NonExperimental Research

1
Non-Experimental Research

• Corey and Lachter chapter 6

2
What is non-experimental research?

• Take people as they are and measure them
• Have already conducted one type Correlation
• (terminology and use of correlational study to
  refer to non-experimental)
• Other types of non-experimental key question
  are groups different from each other

3
First, though

• Two types of pre-non-experimental research to
  address
• Demonstration
• Case study
• Both first step in research
• Demonstration how do people behave in a certain
  situation?
• Just one variable
• Case study how does one person behave?

4
Demonstration

• Why use?
• Asch, as demonstration
• Milgram, as demonstration
• Advantages?
• Disadvantages?
• More modern day possibilities? (either actual or
  that would make sense to do)

5
Case study

• May have unique case
• Psychological disorder
• Unique history
• Like a demonstration, with one person, or like a
  non-experimental version of a single subject
  design how does this person behave
• Can then do follow-up research, to test whether
  certain variables are playing a role

6
Case study

• Examples?
• Advantages?
• Disadvantages?

7
Back to more traditional non-experimental designs

• Correlation
• Key question?
• When use?
• Advantages?
• Disadvantages?

8
Other types of non-experimental

• Static groups
• Ex post facto
• Time series

9
Static groups

• People are in groups that could be changed, but
  come into the study in a pre-existing group
• For ethical reasons, cannot change
• Want to know if two groups differ on some variable

10
Static groups

• Examples?
• Advantages?
• Disadvantages?

11
Ex post facto

• Like static groups, but people are born into the
  group
• Again, for ethical (and practical) reasons,
  cannot change

12
Ex post facto

• Examples?
• Advantages?
• Disadvantages?

13
Time series

• Key question did some event change people?
• Predictor variable before or after event
• Criterion variable some behavior
• Again, cannot manipulate predictor variable
• Why not?

14
Time series

• Examples?
• Advantages?
• Disadvantages?

15
Another way to divide study designs

• Think back to the distinction between
  independent-samples t-tests and dependent-samples
  t-tests
• Between-subjects
• Within-subjects

16
The good and the bad of non-experimental

• What can non-experimental designs tell you about?
• What can they not tell you about?
• What sorts of big questions could you address
  with them?
• What sorts of hypotheses could go with those big
  questions?
• What other causal paths could there be?

17
Applying it all

• Whats the big question, the type of design, and
  what other causal variables?
• Teaching technique and student performance
• Day/evening classes
• Spending habits before and after Katrina
• Traffic patterns in London after charges for
  driving downtown
• Gender and math ability