Ex post facto research
Archival data
Instrumentation bias
Content analysis
Nonreactive
Participant observation
Naturalistic observation
Scatterplot
Correlation coefficient
Positive correlation
Zero correlation
Negative correlation
Coefficient of determination
Restriction of range
Median
Median split
In this chapter, we address the following questions:
1. Why is description a valuable goal in its own right?
2. Why is it difficult to infer causality from correlational designs? (You should know that people may confuse cause and effect (the "chicken-egg" problem that may lead people to believe that diet colas make people heavy [when the reverse--being heavy leads one to drink diet colas--may be true] and overlook the possibility that sometimes aspirins do cause headaches), as well as overlooking third variables as explanations (two variables may be side effects of some third variable [e.g., self-esteem and good grades could be related because they are both side effects of some other factor, such as supportive parents, physical attractiveness, social class, health, etc.).
3. Why must we use scientific methods (rather than informal observation) to accurately describe behavior?
4. What are the primary problems with archival research? ex-post facto research? observational research? research involving psychological tests?
5. How can the data from a correlational design be summarized?
6. What do positive and negative correlation coefficients mean?
7. What common statistical tests can be used on correlational data?
8. When referring to the results of a correlational study, what does statistical significance mean?