalized, we mean “someone’s positive or negative evaluation of themselves”
We search the literature for self-esteem scale or self-esteem measure
In the literature, it looks like Rosenberg’s self-esteem scale is most frequently used

Question 4

Reliability

Answer 4

Reliability?
Can I get the same results more than once?
If I measure alcohol abuse two times within one week, I should get the same results (or very similar results)

Question 5

Validity

Answer 5

Validity?
Am I really measuring alcohol abuse?
Can I predict some behavior?

Question 6

Scales oF Measurements

Answer 6

Nominal- categorizing 
1 = women
2 = men
Ordinal- rank ordering
1 = first place
2 = second place
Interval- rank order but no true zero
1 = strongly disagree; 7 = strongly agree
Ratio- rank order with a true zero
Height, weight

Question 7

Reducing Participant Bias

Answer 7

Hawthorne Effect- subject tries to “do really well”
If you put a lot of effort into the task, then you aren’t acting naturally
Social Desirability
Attempting to display yourself in the best light.
I NEVER drink. I ALWAYS help others.
Reduce bias…
Manipulation checks
Single-blind study

Question 8

Reducing Experimental Bias

Answer 8

What is my agenda?
Experimenter’s race, gender, friendliness, attitude- all affects participant
Reduce experimenter bias…
Mechanize the protocol
Double-blind procedures

Question 9

Forms of Random/Probability Sampling

Answer 9

Simple random sample- each member of the population has an equal chance of being selected as a member of the sample
Stratified sample- the proportions of important subgroups are represented.
Population: 60% female, 40% male
Sample: 60% female, 40% male
Cluster sampling- randomly selected clusters of people who all have something in common.
8th graders from Michigan, Iowa, California