Correlational research

1
Correlational research

• Correlational studies
• Regression and multiple regression
• Factor analysis and cluster analysis

2
Correlational research

• Measured vs. manipulated variables
• Ex post facto studies
• Categorical vs continuous variables

3
Methods of correlation

• Scatterplots
• Correlation coefficients
• Continuous Pearson r
• Categorical
• Spearman r for ranked data on one variable
• Kendalls Tau for two ranked variables
• Cramers statistic, f , for two dichotomous
  variables
• Point biserial for one continuous variable and
  one naturally dichotomous variable
• Biserial for one continuous variable and one
  forced dichotomy
• Tetrachoric for two forced dichotomies

4
Why use correlational designs?

• When it is impossible or unethical to manipulate
  the independent variable.
• When investigating new ideas
• When controlled study would threaten external
  validity
• When we want to predict, using regression

5
Conducting correlational studies

• Control threats to validity
• Control extraneous variables
• Use valid, reliable, and sensitive measures
• Resist the temptation to draw causal conclusions.

6
Correlational studies are also used for

• Reliability assessment .80 minimum
• Inter-rater reliability
• Test-retest reliability
• Equivalent forms reliability
• Split-half or odd-even reliability

7
Validity assessment

• Face validity
• Convergent validity
• Divergent or discriminant validity
• Criterion validity
• Concurrent validity
• Predictive validity

8
Methods of regression

• Regression equations
• Linear regression
• Multiple regression
• Curvilinear regression

9
Uses of regression

• To predict
• Discriminant analysis
• Partial correlation
• Factor analysis